DataQuest

DataQuest

DataQuest

' n ' n a company of i i i l The Dun &Bradstreet Corporation 251 River Oaks Parkway, San Jose, CA 95134-1913

(408) 468-8000 Fax (408) 954-1780 r

November 30,1995

Errata

In Dataquest's 1994 Stepper Market: Reflection of the Growing Strength in Semiconductor

Production, dated November 13,1995 (SEMM-WW-MA-9505), Figure 1 contains an error in the data given for Asia/Pacific and Europe. The correct data is as follows:

Europe 14%

Asia Pacific 24%

Please fQe this notice with the original document in your Semiconductor Equipment,

Manufacturing, and Materials Worldwide binder. Dataquest regrets the error and apologizes for any inconvenience.

For further information, contact Nader Pakdaman, Senior Industry A f S ^ i , wt (408)

46?^^ ±17 or at [email protected]

H:."

.rj.'ir;!. •! -^ -'.^.iYft^^ •• ^,n'

'•>•_ • :^:--^->.j:.nti:^:-;''*:;:'::^-.

V- ^. -: ^ .-••r.'¥^(P<^;= "i ^:;,;

'fV-^; !'.;^ii:;yii.j^; :;:^.afv-<. • ' ,-/..-. .-

- • - - 0 /

4.'^^^.-'.

=jv-

1?; fcjst/^:- •

• ' .

^ "''•^^':}:h

.-?-'

.>Oi.

•fc

^•i-fl

DataQuest

B S fh?DuS&^Bradstrcetcorporaiion 1290 Ridder Park Drive, San Jose, CA 95131-2398

(408) 437-8000 Telex 171973 Fax (408) 437-0292

December 1994

Dear Dataquest Client,

For more than 23 years, Dataquest has helped clients grow and prosper by supplying top-quality market intelligence on the information technology industry. We are proud to be your information partner through providing market analysis, competitive assessments, end-user preferences, and strategic insight for making crucial business and planning decisions.

The enclosed binder is for filing and storing the market research newsletters and reports that you will receive on an on-going basis throughout 1995 as part of your subscription to this research program. The researdi coverage and publications that you can expect to receive in 1995 are detailed on the enclosed program datasheet.

In addition, for those clients who wish to print and file the electronic newsletters and Dataquest Alerts that are part of this research service^ an J^ectronic News bliuler is available on request. To order your copy, singly fill out ^me FaxBadc form located in the binder pocket and fax it to us.

I look forward to working wilh you in our continuing process to im^msve the ccmtent, quality, and delivery of Dataquest's products ami service. U you have questions about our research programs, publications ddivery schedule, or on-g<»ng sCTvices, please call your analyst/account manager, or y(m may call ms direcfly at (408) 437-8293.

SncCTdy,

J c ^ e y A, Byrne

Vice President

Worldwide Marketing

Dataquest Fax Back

To:

City:.

Kimberlie Edwards

San Jose, California

Co.:

Coiintry:.

Dataquest Incorporated

U.S.A.

.Total Pages 1 of 1

Here's How to Order Your Electronic News Binder

Dataquest provides a separate binder called

Electronic News to help you organize your printouts of the electronic newsletters and

Dataquest Alerts that will be sent to you by this research program throughout the year.

Although not all clients will print out electronic news bulletins or file faxes, the Electronic News binder is available by request for those who do.

To order your Electronic News binder, just fill out the form below and fax it back to us. We will mail your binder to you immediately.

Note: If you subscribe to more than one

Dataquest research program, then indicate how many binders you need in the space provided below (plan on one binder per research program), and we'll send them to you in one shipment.

Thank you for helping us serve you better.

Customer Name

Title

Company

Street Address

City

Country

Telephone

Fax

Total number of Electronic News binders needed:

State/Province

Postal/Zip Code

Dataquest

1290 Ridder Park Drive, San Jose, CA 95131-2398

..,•-,W4 .-^r

- • fit -^' * ^ ^

j j ^ ; . . - ) . , " •

_-\^.

jurjmoC tsno«nf«

ittif^'j S.-TT -• _

, ' ; .- !•;*».. jy'>*-"

^.T'tj-CJ*

^^ :" '^

P,.'j9itiitii-

: . r ^ i f ' 3

«i.pv*Tfto-3 -if'S

* ^ ' 1 — i i l . r f

• ' ^ ^ ^ '

^ . •

^ •

. ^ - e ^^fS'' '•'•

• '• . • i r s > s & •''•

»-,fia»"V-^'

..,•1'^.

a

.•v*^^^ -r'-^M-

• • = ? - • -

rts,'-

Global Market Insight for Injonnation

Technology Companies

1995 RESEARCH PROGRAMS

From semiconductors to systems, software to services, telecommunications to document management, Dataquest's scope of expertise provides clients with a clear view of the relationships between information technology segments - relationships that can have a profound impact on making strategic business decisions.

Computer Systems and Peripherals

Computer Systems

Qient/Server Camputing Worldwide

Advanced Desktop, Workstation, and

Qient/Server Distribution Channels

Computer and Qient/Server Systems Europe

PC and Low-End Computer Servers Europe

UNIX and Open Systems Europe

Computer Systems ]cepan

Workstations

Advanced Desktop and Workstation Computing

WorldvMe

Advanced Desktop and Workstation Quarterly

Statistics Worldwide

Workstations Europe

Workstation Quarterly Statistics Europe

Personal Computing

Personal Computers Worldimde

European PC Strategic Service

European (PC) Market Update

Personal Computers Asia/Pacific

Online, Multimedia,

and Software

Emerging Technologies

Multimedia Warldaride

Online Strategies Worldwide

Business Productivity

Qient/Server Software Worldwide

Personal Computing Software Worldwide

European Personal Computing Software

European Personal Operating Systems

Services

Customer Services

Customer ServiceTrends North America

European Customer Services

Professional Services

Professional ServiceTrends North America

Systems Integration and Applications

Development

• Consulting and Education

• Systems Management

Professional ServiceTrends: Vertical Market

Opportunities North America

DataQuest

PCMQA Systems and Peripherals Worldwide

Mobile Computing Worldwide

PC Quarterly Statistics United States

European (PQ Quarterly Statistics

PC Quarterly Statistics ]apan

PC Distribution Channels Worldwide

PC Distribution Channels Europe

Network Distribution Channels Europe-Upgrade

PC Distribution Channels Quarterly Statistics

PC Tracking Services Asia/Pacific

Computer Storage

Removable Storage Worldzoide

Optical Disk Drives Worldwide

Optical Disk Drives Europe

Rigid Disk Drives Worldxxnde

Rigid Disk Drives Europe

Tape Drives Worldwide

Tape Drives Europe

Graphics

Graphics and Disrlavs Worldwide

Workgroup Cotnpniting WorUiwide

European Workgroup Comjmting

Technical Applications

AEC and GB Applications Worldwide

Electronic Design Automation Worldwide

Mechanical CAD/CAM/CAE Worldwide

CAD/CAM/CAE/GIS Europe

CAD/CAM/CAE/GIS Asia/Pacific

European Professional Services

Sector Programs

System Services North America

European Desktop Support Services

Network Integration and Support Services North

America

European Network Integration & Support Services

Software Services North America

Strategic Service Partnering North America

1 9 9 5 R E S E A R C H P R O G R A M S

^;.tif

• i f **^*

Document

Management

Semiconductors

Emerging Technologies

Digital Documents North America

Colour Products Europe

Copiers

Copiers North America

Copiers Europe

Facsimile

Facsimile North America

Printers

Printers North America

Printers Europe

Printer Distribution Chaimels Europe

Printer Tiacking Services Asia/Fadfk

Regional IMarkets

Semiconductors Worldivide

Semiconductor Industry Snapshot

Semiconductors Europe

Semiconductors A^afPadjic

Semiconductors ]«pwn.

Microcomponent Product Planning Europe

Discretes Product Planning Europe

Application lUlarkets

Semiconductor Application Markets Worldwide

Strategic Tracking Services

• China/Hong Kong

Semiconductor Application Markets Europe

Semiconductor Application Markets Asia/Pacific

Communications Semiconductors & Applications

PC Semiconductors and Applications

• Taivjran

• Korea

PC Teardown Anal3rsis

PC Watch Europe

- • Singapore , ,_ Electronic Equipment Production Monitor Europe

Prodycts | : ^ *^

ASiCs'WorZdaJjde - ' =

ASICs Product Planning Europe

Memories WorlamAe

Memory Product Planning Europe

Miciocomponents Worldwide

IVIanufacturing

Semiconductor Equipment, Manufacturing, and

Materials Worldwide

LCD Multi-dient Study

Pricing Trends

Semiconductor Procurement Worldwide

Telecommunications

Networlfing

Netwrorking North America

• Local Area Networks North Arnirica

-a • Digital WANs Nortft Americe ' r, ,jo:

yiodsms North America

Networking Europe . -, .^

•'- • Asynchronous Transfer Mode (ATM) Europe

CaU Centres Europe

Telephones Europe

• PBX/KTS Systems Europe

Public

Public Network Equipment & Services

Wmldimde

Public Network Equipment & Services

iisiinujsu; s

/'":!9(x9 e'jesi.r ^

mat3;.,

r j "

'•:ti:;:

m

71 f

' J) y';» Modems Europe

'«r • Local Area Networks Europe j9- -; ?

) • Private Digital WAN Europe .J, ,,

'jv I a Networking Market Watch—Intellig^i ^Vi>^ ^

"'"^Network Distribution Channels Europe

fc:-!-;v©ice".i noqa!.-'' ' v

-; r fAiice Comorajniciitians No^h Antep'oi

^I*T "Vgice Processing JVortfiAm^^^iat

ii ^vv? Integrated Calls Centers ^pf'.thAmfnca

ip.-t •;. -1 p*e^[j^^witchingSystemp,'^OTth Amer^^

'.. . Vqice Coijt^irtunlcatioRS Europe ' '

.» VoiLe Processing Europe

Dataquest Catalyst

Programs

Cross Technology insight for:

—sr-nrcartciai

• GovemnfentftAgencies lofaubn.- •'••m^a

• IX Busiiiess Strategy and^^i^^^i^gJfi^g^^qzatKg*^

• Public "Nfetwbik'fiiuipment Nort/i America

• Public Network Services North America

Public Network Equipment & Services Europe

• Public Network Equipment Europe

• Public Network Services Europe

Personal

Personal Comnumications North America

Personal Commimications Europe

• Infrastructure and Services Europe

• Terminals Europe

Wireless Communications Europe

• Publishing, Media, and Consulting Finns

—A..K.aTirj.p^)]y)]^giT.£nrgani-7-aHr.Tic

• IXSupxxMrting Organizations

Emerging Markets sor

Centratand E&sferri Buropeaft Mdlftet^ r r^^

• Printers , - ' '• :P'-~: -ii^-}ii. -v [,

' •' Fere6r^'(iMnputers'-"^ ;j sr: • oj ;,i;: -v^'-

iHI i

P".O'J'.r-rz y, e r i l ; ,• m i T W

9 l J •-^-i

DataQuest

•BHB acoinpanyof

JUnThcDun&BiaastieetCorpofatiqn^

Information Technology lUlarlcet Opportunities

• China • Latin America

• Asia/Pacific

Bo^lonAie^ LomtoivAns ,. ^,,- Tokyc

550 Cocfiituate Road Holmere Farm Way"*"" ' Shinkawa Sanko BuikilnB

Ria(ningham,'MA01701-9324>t.j;,HighWyBoml)e,auckinghamshire ethHoor

•UniilaJStates ' ' ' HP-fS^XH • - ' - - ' - • • • ' 1-3-17,Shinkawa

'Ph6ne: (1)508-37»5Sgjj „ - United,Kingdpm ... 3 , Chuo*u,Tokyo 104

.,F^.., (1)508-370-6262 ^ Ptione:'(«) 1494-422^722 Japan

^, Fax (f») 1494yt22742 ,. Phone: (81) 3-556&0411

Taiwan

3IF, No. 87 Sung Chiang Road

Taipei

Taiwan, R.O.C.

Phone: (886)2-509-5390

Fax (886)2-509-5660

Corporate Headquarters

1290 Ridder Park Drive

San Jose. CA 95131-2398

United States

Phone: (1)408-»37-8000

Fax: (1)408-437-0292

©1994 Dataquest Incorporated

PpJs, , Munlcli V ^ tour Franklin. Cedex 11 ,. , , . Krppstadter Strasse,^;-

94942 La Wfense, Paris ' " "816?TMunich80 ^

Phone: (33) 1-41.25-1 &<» Phone: (49) 8 9 « 3 0 « » 0

Fax: (33)1-41-25-18-18 Fax (49)89-930-3277

Asla/PacHIc l/B China Hong Kong Tower

8-12 Hennessy Road

Phono: (852)527-7881

Fax (852)5200079

Korea

Trade Tower. Suite 3806

159Samsung^tong

Kangnam-gu, Seoul 135-729

Korea

Phone: (82)2-551-1331

Fax (82)2-551-1330

Dataquest is a registered trademark of the A.C. Nielsen Company

MKTG 11/94 (PSBCORP-01)

QUOlf

UL

'•i^y 'f

P.

£ j | , v w '

isv/^'iy

i?o-ij;:i ,:..:-:!u'i :•' h i i rrMO-^rn--x,":-

.iVi

5^JEW iniKl ficsir::-;ii iqqA

•;-iC1>U C'TrtJl-jM : ->!t6l>:ir Ci^ i ~ ^ n f c / - - . ^i7i

••••?

. jl-.\B!r;., ;;ii,:.:j£;{ 1 ; , ; j ^ - . . ; , . - i ^:ol:i;. •; ~o:>,

1 ;wiiqqA vi '-rolia'tnc ;jj?i ^^no;; r f-

' • d

')

i" i ; >-:,-•• £ • > -r. j - j b . x .

•-.."/'.6 m r;^j.ob . !

- ,aii:Ii ?r:-B

* ; l O J m j M / K - t j D i l ' - J l s '.;T-irijri.ijV,d -.:•".',. 1

-1

annujosh."

-m

.t .t;;rii. <; ':M Jni'sr^-ijf'JoH -:>i:i_;. iir.->_;,j-. ,!

' • " •% eii,! f : 3 H

-^ ab/S^il srrc-h'-

9\ii -XJ^Kf-i .nc. •

SEMICONQUlTOR EQUIPMENT,

MANUFACTURING & MATERIALS

'Dataquest's SEMM Worldwide program is a qor^prehensive market research service emphasizing an integrated perspective of the interdependence of wafer fabrication equipmentte£hnc4pg}/, semiconducipr^materials applications, andICprocess technology as they relate to the broader issues of semiconductor manufacturing.

Partnering to

'5t!Q-(;:i»'0. ,9^ it -r'-V o:\rfu I

'''.i'r' ' -i-^i

;;-^rt:;^. m-nj 'i ••jrcJiiyi\^- ;< ^ o D Isao.i^'

^•:i^~'.::^:iOiiK)f.n!}i m o ' ? e/ioaTa*-:

""••'kV; .'l&nJ.T

« ! T i i i g r a J ' - iiL. .-ra-i: , .6^1,':^^/-^^^: j ^ - , .

'""^ ifark^ettver^e

As a client,'^6u have direct Through an interactive

A, j.c -„^ 8roMJ<4|^olutions who can provide insights and work with you to tailor the

'iin' t't-iv ,

fi 91/" - i W ,

advice on niarket dynamics, program to meet the unique c 9 3 ' ' '.a '• -2 industry events, and -^ ;-^""''''-needs of your organization.

Information Resource

Centers

Clients have unlimited access to Dataquest's extensive print comtietitiV^'issues.

• > • • . • . • L - . :

Inquiry Support

-,' *n'! Electronic Delivery

Dataquestj3>fiers a variety of and online resource libraries worldwide.

Optional Custom Research

PersonalizeSli\'^iirJ'-supp£irt-'^*0Ois, knowti collectively as is a primary cOTfifbn^nt'blFS^'^^^flta^MJfa'f* On The Desktop, that

Should your needs exceed the scope of this program, your Data^S^st ^lial^'^^'-'^^haiVfe^ftie pdwer to deliver subsc^ptitJh=^ci^{f 13atlt|C[^ Insights directly to

Dataquest offers extensive primary research and

^^5iii Whemer you are on the consulting services.

~r^e^-oi*^-the office.

nrii "i-'oieril ve-'^i^'^r'^sT ^-20-

J2^

• • - ' t i l J t , . fi H'

•) ' L'JC

rrrrt

• Implantation

iftaniifaaiffh^SSrifefe''^ <

•'^-li^^mtolr&ids

• Wafer inspection

..norioqqC Js;nt.'-.' YgciO'-if-'jaT rnoijesri;-sft; Dataque;

•a«2 afispsemss t e ^ % g ^ ^ ^ ,^^^^

. «> Chemical mechanical

TATifVt c f m f i a m r inciffT^f i n t n

" ^" ^ ^ " - c > ^ " ^'^'^ issues relating to the global. r.^.-jyaffier-JFaorication polishing nature of the semiconductor Equipment

Semiconductor Materials

: ),n

^(Wt.

' fsas

.ia^,

* = « i-pi-!" .^: fc if

[ i t li-- -•'>

.&!- ' i J H t •;•:

•^Pi

Vifr;

. I W J U O a ^ " ^ ^ - ij;, .L no • "'a

;>-'^;K-1^ • ,, •

Oi ,';*0.' ..J "C

•JT'" f 'WSd-. r , -•:•_•!. r-a;.v?; -

Topics i n s l ^ - ^ ^ ,3 ^^^I^JSIfasts are pr|®tledrf<j|

s&me of the areas covered in

pur materials research:

. ^rs0i0mp'°<mm^''^''^^^°'^'^'''''''^

Lithbgraphy

• ^Ca^M'pending

• Etch and clean

• Capacity tif^nds

: x»t<-.l,-P#®sition

• FS^fOry ;j>i^^ctivity

"''•^""^^i^ttffcision

•~s8

'.if' : 'i.-iS

• Gases

• Photoresist

• ;,!pther materials

A'fiHH--

3; 3 - ;J-

• Silicon and epitaxial wafers

to iij"-- -^ 1 .^q•^

E)ataQuest

Worldwide Semiconductor Group

W H A T Y O U W I L L R E C E I V E A S A C L I E N T

SEMrCONDUCTOR EQUIPMENT, MANUFACTURING & MATERIALS

Ntertiet Analytis

Focus Reports

Semiconductor Equipment, Manufacturing, «td Materiite Fore«st: This report,

updated twice a year, contains forecasts for semiconductor consvunption and production by region, semiconductor capital spending by region, silicon wafer consumption by re^on, and wafer fabrication equipment consun:q>tion by region and by major equipment category.

L/lMlat0d /n the Second and Fettrtfi Ouartors

Silicon Maritet Share: Market share by company and by region for merchant suppliers, as well as trends in the merchant silicon and epitaxial wafer market, captive silicon production, wafer prices, and silicon square-inch consumption.

Published in the Second Quarter

Wafer Fabrication Equipn»nt Market Share: Market share by company by region for 150 companies that participate in 40 different segments of front-end wafer fab equipment.

Available in the Second Quarter

Wafer Fabrication Facilitiea—North America, Japan, Europe, Asia/Pacific, and ROW;

Pilot and production fab lines by company, location, fab name, products produced, process technology, minimum linewidth, wafer diameter, estimated weifer capacity, square feet of cleanroom, and cleanroom class. (Four separate publiaitions)

Available in the Fourth Quarter

Semiconductor Consumption and Shipment Forecast: Five-year revenue forecasts for the global semiconductor market by region.

Published in the Second and Fourth Quarters

Two in>depth reports on "hot" topics: These reports will focus on topics that provide you with essential information to support your tactical and strategic planning needs. Each report delivers a forward-looking assessment of market issues.

Tof^cs Planned for 1995:

Dedicated Foundry

Planned Fabs

Published in the First Quarter

Published in the Fourth Quarter

Perspectives

Alerts

Ti% DQ IMonday

Report

Timely analysis and commentary on key industry events and issues published in an easyto-read newsletter format on an event-driven basis throughout the year.

Eight issues per Y&w

Dataquest Predicts: In these hard-hitting reports, Dataquest takes a bold, opinionated, often controversial look at key issues, products, emd trends shaping the semiconductor equipment and materials metrkets. The reports rnake predictions about why, when, and how events will happen.

Pul>lished Twice a Ymir

Event-driven news and analysis, delivered by fax, giving a concise overview of significant announcements in the semiconductor industry.

PtOfUshml on an Everd-Oriven Bas^

Weekly news and commentary on semiconductor industry events and issues with a monthly snapshot of semiconductor pricing for 25 key semiconductors in 6 regions. Clients may wish to subscribe to this valuable newsletter for a broad view of this dynamic industry.

Weeidy via Intamet

Conferences Dataquest hosts the industry's preeminent semiconductor conferences in the United States,

Europe, Japan, Taiwan, and Korea. Chents receive special discoimts on conference fees.

riataoniest

aconwgunfof

11ieDun£Braibt(ect Corporation

©1994 Dataquest Incorporated

Corpotaie Headquaitera

1290 Ridder P a * Dme

San Jose, CA 95131 -2398

United States

Ptwne: (1)408-«37-8000

Fax (1)4a&437'0292

Boston V M

550 CochHuate Road

Fiamingham, MA 01701-9324

Urrited States

Phone: (1)50M7O-5555

Fax (1)50B-37O«262 tondenArea

Holmeis Faim Way

High Wycomtie. Buckinghamshire

HP124XH

United Kingdam

Phone: (44) 1494-422722

Fax (44) 1494-«2742

Munich

Kronsjadter Strasse 9

81677 Munich ao

GefTnany

Phone: (49)89-930-9090

Fax (49)89-930^277

Pnts

Tour Franklin, Cedei 11

94042 La O^ense, Paris

France

Phone: (33) 1-41-25-18^)0

Fax (33) 1-41-^18-18

Tokyo

Shinkawa Seiko Bultdng

6th Floor

1-3-17, Shinkawa

ChiK>ku,Tok)OlO*

Japan

Phone: » 1 ) 3 « 6 & 0 4 i i

Fax (81)3-5566-0425

Tahmi

31=, No. 87 Sung Chiang Road

Taipei

Taiwan, R.O.C.

Phone: (86^2-509-5390

Fax (886)2-Sa9-KE0

Korea

Trade Tower, Suite 380t isgSamsung-dong

Kangnam-gu, Seoul 135-729

Korea

Phona (8^2-551-1331

Fax (82)2-551-1330

Oatacpiest is a registered trademark of the-A.C. Nielsen Company

MKTG 9/94 (AKU 19B)

»

Dataquest

Dataquest^ j ^ ^ j ^ ^

1290 Bidder Park Drive * Sam Jose * CA * 9S131 -2394 * Phone 4 0 ^ ^ 7 - ^ 0 0 0 * Fact 'lOA^37-0292

The E3 Show: Video Game Showdown in Los Angeles

The Electronic Entertaiiunent Expo (E3) quite simply was a showcase for h o m e video game entertainment. The show, held May 11 through 13 in Los Angeles, is advertised as encompassing everything that is electronic, entertaining, and interactive, but the focus this year was on next-generation home video game hardware and software. The Sega Saturn and Sony

Playstation were the darlings of the show, witii innumerable software vendors using these 32bit systems to showcase their latest efforts. This Dataquest Alert is a short description of the n e w h a r d w a r e platforms from Sega, Sony, Nintendo, Atari, a n d 3DO.

"Who Has the Best Game in Town?"

This was the question on everyone's lips. Unfortunately, this question is more complex than it appears to bfe. Having the best game in town requires a combination of advanced features built into the hardware and popular software titles. Sony, Sega, 3DO, Atari, and Nintendo were all showing advanced hardware, although Nintendo's Ultra-64 w a s conspicuously absent. The platforms, see pataqUe^t!Eujife3to|ftg-'^iukBhla"MW^arsi«p Vi^^e>Gifi^:i^i#ir^«)i!^/,''/^m|dr'.,-i,-s • ' ^ ^ i .

Technology at the Ned r^diitk^A^fd^'^f^

materials for Ihfi £k?gaS^^m>4ild^i:^f'3f'M^ ^ focusing on the sepicondiigtct?tedKtloIpg^landMdiiiet o|)porttmltiia;ii, t' "^"'^'r^v-'i,f!*. "^^^4 -i^- *m ". ut

• ' ' • • • " ' • - ' • • ; • ' • • - ' v y ' ' - ^ ' : ; ^ . . • • • V i r r ^ i i ^ : , - , , 1 ' " ' • • •

.:SU

Sega Saturn

Sega w o w e d the crowd hy aririounciiig t h a t i h e Satuirrwi¥ hov*- available SrfT&^ted quantities" at three national retailers: Tbys-R U^, Electronics ffoutiq^e; a n d ' N e c ^ V r ( ^ a

• • j i •

« / > l H V i ^ t f T t-if^-m-r%iyv^\T f ^ r Qjr*ft-c*TsTifcA.flJ'J'rN' '»^'i-i^' H'>Wt^.4fVj^jakj^ ^' I'In.'^'tn. et'f»iJr***^>wJS-'j'.,--k*V-fc'**-iJ>t'*^ ^ 4 N U . T & ' Q J — H Jwl^ii^*^. A

Saturn in japan sinci; fh^Onxisirfias 1994 iseasdh. ^stirnates of tolal saleis4\oVsr"^«iiofutS90tPPO units. The Saturn h a CDfeROM-based gaine uriffS-Vith a 32-bit ai-chftectoia: Tl^aiimouncemeat i n c l u d e d a itarg^.jf)rieeof $ 3 9 9 ' f h a t i A s l \ i ! d ^ ^ ^ b i ^ '^ beUeves Ihis price to be agg^£dS7tf'e> baSi^a o^a%^*Wn fe^^ •' ' \ includirig a c o s l ^ faiE-ol^:daaWiafei n ^ ^ g i f Sati.trii f^^^

<e»ft

• ' • ' • • ' • ' • • '• •• -^ ' • - u , •• ' ' ' :• '. i • , : '

June 2, 1995 • '. ®1995 Dataquest Incorporated

«

Dataquest Alert Semiconductor Application Markets Worldwide of RAM, advanced sotind processing including MIDI capabilities, a 320-KB/sec CD-ROM drive, a n d expansion options.

Sony Playstation

Just after the announcement b y Sega, Sony Computer Entertainment announced that its

Playstation platform w o u l d sell for $299 w h e n it arrives in stores on September 9,1995.

Dataquest believes this system, like the Sega Saturn, is priced aggressively, based on a teardown anetlysis of tiiese units. The Playstation is CD-ROM-based a n d h a s a highly integrated design. In contrast to the Sega Saturn, the Playstation h a s the control electronics for the CD-ROM drive on the main board, which makes the unit very compact. The small size of the case is misleading, however, because the Playstation has lots of processing power with a

MIPS RSOOO core for the main processor, in addition to a graphics processor, a geometry engine, a sotmd processor, a n d a data decompression engine for JPEG decoding. Oflier specifications include 2.5MB of DRAM, 1MB of VRAM, and an array of video output options.

Table 1 compares the Sega Saturn a n d the Sony Playstation.

N i n t e n d o Virtual B o y

Nintendo's Virtual Boy game unit breaks the mold of the console-attached-to-your-TV standard. It is difficult to classify this product as either a handheld device like the Game Boy or a console device like the Super NES because it delivers an extremely different experience.

Virtual Boy provides a visually immersive, three-dimensional gaming experience using a selfcontained display. The unit looks like a pair of large goggles wired to a h a n d h e l d controller. In the exhibit haU, Nintendo showed the Virtual Boy m o u n t e d on a small stand. To understand h o w the Virtual Boy was displayed, imagine a pair of large goggles m o u n t e d on a stand so that a user can lean forward a n d look into tt\e goggles. Virtual Boy is portable, b u t is designed to be u s e d in a stationary position. Walking arotmd with the goggles on w^ould be dangerous. The imit is deceptively lightweight. Although the battery pack is attached to the hemd controller, aU of the main electronic components are contained in the goggles. A 32-bit RISC processor a n d a dual LED display create a smooth, three-dimensional display. The display, designed by

Reflection Technology in Walttiam, Massachusetts, has an LED strip a n d a mirror assem^bly for each eye. Unfortunately, the colors on the Virtual Boy are limited to a few^ shades of red, but the images are crisp a n d offer surprising deptti because each eye receives a different image.

Previous concerns over image flicker associated with this t5rpe of display appear to be tmwarranted, based on our "test drive" of ti\e tmit. The Nintendo Virtual Boy delivers a flickerfree image with sharp definition. The 3-D nature of the titles displayed involved vertical graphic planes at different apparent distances from the viewer in addition to true tiireedimensional graphics where objects move fluidly toward or away from the viewer.

A t a r i J a g u a r V R

The Atari. Jaguar VR is the Jaguar game unit with a head-motmted display a n d an additional unit for motion tracking. This Jaguar VR platfomn is visually immersive a n d offers a n e w level of virtual reality in homie gaming systems. The hardware uses infrared tracking to measure real-time movement of botti the head-mounted display a n d an optional joystick controller.

,,v\«^.

. « • » ' •

t

June 2, 1995 ®1995 Dataquest Incorporated

Dataquest Alert

Semiconductor Application Markets Worldwide

W h e n a user plays a game with the Jaguar VR system, the display changes as the h e a d turns.

For example, a user playing a game with a first-person perspective (seeing w h a t the character w^ould see rather than seeing the character on the screen) can look to the right simply by turning his or her head to the right. Head motion is tracked along two degrees of freedom.

(Left and right rotation is one degree of freedom; u p and down rotation is the second degree of freedom.) The optional joystick is also tracked in two degrees of freedom, but the specific degrees of freedom can vary from one software title to the next. The head-moxmted display uses a single, 104,000-pixel LCD display a n d then splits the image for each eye, so the effect is immersive, but not truly three-dimensional. For safety reasons, the head-mounted display will shut itself off if the player tries to walk around. The game m u s t be played while sitting or standing in one spot.

T a b l e 1

Hardware C o m p a r i s o n of t h e S e g a Saturn a n d S o n y P l a y s t a t i o n

Feature

Nvunber of Processors

M a i n CPU

CPU Q o c k Frequency

(MHz)

Memory

Sega Sahum

Eight

Hitachi SH-2 (two)

28.6

Sony Playstation

Five

MIPS R3000 core from LSI Logic

33

Graphics

Sotmd

2MB main

1.5MB video

540KB audio

540KB CD-ROM cache

2MB main

1MB video

512KB audio

32KB nonvolatile

Two processors

200,000 texture-mapped polygons/sec 180,000 texture-mapped polygons/sec

500,000 flat, shaded polygons/sec

T w o processors

360,000 flat, shaded polygons/sec

16 million colors

720x576 nvax resolution

Yamaha F H l DSP

Motorola 68EC000 processor

32 PCM channels

8 FM channels

44.1-KHz sampling

Proprietary

16 million colors

640x480 m a x resolution

Sotind processor

Built-in digital effects such as reverb a n d envelope

24 PCM channels

44.1-KHz sampling

JPEG Decompression

Standards

CD-ROM Drive

Video O u t p u t

2x speed

Composite video standard

Optional NTSC, S-video, RGB, a n d

HDTV

2x s p e e d

Composite video

S-video

RGB output

5V p o w e r for external RF converter

Source: Dataquest (Tune 1995) t

June 2, 1995

®1995 Dataquest Incorporated

# f

Dataquest Alert Semiconductor Application Markets Worldwide

3DOM2

3DO u s e d the show to announce m a n y of the specifications of its next hardw^are offering, which is called M2. 3DO was the first company to produce a 32-bit, CD-ROM-based game platform, a n d the M2 is its next step. Although 3DO w^as not showing M2 hardw^are o n the show floor, it d i d share irvany technical details. The core of the M2 technology is a 66-MHz

PowerPC 602 CPU with 10 custom coprocessors. Other specifications include a 64-bit memory b u s with 6MB (48Mb) of m e m o r y including synchronous DRAM a n d ROM, an MPEG engine supporting both MPEG-1 a n d JPEG decompression, alpha-channel graphics processing, a 66-

M H z DSP for audio processing (including MPEG audio decompression), a n d backward compatibility w i t h existing 3DO software titles.

Dataquest Perspective

Video game units are becoming fixed-function PCs in terms of both processing power a n d industry-standard features. The n u m b e r a n d complexity of semiconductor devices in each unit are overwhelming, a n d those devices represent ttie cutting edge of consumer technology. AH of these tinits also are targeted to the price-sensitive a n d highly competitive consumer electronics market, which makes the video game hardware market a low-margin business w h e r e h i g h volxunes are required to recover start-up costs. Based on a teardown analysis of t w o game units, Dataquest believes the hardware platforms are sold at very low margins a n d possibly sold at a loss to capture early market share in the emerging 32-bit game market. The profits appear to be tied to software sales.

The market dynamics and pricing policies for video game h a r d w a r e m a k e this a high-rew^ard a n d high-risk market for semiconductor n\anufacturers. Custom ICs are the rule rather than the exception inside these boxes, because each must have cutting-edge features in a highly integrated design. Standard chip products with these features are either n o t available or are not as integrated as they need to be to meet aggressive cost goals. Hetrdware designs for tivese game units are a mix of technologies from different vendors, but most of the technologies are integrated in ASICs rather than purchased as standard components. However, Sega^ Nintendo,

Atari, 3DO, a n d Sony are not defining the compression standards and sampling rates for multimedia technology; they are picking and choosing among the standards to deliver the greatest punch at the lowest price. For these reasons, semiconductor companies that w^ant their core processor, MPEG decoder, or audio synthesizer in these units m u s t w o r k with the game developers from the beginning. The high level of integration required to m a k e these game platforms profitable does n o t allow substitution of seiruconductor technologies in the middle of the design process.

The high product voltimes for successful home: video game platforms m a k e this market irresistible for m2iny semiconductor companies, in spite of the risks. A critical step for a semiconductor manufacturer is to work on reducing the risk once a decision to p u r s u e this market has been reached. Silicon Graphics is w^orking to reduce the market risk of its Magic

Carpet architecture by targeting multiple markets with the same core technology. For example, a proprietary version of the Magic Carpet architecture is the core technology for Nintendo's

Ultra-64 video g a m e platform, but Silicon Graphics is also releasing an "open systems" version of the architecture for the interactive set-top box market. In olher words, Silicon Graphics h a s

June 2, 1995 ®1995 Dataquest Incorporated

«

Dataquest Alert Semiconductor Application Markets Worldwide leveraged its design efforts to create opportunities in two different meirkets. Three alliance partners (AT&T Network Systems, Samsung Electronics, and Philips Electronics N.V) already have annotmced plans to use the Magic Carpet architecture in set-top box products. Silicon

Graphics has successfully reduced the meirket risk of this architecture by pursuing both of these growing opportvinities, and it plans to promote the Magic Carpet architecture for additional consumer products.

Summary

Overall, the home video game market provides a tremendous opportunity for semiconductor companies to enjoy high volumes and high profits. There is a high risk of choosing the "wrong" platform, however. For these reasons, it is important to leverage design efforts into more tiian one specific electronic system. The example given in this document is Silicon Graphics' foray into both home video games and set-top boxes with the Magic Carpet architecture, but opportunities also exist in tiie PC multimedia market. Success in any one of these markets will quickly lead to the million-unit volumes that every semiconductor company craves.

By Geoff Ballew

f

June 2, 1995 ®1995 Dataquest Incorporated

Perspective

Semiconductor A p p l i c a t i o n Markets W o r l d w i d e

Dataquest Predicts

Electronics Industry Works into a Groove for

Major League Growth

AbStrSCtl In the latest semiannual forecast of electronics equipment production, Dataquest

predicts worldwide revenue will grow by 11.8 percent in 1995 to reach $791 billion. A strong growth rate is projected to carry the industry to $1.2 trillion by 2000, an upward revision from Dataquest's spring forecast. This report also provides details on forecast production revenue and growth rates for equipment subcategories and for geographic regions. Details of this forecast are presented in Dataquest's Semiannual Worldwide and 'North American Electronic Equipment Production Forecast (SAMM-WW-MS-9504), which will be published in

October. Please see this publication for complete details of this forecast,

By Dale Ford

Dataquest Predicts

The electronics industry moved onto a significantly higher revenue growth curve beginning with the 1993 recovery. In the latest semiannual forecast of electronics equipment production, Dataquest predicts revenue from worldwide electronics equipment production will grow by H . 8 percent in 1995 to reach $791 biUion. This growth follows a 10.6 percent revenue increase in

1994, according to revised statistics.

Dataquest predicts electronics equipment revenue wiU grow at a 9.0 percent compound annual growth rate (CAGR) for the next five years, driving overall revenue to $1.2 trillion by the year 2000. Worldwide industry revenue will top the $1 trillion mark in 1998, two years earlier than previously forecast. This growth in electronics equipment production will drive a semiconductor market that Dataquest forecasts will climb to $331 billion by

2000. The new growth path of the electronics industry is best illustrated in

Dataquest

Program: Semiconductor Application iVlarkets Worldwide

Product Code:SAMM-WW-PD-9502

Publication Date: October 30,1995

Filing: Perspective

(For Cross-Industry, tile in the Semiconductors, Volume 3 of 3 binder behind the

Semiconductor Application Markets Worldwide name)

Semiconductor Application Markets Worldwide

Figure 1. From 1988 to 1993, electronics equipment industry revenue grew at a rate of 5.1 percent CAGR. A clear inflection point is seen between 1993 and 1994 as the industry moved to a much higher level of growth, now forecast to be 9.3 percent CAGR from 1993 to 2000.

If the electronics industry were compared to a professional baseball pitcher, it could be said the pitcher began the game throwing a combination of fastballs, change-ups, spitballs, sinkers, sliders, and even some curves. Now in the last two innings (years), he has moved into "a zone" and begun throwing blazing fastbaUs. Naturally his manager is pleased, but he is also concerned that the pitcher's arm will soon tire. Dataquest predicts this kid is a champ with staying power and the right stuff to keep throwing smoking fastb^ls for at least the next five innings (years).

Who's Hot, Who's Big

Dataquest's new forecast shows major u p w a r d revisions in the forecast revenue growth for computers, premise telecommunications equipment, mobile communications equipment, video consumer products, and appliances. The revenue forecast for each of the major equipment categories is presented in Figure 2. As expected, the PC market plays a donunant role in shaping the growth of industry revenue. In 1994, PCs accounted for

11.3 percent of total electronic equipment industry revenue with $80 bilHon in factory revenue. By 1999, the PC market is forecast to account for

15.2 percent of total industry revenue and generate $165 billion in sales. It should be noted that this revenue does not include standard peripherals

Figure 1

W o r l d w i d e Electronic E q u i p m e n t Production R e v e n u e Forecast

Billions of U.S. Dollars

1,200 -t

1,196

1,000-

8 0 0 -

6 0 0 -

1994-1999 CAGR = 9.0%

m

Transportation

m

Military/Civil

Aerospace

D

S

industrial

Consumer

1 ^

WB

Communications

Data Processing

^

4 0 0 -

200-

T I r ^ n I i " " T T r — i 1 — r

1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Source: Dataquest (October 1995)

SAMM-WW-PD-9502 ©1995 Dataquest Incorporated

9505759

October 30,1995

Semiconductor Application Markets Worldwide

Figure 2

Worldwide Growth of Electronic Equipment Production Revenue by Major Category

Data Processing

Computers

W//////////^^^^^

Datastorage

B I H B 36

Input/Output

ms^mm^^

59

Dedicated System

WSS^

29

Other Data Processing

1 7

Communications

Premise Telecom

1

T^-'- ; : : | l l i | p i i S i 1

76

Public Telecom

^i^mssm^

1 51

Mobile Communications

m^^ss^^^^M

74

Broadcast and Studio

9 8

Other Communications n | ^ 13

Consumer 1

Audio

• S P H 38

Video

msmms^^^i^

I 80

Personal Electronics

• H B 29

Appliances

^/zyj7//y///jyy.

M{^h.i'i'^v^

72

Other Consumer

H B 21

Industrial

^ 1994

• 1999

Security and Energy Management

P M 21

Manufacturer Systems/Instruments

w^mmz^^^^

.:.:;. i:-;J^a!yS

92

Medical Equipment

l ^ g 22

Other Industrial

gjiffl 27

lUlilitary/Civil Aerospace

^^smmssssssmM

64

Transportation

m^mgfk

47

1 1 I \ i

C

%

) 50 100 ^SM !S^' 2i£

BEIIions of U.S. Dollars

g; lOETEO

Source: Dataquest (October 1995)

SAMM-WW-PD-9502 ©1995 Dataquest Incorporated October 30,1995

Semiconductor Application Markets Worldwide such as rigid disk drives and CD-ROM drives, among others, but it does include memory SIMMs sold through aftermarket channels into the PC market. As seen in Figure 3, other data processing categories such as data storage are forecast to experience comparatively slower growth rates due to severe price erosion. However, the price decline in these categories is a key factor in enabling complete, cost-competitive PC packages to stimulate the growth of both business and consumer PC markets. Overall, the data processing category is forecast to grow at a 9.0 percent CAGR during the next five years, the highest growth rate next to the communications category.

The communications category is expected to experience the highest revenue growth during the next five years with a forecast revenue growth of

12.2 percent CAGR. Mobile communications and premise communications are projected to lead the way with increases of 16.8 percent and 12.7 percent

CAGR, respectively. Revenue from mobile communications equipment should more than double from $34.1 bilUon worldwide in 1994 to $74.3 bilhon in 1999.

The Asia/Pacific region continues on its path to replace North America as the leading producer of electronic equipment by 1999. Figure 4 shows that in 1995 North America is expected to generate 31 percent of the worldwide electronic equipment revenue, followed by Japan with 27 percent, Europe with 22 percent, and Asia/Pacific with 20 percent. However, by 1999 Asia/

Pacific is forecast to produce 27.4 percent of worldwide electronic equipment revenue with $297 biUion, while North America is forecast to account for 26.6 percent of the revenue with $289 billion. It is also notable that revenue growth in the Asia/Pacific region is not driven by PCs, data storage, and consumer products alone. Communications, particularly mobile communications, will add momentum to the regional growth in production as the local markets experience dramatic expansion.

Dataquest's semiannual forecast of worldwide and regional electronic equipment production draws on extensive databases that cover a range of areas from data processing to transportation. Information for these databases is developed by Dataquest analysts in North America, Europe, Japan, and the Asia/Pacific region. Through a coordinated effort, this information is compiled and analyzed to develop a comprehensive view of the worldwide electronics industry. Details of this forecast are presented in

Dataquest's Semiannual Worldwide and North American Electronic Equipment

Production Forecast (SAMM-WW-MS-9504), which will be published in

October. Please see this publication for complete details of this forecast.

SAMM-WW-PD-9502 ©1995 Dataquest Incorporated October 30,1995

C/3

>

CD

tn

o o

T 3

O

B>

!-••

CD

@

<o

CO

en o

o

I

<o

U i

o

t>o

Figure 3

Worldwide Electronic Equipment Growth Forecast by Major Category

Forecast CAGR 1994-1999 (Percent)

18

12-

6 -

BS

ODP

SEM

OCN^T^

PE

OCM

ME

AU

DSY

DST

MC

PUT

10

PBT

VI

AP

VA

CO

(Overall Average Forcast Growth Rate)

11/SI

ME

01

AU

VI

PE

AP

OCN

MCA

TR

CO

DST

10

DSY

ODP

PRT

PUT

MC

BS

OCM

SEM

MSI

Comput

Data Sto

Input^Ou

Dedicate

Ottier Da

Premise

Public T

Mobile C

Broadca

Other Co

Security

Manage

Manufac

Systems

Medical

Other In

Audio

Video

Persona

Applianc

Olher Co

Military/C

Transpo

0 0

_o

CO

CO

cn

O

O

f - ^

o cr

CD

Source: Dataquest (OctotDer 1995)

MCX

-/M

30,000 60,000

130,000

1994 Revenue (Millions of U.S. Dollars)

Semiconductor Application Markets Worldwide

Figure 4

Estimated 1995 Regional Production of Electronics Equipment

North America Europe

Japan

Data Processing

H Communications

S Consumer

i i i Industrial i m Military/Civil Aerospace

H Transportation

Asia/Pacific

27%

Total 1995 Revenue = $791 Billion

Source: Dataquest (October 1995)

SAMM-WW-PD-9502 ©1995 Dataquest Incorporated

9SQS762

October 30,1995

Semiconductor Application Markets Worldwide

For More Information...

Dale Ford, Senior Irtdustry Analyst (408) 468-8311

Internet address [email protected]

Via fax (408) 954-1780

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in coniidence by our clients. Reproduction or disclosure in whole or in

U 3 l 3 Q U c S l P " t ' ° other parties shall be made upon the written and express consent of Dataquest

©1995 Dataquest Incorporated—Reproduction Prohibited aoompanyoi ^ ^ '

TheOunSBradsticet Corporation Dataquest is a registered trademark of A.C. Nielsen Company

Perspective

Semiconductor Application Markets Worldwide

Dataquest Predicts

The Worldwide Electronics Industry: On Target for

$1 Trillion by 2000

AbstrSCt: Building on momentum that began in 1993, the worldwide electronics industry

surged ahead 9.0 percent to nearly reach $690 billion in 1994. Dataquest's most recent electronic equipment forecast predicts continued strong growth of 6.5 percent during 1995, which will push production to $734 billion. This same forecast projects the worldwide electronics industry to have a compound annual growth rate of 7.0 percent during the next five years to drive the industry to $967 billion in 1999. If this same growth is sustained an additional year, the worldwide production of electronic equipment will top $1 trillion in the year 2000. This report provides a summary of the global and regional forecasts for the data processing communications, consumer, industrial, military/civil aerospace, and transportation electronics industries. The information presented highlights important regional and market trends and identifies significant economic factors influencing these trends. Also, an updated strategic planning guide plots product opportunities in each of the major industry segments and explains the implications of future product developments for semiconductor products.

By Dale Ford

Electronics Production Experiences Solid, Broad-Based Growth

The dramatic 9.0 percent growth of the worldwide electroiiics industry during 1994 was driven by soHd production increases in all regions and all but one of the major industry segments. The electronics industry nearly reached $690 billion in 1994 and is expected to hit $734 billion in 1995.

Dataquest Predicts

Dataquest predicts that worldwide electronics industry will have a compound annual growth rate (CAGR) of 7.0 percent in the next five years, driving the industry to $967 biUion in 1999. Sustained growth for an additional year at the same rate will put worldwide production of electronic equipment over $1 trillion by the year 2000.

EtataQuest

Program: Semiconductor Application Marl<ets Worldwide

Product Cod8:SAI\/ll\/l-WW-PD-9501

Publication Date: IVIay 8,1995

Filing: Perspective

Semiconductor Application Markets Worldwide

The communications and data processing industries led all industries with amazing growth of 11.4 percent and 11.3 percent, respectively. Mobile communications continued to fuel growth in the commtmications segment, while the surprisingly strong growth of personal computers boosted the data processing industry to new heights. The only industry that failed to show major growth was the inilitary/civil aerospace industry, which contracted 2.6 percent in 1994 because of continued cuts in defense spending.

Figures 1 and 2 summarize Dataquest's worldwide electronics industry forecast. The Asia/Pacific region continued its robust growth as production grew 17.6 percent in 1994 to $135 billion. Rebounding from a severe recession, Japan's growth rate surpassed North America and Europe in 1994 as its production grew 9.0 percent to $190 billion. The soaring yen has also boosted the dollar-based growth of Japanese production in 1994. This article includes figures that show the continuing shift in global production patterns.

The combination of a dynamic electronics industry and shifting global and regional economic conditions creates tremendous challenges for projecting the development of the global electronics industry. Dataquest's recent forecast of worldwide and regional electronic equipment production draws on extensive databases covering data processing to transportation. Information for these databases is developed by Dataquest analysts in North America, Europe, Japan, and the Asia/Pacific region. Through a coordinated effort, this information is compiled and analyzed to develop a comprehensive view of the worldwide electronics industry. Detailed results of this project are presented in Dataquest's Worldwide and North American Electronic

Equipment Production Forecast, dated April 24,1995 (SAMM-WW-MS-9501).

Please see this publication for complete details of this forecast.

Figure 1

W o r l d w i d e Electronic E q u i p m e n t Production R e v e n u e Forecast

Billions of U.S. Dollars

1,200-ff

m

Transportation

1994-1999 CAGR = 7.0%

1,000-

^

Military/Civil

Aerospace

800-

n

Consumer

600-

s

Industrial

400-

n

Communications

200-

Data Processing

^

1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Source: Dataquest (May 1995)

SAMM-WW-PD-9501

©1995 Dataquest Incorporated

9SD1920

May 8,1995

Semiconductor Application Markets Worldwide

Figure 2

W o r l d ^ d e Growth of Electronic Equipment Production Revenue by Major Category

Data Processing

Computers

Data Storage

Input/Output

Dedicated System

Other Data Processing

Communications

Premise Telecom

Public Telecom

Mobile Communications

Broadcast and Studio

Other Communications

Consumer

Audio

Video

Personal Electronics

Appliances

Other Consumer

Industrial

Transportation

Military/Civil Aerospace

Billions of U.S. Dollars

1 r

140 160 lao

9601 Sei

Source: Dataquest (May 1995)

Global Economics: Continued Strength, Moderate Growth Expected

Current and expected worldwide macroeconomic conditions are assessed and forecast in developing Dataquest's electronic equipment production forecast. Information from The Dun & Bradstreet Corporation, including gross domestic product (GDP) forecasts listed in this section, is used to develop a macroeconomic forecast for the world's major economies. This forecast identifies trends in the economic health of the world's leading consumers and producers of electronic equipment and provides a foundation for projecting electronic equipment production on a regional basis.

Worldwide business optimism for first-quarter 1995 sales edged higher, while expectations for higher prices and profits moved toward six-year peaks, according to The Dun & Bradstreet Corporation's latest quarterly survey of more than 11,000 business executives in 16 countries. Although there has been recent volatility in some markets in response to proposed

U.S. trade sanctions against China and the recent depreciation of the

SAI\/1M-WW-PD-9501 ©1995 Dataquest Incorporated

• May 8,1995

Semiconductor Application Markets Worldwide

SAMM-WW-PD-9501

Mexican peso, the strength of the survey responses indicates that the current global economic expansion will be able to overcome setbacks without a broad negative impact. Specific trends and assumptions for the major regions in this forecast are outlined in the following sections.

North America

Table 1 shows actual gross domestic product (GDP) growth in 1994, and

1995 to 1996 projections.

Table 1

Nortti A m e r i c a n Gross D o m e s t i c Product G r o w t h Percentages

United States

Canada

Mexico

Source; The Dun & Bradstreet Corporation

1994

4.0

4.2

2.8

1995

2.9

3.7

2.0

1996

3.0

3.2

2.5

Assumptions for the North American region include the following:

• The first-quarter 1995 survey suggests that executives in North America and Europe believe their economies' respective expansions are nearing their peaks, and that inflationary pressures will grow stronger as idle production capacity is exhausted. As the second quarter approaches, prices will merit closer inspection in order to gauge the impact of recent interest rate hikes in the United States.

• Rising interest rates are also expected to moderate growth in the consumer electronics, home PC, and automobile markets during 1995.

Higher interest rates have already slowed the housing market, which will reduce sales of consumer electronics into new homes.

• Prior to the sharp decline in the value of the peso, executives in Mexico had reported expectations for record growth in sales, prices, profits, and inventories as businesses in the region prepared for unprecedented expansion under the North American Free Trade Agreement (NAFTA).

Unfortunately, the recession that has gripped Mexico in the wake of the collapsing peso changes everything. Although there is still tremendous promise in the region, Httle of the previously expected growth will be reahzed tmtil the peso stabilizes. As one of the United States' largest trading partners, and also as a critical export market for Canada, Mexico's volatility will doubtless h a v e some damping impact on expected growth across North America.

• Uncertainty surrounding government rules and regulations for the information superhighway continues to present roadblocks to corporate development efforts. Expansion of investment has been delayed by many companies because of poor results in many trials across country.

Japan

Assumptions for the Japanese region include the following:

• Japan's real GDP is forecast to grow 2.0 percent in 1995, u p from 0.8 percent growth in 1994, and to accelerate to 3.5 percent in 1996.

• Japan experienced steady economic growth during the last three quarters of 1994. However, Japanese executives predicted that business conditions would only improve marginally during the first quarter of 1995. Strong

©1995 Dataquest Incorporated

May 8,1995

Semiconductor Application Markets Worldwide

gains in expected sales and employment levels were offset by slightly lower expectations for net profits. The yen's appreciation against

Western currencies proved a hindrance to exports and has frustrated

Japan's nascent recovery, at least for the first quarter of 1995.

• Industrial electronics should benefit from a resumption of capital spending by Japan's manufacturing industries, in keeping with Japan's overall economic recovery. Industrial electronics growth is being particularly spurred by record investment spending in the semiconductor industry on manufacturing equipment, and on test and measurement equipment.

• Production for domestic consumption and export will continue to shift out of Japan. The further strengthening of the yen will reinforce this shift.

• Japan will continue leadership in significant high-growth electronics equipment areas such as notebook PCs, personal digital assistants, digital cellular telephones, video games, video cameras, and optical disk drives.

Europe

Table 2 shows actual GDP growth in 1994 and 1995 to 1996 projections.

Table 2

European Gross Domestic Product Growth Percentages

Germany

United Kingdom

France

Italy

Source: The Dun & Bradstreet Corporation

1994

2.8

4.0

2.4

2.3

1995

3.1

3.2

3.2

2.9

19%

3.5

3.0

2.9

3.2

Assumptions for the European region include the following:

• Europe is enjoying the benefits of a mature economic expansion, and most indexes are positive for the duration of the first quarter of 1995. The region has expanded continuously from the first quarter of 1993. European manufacturers made significant productivity gains in 1994, but continued expansion could increase pressure on central banks in the region to raise interest rates.

• Although growth in various segments of the European telecommunications market is mixed, European telecommrmications mantif acturing companies are benefiting from a significant presence in the global market emd rising exports to the fast-growing Asia/Pacific market. In the mobile arena, GSM cellular telephones will drive a huge market in Europe as penetration rates approach those in the United States. The continued success of GSM in non-European countries also will bolster production in the near term until it shifts closer to the regions of consumption.

• Strong rates of growth are anticipated throughout Europe as overall expectations for sales and profits continue to be optimistic.

• With a strengthened economy, household spending on consumer electronics will continue in 1995.

Asia/Pacific

Table 3 shows actual GDP growth in 1994 and 1995 to 1996 projections.

SAMM-WW-PD-9501

©1995 Dataquest Incorporated May 8,1995

Semiconductor Application Markets Worldwide

Table 3

A s i ^ P a c i f ic Gross D o m e s t i c Product G r o w t h Percentages

Taiwan

South Korea

Singapore

1994

Hong Kong

Source; The Dun & Bradstreet Corporation

6.3

8.2

10.0

5.5

1995

6.3

15

8.0

5.5

1996

6.5

7.5

8.0

5.0

Assumptions for the Asia/Pacific region include the following:

• The Dynamic Asian Economies (DAE) — Korea, Taiwan, Thailand, Hong

Kong, Singapore, and Malaysia—have attracted growing attention for their strong economic performance. The strong growth patterns of 1993 and 1994 are projected to continue into 1995 and 1996. Significant increases in local regional consumption and liberal trade policies that fuel exports persist as key drivers in these economies.

• Because of the continued importance of exports to the world's mature economies for the Asia/Pacific region, the economic performance of

Asia/Pacific countries is influenced strongly by economic activity in areas such as the United States and Europe. The impact of decelerating growth in the United States is expected to be partially offset by accelerating growth in western Europe in 1995. AlsOj. tihe currencies of major economies in this region are tied to the U.S. dollar, so their exports are becoming cheaper (in some cases) as the dollar depreciates against the yen and deutsche mark. Conditions in these export markets paint a bright picture for the Asia/Pacific region in 1995.

• Low-cost production opportunities will continue to generate growing captive and contract manufacturing operations in Asia/Pacific countries such as Indonesia, Malaysia, Thailand, and The People's Republic of

China.

• Investment in consumer electronics production, particularly from Japan, will follow an upward trend. Local demand for consumer products in many countries in the region is just beginning to grow.

• The booming telecommunications growth in China and the Asia/Pacific region will expand. Tht; industry has leapfrogged from low-end telephones and answering machines to include high-end equipment plants.

Major telecommunications players have been encouraged to invest in local manufacturing in exchange for market access.

Regional Production Trends: Shifting to Asia/Pacific Region

The influence of the economic factors described is seen in the shifting global production trends in the electronics industry. Although overall production in all regions will grow, the Asia/Pacific region is predicted to continue its dramatic ascendancy in electronic production as it serves growing markets in its own region and provides cost-effective manufacturing solutions for export markets. As a result, the Asia/ Pacific region WLU edge out North

America as the leader in revenue from electronic equipment production by

1999. Figure 3 shows the clear pattern of shifting production to the Asia/

Pacific region. The major underlying trends in data processing, communications, and consumer equipment that sum to this overall movement are shown in Figures 4, 5, and 6.

SAMM-WW-PD-9501 ©1995 Dataquest Incorporated May 8,1995

Semiconductor Application Markets Worldwide

Figures

Regional Share of Worldwide Electronic Equipment Production Revenue

Percent

40-

1 0 -

North America

Japan

— i —

1997

1994

1995

1996

Source: Dataquest (May 1995)

Figure 4

Regional Share of Data Processing Equipment Production

Percent

40-

1998

Europe

Asia/PacHic

1999

9501922

1 0 -

1994 1995

Source: Dataquest (May 1995)

SAIVIIVI-WW-PD-9501

1996

North America

Japan

— I —

1997

©1995 Dataquest Incorporated

Europe

Asia/Pacific

1998 1999

9501923

IVIay 8,1995

Semiconductor Application IVIarkets Worldwide

Figures

Regional Share of Communications Equipment Production

Percent

40-

3 0 -

40

3 5 -

3 0 -

2 5 -

2 0 -

''1

1 0 -

5 -

• -

2 0 -

• -

1 0 -

North America

Japan

— I —

1997 1994

1995

Source: Dataquest (May 1995)

1996

Figure 6

Regional Share of Consumer Electronics Equipment Production

Percent

1998

Europe

Asia/Pacific

1999

9501924

North America

Asia/Pacific

1994

1995

Source: Dataquest (May 1995)

Japan

- * — Europe

1996

1997

1998

1999

9501925

SAI\/IM-WW-PD-9501 ©1995 Dataquest Incorporated

May 8,1995

Semiconductor Application l\/larl<ets Worldwide

Market Opportunities in tlie 1990s

Mergers, alliances, trial markets, and multifunctional products have developed at a quickened pace as companies pursue market opportunities created by overlapping industries in computers, commimications, and consumer products. Indeed, there are few product segments that wiU not be impacted by the growing overlap of technologies and markets. The following figures and highlighted trends provide an overall perspective on the anticipated growth of major segments in the data processing, communications, and consumer industries. Key trends are also listed for the industrial, military/civil aerospace, and transportation industries.

Data Processing Equipment IVIarlcet Trends

The worldwide data processing equipment market is forecast to grow at a

7.2 percent CAGR from $214 billion in 1994 to $304 billion in 1999. Figure 7 shows the forecast growth and size of major segments within this market.

Key Trends

Key trends are as follows:

• Competitive forces among the Pentium, PowerPC, and the other RISC processor camps continue to drive the price, performance, and pace of product development in the PC and workstation markets. As a result, businesses and consumers are gaining access to soaring computing power at stable price points.

Figure 7

Data Processing Equipment Production Forecast by Major Category

1993 Worldwide Revenue (Billions of U.S. Dollars)

Computers

100-

80-

60-

40-

Input/Output •

Data Storage •

• Dedicated Systems

20»

• Others

0-

1 1 1 < 1 i 1 1

0 1 2 a 4 5 6 7

0 9

Forecast CAGR 1994-1999 (Percent)

9501926

Source: Dataquest (May 1995)

SAMM-WW-PD-9501 ©1995 Dataquest Incorporated May 8,1995

10

Semiconductor Application Markets Worldwide

• Mobile products such as notebook, subnotebook, and palmtop PCs are leading the growth of the PC market. Continued sluggishness and a conversion to less-expensive systems based on multiprocessing technology in the mainfram.e and midrange markets is offsetting strong growth by

PCs in the computer market. Also, the success of the PC is eroding the potential workstation market.

• The expansion of the home PC market and niultimedia PCs has provided added momentum to the PC market. Both of these factors lead to a significant increase in growth of the personal computer market during 1994.

These drivers will continue to play a fundamental role in market growth over the forecast period. The rapid expansion of interest in multimedia is also bolstering growth in the storage, display, and add-in board segments such as sound, graphics, and video boards.

• Increasingly powerful operating systems and applications programs are pushing the appetite for storage capacity. Together with multimedia products, they are creating demands for higher storage capacity in rigid and optical disk drives. However, major price erosion has kept revenue flat or declining in all major storage categories. This trend wiU depress future revenue growth.

Communications Equipment iVIaricet Trends

The worldwide communications equipment market is forecast to grow at an 8.1 percent CAGR from $127 billion in 1994 to $188 bilhon in 1999.

Figure 8 shows the forecast growth and size of major segments within this market.

Figures

Conununications Equipment Production Forecast b y Major Category

1993 Worldwide Revenue (Billions of U.S. Dollars)

Ar\

35>

Public Telecom • Premise Telecom • ^ . ' ^ ° ' ' ' ' ® H

Communications

30-

25-

20-

15-

10-

5 n -

• Others

• Broadcast and Studio

1 I 1 1 1

Forecast CAGR 1994-1999 (Percent)

BS0iaE7

Source: Dataquest (May 1995)

SAM!\/1-WW-PD-9501

©1995 Dataquest Incorporated IVIay 8,1995

Semiconductor Application Markets Worldwide 11

Key Trends

Key trends are as foUows:

• A continued surge in demand for wireless voice and data communications products including cellular phones, wireless modems, and repopularized pagers is propelling the mobile communications markets to high growth rates.

• The growth of mobile communications markets throughout the world exceeded even the most optimistic forecast for 1994. For example, the number of cellular subscribers in the United States grew by more than

40 percent in 1994. Dataquest is very optimistic about the continued growth of wireless markets in North America and around the world. In the United States, the creation of a new personal communications services (PCS) market that wiU increase competition with existing wireless communications services is expected to fuel even greater growth. The

GSM system in Europe is driving strong growth in that region, and the

Personal Handy Phone (PHP) and Pacific Digital Cellular system is showing great promise in Japan. However, PCS/personal communication network (PCN) is a very cost-conscious market and there will be significant price pressure on end-user equipment. This pressure will be reflected to the factory, where significant average selling price (ASP) erosion is expected over the forecast period. Even with these price pressures, strong growth is forecast for mobile equipment revenue.

• The strong subscriber growth in the mobile communications market will act as a driver for new infrastructure equipment and mobile base stations. In addition to serving the needs of their local markets, the

United States and Europe are major exporters of mobile communications equipment to other developing regions of the world. For this reason, the growth of base station production will outpace even the growth of the local mark'sts in these areas.

• A trend toward digital cordless phones, smart phones, and answering machines is developing.

• Digital radio and TV studio upgrades are stimulating higher growth in broadcast and studio equipment.

• Increased capital spending is stimulating premise networking equipment production. Near-term dememd is for higher-speed network^g products such as switched Ethernet and FDDI. Intermediate and long-term production will be carried on for 100-Mb/sec and ATM technologies.

• The modem market is being propelled by record subscriptions to online services and Internet. ISDN and T Carrier equipment production is increasing because of higher-speed WAN requirements.

• Public telecom and cable infrastructure equipment companies (for example, SONET) are receiving record orders to support the rollout of highbandwidth interactive video services to the residential market.

• Fiber-optic-based synchronous digital hierarchy (SDH)/SONET transmission systems are being installed worldwide. Asynchronous transfer mode (ATM) and asymmetric digital subscriber loop (ADSL) are representative of new technologies under evaluation.

• Networking and internetworking equipment continues strong growth as network systems are deemed an integral part of computer hardware additions and upgrades driving purchases of hubs, routers, and bridges.

SAIVIM-WW-PD-gSOl ©1995 Dataquest Incorporated May 8,1995

12

Semiconductor Application Markets Worldwide

• Call processing equipment continues in popularity in North America and expansion in other regions.

Consumer Electronic Equipment Marlcet Trends

The worldwide consumer electronic equipment market is forecast to grow at a 7.3 percent CAGR from $155 biUion in 1994 to $220 billion in 1999.

Figure 9 shows the forecast growth and size of major segments within this market.

Key Trends

Key trends are as follows:

• Much of the consumer electronics market has entered "replacement" mode in Group of Seven (G7) countries. Growth in consumer electronics sales in newly industrialized countries is adding strength to this market.

• A seemingly insatiable appetite for entertainment is driving growing purchases of items such as VCRs and video game players. Digital video and audio products are replacing past-generation analog systems. DSPbased TVs, VCRs, camcorders, and audio products including Digital

Compact Cassette (DCC) and Mini Disc (MD) are emerging. The DCC and MD products have enjoyed early success in Japan, and their promoters beheve this growth will spread to other regions as more supporting

"software" is introduced. Also, a new genre of interactive consumer devices such as CD-ROM "infotainment" systems and digital videodisk

(DVD) is emerging.

Figure 9

Consumer Electronic Equipment Production Forecast by Major Category

1993 Worldwide Revenue (Billions of U.S. Dollars)

Appliances • V i d e o !

45-

4 0 -

3 5 -

3 0 -

2 5 -

2 0 -

15-

10-

5 -

Others •

Personal Electronics •

Audio!

U 1 1 1 1

0 2 4 6 8

IQ

Forecast CAGR 1994-1999 (Percent)

Source: Dataquest (May 1995)

1 2

9501926

SAMM-WW-PD-9501

©1995 Dataquest Incorporated May 8,1995

Semiconductor Application Markets Worldwide 13

• The competition over the set-top decoder box market is heating u p as companies vie for the opportimity to bring the information superhighway into the living room with their hardware. Sales of direct broadcast satellite systems soared in 1994. Sales of cable TV set-top boxes are expected to ramp u p also. Very strong growth is forecast for this network.

• Appliances continue solid growth with strengthening economies and new household forniation.

Industrial, Aerospace, and Transportation Electronics Equipment

IVIarlcet Trends

The worldwide industrial electronics equipment market is forecast to grow at a 7.1 percent CAGR from $105 billion in 1994 to $148 billion in 1999.

Highlights are as follows:

• Industrial electronics production has been fueled by broad-based growth of business spending on capital equipment. The continuing trend of downsizing has driven companies to seek implementation of more efficient processes. Recent increases in interest rates have lowered the current outlook for growth in industrial equipment production slightly below what it otherwise would have been.

• There is a growing trend toward the use of standard computing equipment in industrial control applications.

• Increased pressure from managed health care providers and insurance companies has h a d a major impact on the medical electronics markets.

Major equipment manufacturers are revising their product development strategies to provide more cost-effective solutions for an extremely costconscious market. In a period of major cost-cutting in the health care industry, medical equipment providers are working to position their products as a "solution" to the problem of spiraling costs. This environment will moderate revenue growth in this segment over the forecast period.

The worldwide transportation electronics equipment market is predicted to grow at an 8.6 percent CAGR from $30 billion in 1994 to $45 billion in 1999.

Highlights are as follows:

• Automotive production in the United States and Europe is expected to slow over the next two years as higher interest rates impact the market.

The Asia/Pacific market for automobiles is booming.

• Automotive electronics content is rising. Products enabling conapUance with more rigorous antipollution regulations and fuel-efficiency standards, antilock brakes, air bags, navigation systems, and security systems are leading the way. Comprehensive driver information systems will add to the growth in the long term.

The mihtary/civil aerospace electronics market is expected to grow at a modest 1.5 percent CAGR from $58 billion in 1994 to $62 billion in 1999 for the following reason:

• Military spending cuts are expected to level out after several years of deep cuts. A cyclical upturn is expected in civil aviation avionics production. Civilian space electronic production is expected to remain robust amid a flurry of orders for low-earth-orbit satellite telephone and direct broadcast satellites.

SAMI\/I-WW-PD-9501 ©1995 Dataquest Incorporated May 8,1995

14 Semiconductor Application Markets Worldwide

A Spectrum of Opportunities

Tables 4 through 9 show updated planning guides with the 1994 to 1999 outlook for selected worldwide electronic products. In general, the declining stage indicates that the product or function is being replaced by another. Slow growth of zero to 6 percent indicates that a mature product is on the way to decline or is in a replacement/saturated market mode.

Emerging growth of greater than 12 percent and moderate growth of 7 to

12 percent represent the market behavior of products going through introduction and maturity.

Projected semiconductor trends for computer systems are as follows:

• Memory

Q PCs with 7MB average DRAM in 1994 will move to 20MB DRAM in

1999. The preferred configuration will be xl6.

• Increased opportunities wiU emerge for fast/synchronous DRAMs as

MPU speeds outstrip conventional memory architectures. a Bursting, MPU-specific SRAM wiU be used for cache design. Average size will move from 256KB to 1MB by 1999.

• MPU/MCU

Q New generations of Pentium, PowerPC, Alpha, PA-RISC, SPARC, and

MIPS processors wiU drive future PCs, workstations, and servers.

• Proprietary processors such as ARM and Dragon will continue to be used in palmtop devices.

• ASIC/ASSP a New chipsets to support new MPU generations will continue to emerge. a The PCI bus standard will dominate with Card Bus emerging.

Q The PCMCIA interface will dominate in the mobile market.

Table 4

Computers: 1994 to 1999 Unit Growth

Declining (Negative)

386 PC

Proprietary midrange

Proprietary workstations

Proprietary mainframe

Proprietary supercomputer

Slow (0 to 6%) Moderate (7 to 12%)

486 PC

Emerging (More than 12%)

Multimedia PCs and workstations

Open workstation/server Power PC—PCs/workstations/ servers

RISC mainframe

Pentium (P6, P 7 ) - P C s / s e r v e r s

Alpha, PA-RISC, SPARC;

M I P S - PCs/workstations/ servers

RISC supercomputer Notebook/ subnotebook PCs

Palmtop computers (for example, PDAs)

Pen-based organizers

Board computers

Source: Dataquest (May 1995)

SAMM-WW-PD-9501 ©1995 Dataquest Incorporated May 8,1995

Semiconductor Application Markets Worldwide

15

a Mixed-signal graphics on a single chip (with RAMDAC) will implement high-resolution, surpassing 1 million pixels with 24-bit color, 3-D graphics, and digital video.

Q Mixed-signal I/O chips will support high-speed storage and peripherals communications (enhanced IDE, SCSI, P1394/Fiber Channel, SSA).

• Soimd emd video codecs and I/O jfunctions will be used in multimedia systems.

Projected semiconductor trends for computer peripherals and boards are as follows:

• Memory a Storage systenns (RDD/tape) will move from SRAM to DRAM. a 20MB SRAM/DRAM memory cards and 100MB flash memory cards will be used for storage. a Graphics cards will move from average 0.5MB to 2MB DRAM/VRAM.

• MPU/MCU a Storage systems will move from 8-bit to more sophisticated 16- and

32-bit processors with digital signal processing (DSP) capability.

Table 5

Computer Peripherals, Boards, and Functions: 1994 to 1999 Unit Growth

Slow (0 to 6%) Moderate (7 to 12%) Emerging (More than 12%) Declining (Negative)

Storage

More than 5.25-inch HDD More than 5.25-inch optical drive

3.5-inch HDD (greater than

400MB)

2.5-inch, 1.8-inch RDD

C D - R O M / R / X A / 1

Magneto-optical drives

Disk arrays (RAID)

Solid-state drives/PCMCIA memory cards

1/4-inch cartridge/4mm helical scan tape drives

Multimedia/ graphics

(boards or functions)

C G A / E G A / s t a n d a r d VGA Analog video

Windows graphics accelerators

Accelerated digital video

Sound (8 bit) Sound (16 bit)

Speech processing

Terminals

Alphanumeric graphics

Source: Dataquest (May 1995)

X windows

SAI\/IIV1-WW-PD-9501

©1995 Dataquest Incorporated May 8,1995

16

Semiconductor Application Markets Worldwide

• ASIC/ASSP

• Mixed-signal, highly integrated storage controller chipsets increasingly wiU be used for controls and data path. PRML, Fiber Channel, and SSA are emerging technologies. a Single-chip graphics controllers that integrate RAMDAC, digital video, and timing, among others, wiU predominate in mainstream applications.

Projected semiconductor trends for other data processing equipment are as follows:

• Memory

Q High-end printer/copiers will move to 4MB to 8MB memory. a Flash will replace ROM/EPROM in systems (4MB or greater).

• M P U / M C U

a Single-chip controllers will be used in low-end printers and handheld terminals.

• ASIC/ASSP

• Mixed-signal chipsets wiU be used in most systems.

Table 6

Other Data Processing Equipment: 1994 to 1999 U n i t G r o w t h

Declining (Negative)

Word processors

Slow (0 to 6%)

Dot matrix printers

Moderate (7 to 10%)

Laser printers

(1 to 10 ppm)

Ink jet printers Plain paper printers

Point-of-sale terminals

Bank/teller systems

Emerging (More t h a n 12%)

Laser printers

(more than 10 ppm)

Color printers

Host-based printers

Personal PPCs

Color/digital copiers

Scanners

Handheld terminals

Smart cards

Smart/crypto cards

Source: Dataquest (May 1995)

Projected semiconductor trends for communications equipment are as foUows:

• Memory a LAN systems increasingly will employ flash (1MB or greater) and

DRAM/SRAM (4MB or greater) memory. a Public/WAN systems will use denser and higher-speed DRAM/

SRAM and special function memory. a Low-power flash memory increasingly will be used in mobile devices.

SAMI\/I-WW-PD-9501

©1995 Dataquest Incorporated May 8,1995

Semiconductor Application Markets Worldwide

17

Table 7

Communications Equipment: 1994 to 1999 Unit Growth

Slow (0 to 6%)

Moderate (7 to 12%) Declining (Negative)

Networking/ Image

Communications

Previous LAN standards (NICs)

Bridges

Previous modems

Mobile commutucations

Corded phones

Traditional hubs

Analog cellular

Ethernet/ token-ring cards

Routers, T / E carrier systems

Standalone fax

V.32/bis modem

Analog cordless

One-way pagers

Transmission/ broadcast

Digital microwave Digital local loop systems

VSAT

' •

Emerging (More than 12%)

Fast Ethernet/Any LAN

NICs FDDI NICs, ATM NICs

Wireless LAN, NICs, and hubs

Switched and 100-Mbps hubs

ATM switches

Frame Relay systems

ISDN adapters

F a x / m o d e m cards/PCMCIA

V.34 m o d e m

Videoconferencing

Video phones

Digital cordless

Base stations

P C N / P C S / d i g i t a l ceUular terminals

Satellite phones

Wireless data communication

Two-way pagers

(narrowband PCS)

SONET/SDH transmission

Digital cable/DBS transmission

ADSL/HDSL transmission

HDTV upgrade

Switching and call processing

Analog line cards

Source: Dataquest (May 1995)

PBX/CO switch

Voice response systems

Automatic call distribution systems

Voice mail systems

Wireless PBX

N-ISDN, ATM

SAWIM-WW-PD-9501

©1995 Dataquest Incorporated May 8,1995

18 Semiconductor Application Markets Worldwide

• M P U / M C U

Q 32-bit MPUs will proliferate in LAN internetworking systems such as hubs.

Q Mobile systems will move from 8- to 32-bit MPUs and DSPs.

• ASIC/ASSP

Q PC and workstation connectivity increasingly will be implemented with single-chip, mixed-signal ICs. Each standard will command its own dedicated versions (Ethernet, including 100-Mbps versions;

Token Ring; FDDl; AnyLAN; ATM). a Advanced transceiver technology will be used for high-speed LANs

(more than 100 Mbps). a CMOS/BiCMOS technology and advanced macro cell libraries wiU be used for public/WAN devices.

• More highly integrated mixed-signal baseband chipsets will be used in mobile applications to implement various standards such as GSM,

TDMA, and CDMA.

• Other a Mobile devices wiU employ a new generation of integrated, lowpower, radio frequency/infrared transceivers.

• GaAs will find some use in transmit/receive fxmctions.

• Many passive devices such as filters and active discretes will be incorporated into more integrated ICs.

Projected semiconductor trends for consumer equipment are as follows:

• Memory

• SRAM/DRAM memory will be used for the first time in many systems.

• M P U / M C U

• Processors ranging from 4 to 32 bits and DSP MPUs increasingly will be used in systems.

• ASIC/ASSP

• New designs will employ digital/mixed-signal chipsets. a There is an overall trend from analog to digital technology.

SAMM-WW-PD-9501 ©1995 Dataquest Incorporated May 8,1995

Semiconductor Application Markets Worldwide 19

Table 8

Consumer Equipment: 1994 to 1999 Unit Growth

Moderate (7 to 12%)

Declining (Negative) Slow (0 to 6%)

Video

B / W T V Less than 30-inch color TV

VCR

Camcorder

Audio

More than 30-inch color TV

Laser disc

Emerging (More than 12%)

Digital/HDTV

Digital/HDTV VCR

Digital/HDTV camcorder

HDTV laser disc

Digital Video Disc (DVD) systems

Video karaoke

Digital cable and satellite codec/ converters

Analog cable and satellite converters

Systems/ components Embedded CD personal stereo

Digital Compact Cassette

(DCC)

Musical instruments

Mini Disc (MD) digital audio

Personal

8-bit video game

Calculators

16-bit video game

32/64-bit interactive players

Watches CD-based interactive players

Miscellaneous

White goods

Microwave ovens

Source: Dataquest (May 1995)

HVAC systems

Security systems

Mobile GPS terminals

Smart appliances

Projected semiconductor trends for automotive equipment are as follows:

'

• Memory

Q EPROM memory will be displaced by flash memory.

• MPU/MCU a 32-bit processors increasingly will be used in drive-train systems. 8and 16-bit MCUs with integrated analog interfaces and power controls will be increasingly popular.

• ASIC/ASSP

Q Systems will incorporate mixed-signal/power ASICs.

• Others

• Solid-state sensors will be used in systems.

SAMM-WW-PD-9501 ©1995 Dataquest Incorporated

May 8,1995

20 Semiconductor Application Markets Worldwide

Table 9

A u t o m o t i v e E q u i p m e n t 1994 to 1999 U n i t G r o w t h

Declining (Negative)

Discrete engine controls

Slow (0 to 6%)

8/16-bit engine controls

A M / F M cassette stereo

Moderate (7 to 12%)

Solid-state switches

Motor controls

Temperature controls

CD embedded stereo

Security controls

Emerging (More than 12%)

32-bit power train controls

Solid-state sensors

Air bags

Four-wheel antilock braking systems

Active suspension

Electronic steering

Mini Disk DCC

Navigation/GPS

Intelligent Vehicle Highway

System

Electric-powered vehicle systems

Source: Dataquest (May 1995)

Dataquest Perspective

This updated summary of Dataquest's worldwide electronics equipment production forecast has highlighted major regional and market segment trends. As the market grows toward the $1 trillion mark, major opportunities and obstacles await competitors in the various market segments. As before, focus will be critical for semiconductor players in this dynamic and rapid-paced industry. The skills and resources required to successfully compete in growing markets will require concentrated attention on specific products and applications. As the challenges associated with converging markets continue to increase, more companies are entering into strategic alliances and partnerships that allow them to access new skills and technologies while leveraging their current core competencies. When executed properly, this approach wiU enable manufacturers to compete more successfully in bringing differentiated products to the market and realize profitable returns on their investment.

For More information...

Dale Ford, Industry Analyst (408) 437-8311

Internet address [email protected]

Via fax (408) 437-0292

The content of this leport represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in

Ll3l3QUCSl po't to other parties shall be made upon the written and express consent of Dataquest

_ , ©1995 Dataquest Incorporated—Reproduction Prohibited

m m aoompanyot ^ '^ '^

MSmm TheOun&Biadsticct Corporation Dataquest is a registered trademark of A.C. Nielsen Company

K

Perspective

S e m i c o n d u c t o r A p p l i c a t i o n Markets W o r l d w i d e

Research Brief

Low-Cost, High-Performance Connections:

Is 1394 a Serial Killer?

AbStrSCti IEEE bus standard 1394 is Hearing market rollout. It appears that advanced

digital consumer electronics will be the early adopter of this high-speed peripherals interconnect technology.

By Greg Sheppard

IEEE Bus Standard 1394

As computer and consumer systems migrate to the high-speed world of digital multimedia, a new opportunity has arisen for silicon suppliers to reduce system costs. IEEE bus standard 1394 (also known as High-

Performance Serial Bus) is proposed as a method of cost-effective highspeed interconnection between various systems, including PCs and accompanying storage and i n p u t / o u t p u t (I/O) peripherals, as well as advanced consumer audio, video, and interactive systems.

The standard is near final balloting etnd approval by the IEEE constituency.

It is expected to be finalized this ]wne or at the latest this September. At this time it is believed to be in its final form thus making a stable specification for developing semiconductor products.

An All-in-One Connector

1394 transmits data at rates of 100,200, or 400 Mbps over a flexible sixconductor wire. There are two signal pairs and one power pair; the latter can support power-down situations. The specification supports "live" insertion and can support u p to 63 nodes by daisy chaining (up to 16 hops) and

OataQuest

Program: Semiconductor Application Markets Worldwide

Product Cod8:SAMIVI-WW-RB-9501

Publication Date: February 27,1995

Filing: Perspective

Semiconductor Application Markets Worldwide

branching. Because of the high data rates and relative simplicity and economy, 1394 has the potential to become a universal interface for all types of office and consumer equipment.

1394 is positioned to compete with entrenched standards like SCSI (and its serial version) but is also doing battle in certain applications with new technologies like SSA and Fiber Channel.

Serial Players

Texas Instruments (Tl), Apple Computer (Apple's implementation is known as FireWire), and IBM have taken most of the initiative from a market development standpoint to get 1394 going. TI has a two-chip set on the market that implements the physical/transceiver functions and the linklayer function. Molex is on board as a supplier of cormectors and cables.

A group known as the 1394 Trade Association has been formed to perpetuate the standard into multiple markets. Additional participants include

Adaptec, AMD, Cypress Semiconductor, Fuji Film Microdevices, Lexmark,

Maxtor, Microsoft, NCR Microelectronics, National Semiconductor, Philips,

Seagate, Sony, and Toshiba.

Advantages: Plug-and-Play on the Cheap

Table 1 presents an overview of the different potential applications and our assessment of 1394's chance of use in each. This is based on a poll of different Dataquest analysts and industry players. Other uses could also emerge as connection prices come down. Key patrons already committed to this technology include Sony, for an emerging digital camera line for desktop videoconferencing; Fuji Film, for digital cameras; and the European Digital

Video Broadcast (DVB) group. DVB is Europe's response to the U.S. HDTV program, and based on initial support, it has a good chance of defining much of the world's consumer digital TV standards. There is a good probability that future digital camcorders/cameras, digital VCRs, and the emerging digital video disk will all have a 1394 interface. These groups were attracted to the low cost and universal connector concept advantages of

1394. A lesser-known advantage is that, unlike SCSI, 1394 does not require an active terminator.

Legacy Means Lethargy

Unlike digital consumer equipment where there is a chance for a fresh start, computers — and their storage and document management (such as printers and scanners) interfaces — are bound greatly by backward compatibility. To change a standard here, both the host (embedded or adapter card) and peripheral must have 1394 interfaces. Unless a few key players decide to break away, penetration of 1394 into these areas remains unlikely. The storage community is going with enhanced IDE and SCSI for now, while

Fiber Channel and SSA are on the horizon in some projects. The document management industry has given no indication that it will be implementing

1394, and no PC makers have committed to 1394.

When desktop videoconferencing grows in importance, the camera interface could stimulate broader use in PCs and workstations.

i

SAMM-WW-RB-9501 ©1995 Dataquest Incorporated February 27,1995

i

Semiconductor Application Markets Worldwide

Table 1

Chances of 1394*s Success in Various Applications

System

RDD/Optical Storage

Chance of

20 Percent

Penetration

by 1999 (%)

10

Printer

Scanner

Audio Systems

Advanced/HDTV

Digital VCR/Video Disk

Video Game/Video CD

Camcorder/Still Camera

Advanced Set-Top Box

Source; Dataquest (February 1995)

10

10

20

100

100

50

100

100

Alternatives

Enhanced IDE, SCSIx,

SSA, Fiber Channel

Centronics, high-speed parallel

Enhanced IDE, SCSIx

Proprietary

Proprietary

Proprietary

Analog

PCM, Analog

Proprietary

Dataquest Perspective

There appears to be significant momentum behind 1394, and the chance of broad deployment in the digital consumer electronics community in particular is good. Computer industry usage remains uncertain but could change dramatically if segment leaders in storage and document management change their position. The net of it is that the chip opportunity for

1394 could hit 1 million units by 1996 and 15 million units by 1999 in consumer systems alone. We expect the chipset average selling prices to drop into the $7 range by 1999, driving an annual worldwide market of over $100 million.

SAMM-WW-RB-9501 ©1995 Dataquest Incorporated February 27,1995

Semiconductor Application IVIarkets Worldwide |

For More Information...

Greg Sheppard, Director/Principal Analyst (408) 437-8261

Internet address [email protected]

Via fax (408) 437-0292

The content of this report represents our interpretation arul analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in

U3l3QUCSl part to other parties shall be made upon the written and express consent of Dataquest

, ©1995 Dataquest Incorporated—Reproduction Prohibited n n a company of ^ ^ '

tmU ThcDun&BradstreetCorpoiation Dataquest is a registered trademark of A.C. Nielsen Company

SEMICONDUCTOREQUIPMENT, MANUFACTURING, AND MATERIALS

WORLDWIDE 1995

TABLE OF CONTENTS

PERSPECTIVES

9501 2/20/95

9501 6/19/95

The Cyclicality of the Equipment Market: What Makes It

Tick?(DQ Predicts)

The 1995-1996 DRAM Fab Outlook—Can PCs Absorb That

Much Needed DRAM Capacity? (Telebriefing Summary)

MARKET ANALYSIS

9501 2/27/95

9502

9503

9504

9505

9501

9502

9503

9504

9505

9506

9501

9502

3/27/95

3/22/95

10/23/95

11/13/95

6/19/95

6/26/95

12/25/95

12/25/95

12/25/95

12/25/95

7/31/95

1/22/96

1994 North American Semiconductor Gas Market: Strong Year for

Specialty Gases (Market Analysis)

A Market Perspective on the SIA Lithography Road Map

(Market Analysis)

The Move toward 300mm Wafers: The Issues Surroimding When and

How—^Money Will Talk (Market Analysis)

Lithography Stepper Capacity: Bottleneck of Semiconductor

Production (Market Analysis)

1994 Stepper Market: Reflection of the Growing Strength in

Semiconductor Production (Market Analysis)

1994 Silicon Wafer Market Share Estimates (Market Statistics)

1994 Wafer Fab Equipment Market Share Estimates

(Market Statistics)

Americas Fab Database (Market Statistics)

Japan Fab Database (Market Statistics)

Europe, Africa, and the Middle East Fab Database (Market Statistics)

Asia/Pacific Fab Database (Market Statistics)

Midyear 1995 Forecast:: Capital Spending, Wafer Fab Equipment, and Silicon (Market Trends)

Year-End 1995 Forecast: Capital Spending, Wafer Fab Equipment, and Silicon (Market Trends)

V

/

Dataquest^ j^ E R T

251 River Oaks Parkway • San Jose • CA • 95134-1913 • Phone 408-468-8000 • Fax 408-954-1780

Wafer Fab Equipment Forecast Update: The More Things

Change, the IHore They Remain the Same

In this interim update, we have modified the April 1995 forecast for front-end wafer fab equipment in light of the recently announced revision of Dataquest's semiconductor forecast.

In short, the strength in the semiconductor market has translated into a strong average growth trend in equipment. We have kept the assumption that the equipment market has a cyclical nature, anticipating the migration to the 16Mb DRAM.

This Dataquest Alert presents the top-level regional update as a reference for our clients in advance of the next regularly scheduled forecast update in December.

The Forecast

Table 1 shows the top-line regional forecast. The compotmd average grow^th rate (CAGR) for

1994 to 2000 calls for a 20.5 percent growth rate. This is slightly higher than the semiconductor growth rate for the same period. In the semiconductor forecast, we estimated stable growth in every region to the end of the decade. However, the dynamics of supply and demand in production capacity play an important role in the equipment market and modulate the year-toyear growth rates.

The globalization of production, particularly in the DRAM market, caUs for relatively uniform swings in the equipment market across all regions. This holds true even for the Asia/Pacific market, as both DRAM spending and semiconductor contract manufacturing (SCM), also known as fotmdry, have been responsible for the hypergrowth in this region. However, anrplitude variations in the growth rates for equipment spending in Asia/Pacific and Europe will be more volatile than in the more mature markets of North America and Japan.

Currently the Dataquest fab database contains well over 100 announcements for new fabs (or major expansions and upgrades) to come on line between 1995 and 1998. The current growth in capacity began in 1993 and has been responsible for building the backlogs of equipment suppliers. For some equipment segments, such as lithography steppers, the backlogs stretch into 1997. Equipment suppliers, in turn, have built u p their capacity to respond to this strong demand. In 1996, suppliers will continue not only to supply fabs, but also to build capacity to contain and manage their backlogs. This shipment m o m e n t u m should carry into 1997, the turning point in our forecast.

October 24,1995 ©1995 Dataquest Incorporated

V

*' .

#

Dataquest Alert Semiconductor Equipment, Manufacturing & Materials Worldwide

By 1997, we estimate that DRAM suppliers will be shifting most of their "convertible" capacity from 4Mb to 16Mb production, when device shrinks and process improvements have driven up yields. Therefore, strong DRAM demand will be met by a larger supply of 16Mb devices.

This will drive the price per bit down and, in turn, slow capital spending. The forecast shows this softening in the market in 1998 as a slightly negative growth rate. We have also studied the supply-demand scenario in the SCM arena, with a strong emphasis on its impact in the

Asia/Pacific region. In short, until the latter part of 1997, we see tight supply of foundry capacity. With tiie new fabs coming on line in 1997, we expect growth in capacity of this type to slow by the end of 1998.

With sub-0.3 micron DRAM and logic devices ramping up in 1999, growth rates shoiild turn to the positive side again. The forecast period ends with a strong growth rate predicted for 2000.

The equipment market will consume about 10 percent of overall semiconductor revenue in

2000, nearly $33 biUion.

Table 1

1994 1995 1996 1997 1998

Total Wafer Fab

Eqmpment

Growth (%)

North America

Growth (%)

Japan

Growth (%)

Europe

Growth (%)

Asia/Padflc-ROW

10,756

49.8

3,141

49.8

3,668

45.1

1,385

23.9

2,562

Growth (%) 87.6

Source: Dataquest (October 1995)

16,672

55.0

4,553

45.0

4,995

36.2

2,391

72.6

4,733

84.7

23,109

38.6

5,748

26.2

6,742

35.0

3,031

26.8

7,588

60.3

25,430

10.0

6,330

10.1

7,194

6.7

3,422

12.9

8,483

11.8

23,851

-6.2

6,347

0.3

7,151

-0.6

3,268

-4.5

7,087

-16.5

1999

26,247

2000

32,956

CAGR (%)

1994-2000

20.5

10.0

7,073

11.4

7,850

9.8

3,508

7.4

7,816

10.3

25.6

8,555

21.0

9,587

22.1

4,185

19.3

10,629

36.0

18.2

17.4

20.2

26.8

We wiU provide updated segment forecasts relative to the above adjustments when we have completed our year-end survey of equipment suppHers. Until our December forecast is released, we recommend that clients simply ratio the values above to the segment forecasts presented in July.

By Nader Pakdaman, Calvin Chang, Clark Fuhs, and Yasumoto Shimizu

October 24,1995 ©1995 Dataquest Incorporated

; . ^

Dataquest^ ^ ^ j^ 2^

1290 Bidder Park Drive * San Jose * CA * 95131 -239« * Phone 4 0 « ^ 7 - « 0 0 0 • Fax 405^37-0292

Can the $10.8 Billion Wafer Fab Equipment Market in 1994

Grow to Nearly $20 Billion in 1996? Believe It!

We knew a few months ago that the 1995 market would surprise us with its strength. A monthly leading indicator w^e have been developing went crazy in April a n d h a s not rested since. Our May capital spending survey confirmed our indicator—that 1995 is basically 1994 all over again.

This Dataquest Alert w^iU be followed by publication of the Semiconductor Equipment,

Manufacturing, and Materials Worldw^ide services 1995 Midyear Forecast, which SEMM clients should receive w^ithin the next few weeks. We will also review this subject in an upcoming

Dataquest Teleconference briefing on July 7, 2995, at 8:30 a.m. PDT. Please contact Jenny Williams at

408-437-8263 by July 6 if you would like to be included in the teleconference.

Still Booming in 1995... and into 1996

Our capital spending survey just completed show^s 60 percent grow^fh over 1994 levels, but not all of this grow^th w^ill trickle immediately into equipment in 1995. Several n e w projects in

Taiwan are under w^ay, v/hich vv'iLl place the bulk of equipment into the fabs in 1996. European companies, primarily driven by Siemens, and multinational companies make Europe the second fastest-grow^ing region. Why the boom in Europe? With the region's economies recovering and the PC boom ongoing, Europe has attracted PC production, particularly in the

United Kingdom. Semiconductor production has moved along with the PC, with Intel and

DRAM producers w^orldw^ide taking part. The acceleration of telecommunications-related semiconductor production also benefits European companies such as Philips and Ericsson.

But the story behind the forecast of 52 percent wafer fab equipment growth in 1995, foUow^ed by 22 percent grow^th in 1996, is the continued heavy investment in DRAM capacity. The lack of availability of the critical 16Mb DRAM part at high yields, which restrains the PC market's conversion from the popular 4Mb generation parts, has forced continued investment in capacity. Persistent high prices for DRAMs has kept DRAM manufacturers profitable, w^hich in turn has spurred strong spending.

Spending by the three largest Korean companies is expected to increase 124 percent in 1995 to a combined total of $5.6 billion. Japanese company spending w^iU increase 44 percent as w^ell.

Three new projects in Taiwan involve new DRAM players (some with Japanese technology

July 6, 1995 ®1995 Dataquest Incorporated

Dataquest Alert

Semiconductor Equipment, Manufacturing, and Materials Worldwide partners): N a n Ya Plastics, Vanguard, and PowerChip. A n d IBM a n d Siemens are back, each spending an estimated $1 billion in 1995.

What about 1997?

As w^e have mentioned in the past, DRAM investment w^iU inevitably decline as capacity currently running 4Mb DRAMs converts to 16Mb parts, effectively doubling or tripling the bit capacity of the manufacturing line.

What this means is that the "pause" in the equipment vdU occur eventually, triggered by the conversion of the end-use market away from the 4Mb generation tow^ard the 16Mb DRAM.

This wiU h a p p e n only w h e n the 1x16 configuration 16Mb DRAM becomes widely available at reasonable yields, about 60 to 65 percent, and becomes economical to produce. This is not expected to occur until well into 1996. We now^ look toward 1997 as being the toughest year for the equipment industry, but w^e do not expect it to be down significantly because PC tmit shipments are expected to remain strong. PC xunits drive about one-third of the semiconductor market today, a strong underlying trend likely to limit the extent and duration of the slow^down.

The Forecast

Table 1 shows the regional topline wafer fab equipment forecast through the year 2000. Further details w^iU be reviewed at the upcoming teleconference and in the forthcoming report.

Table 1

Wafer Fab E q u i p m e n t Forecast, 1994-2000 ( M i U i o n s of U.S. DoUars)

Total Wafer Fab

Eqmpment

Percent Growth

1994

10,755

56.4

1995

16,340

1996

19,854

51.9 21.5

1997

18,888

-4.9

1998

19,323

1999

22,495

2000

29,701

2.3 16.4 32.0

1994-2000

CAGR (%)

18.4

N o r t h America

Percent Growth

Japan

Percent Growth

Eiirope

Percent Growth

Asia/Padfic-ROW

Percent Growth

Source: Dataquest (July 1995)

3,141

47.5

3,668

49.1

1,385

41.6

2,562

95.7

4,409

40.4

5,008

36.5

2,341

69.0

4,582

78.8

5,040

14.3

5,459

9.0

2,740

17.0

6,615

44.4

4,966

-1.5

4,931

-9.7

2,696

-1.6

6,296

-4.8

5,160

3.9

4,953

0.4

2,842

5.4

6,369

1.2

6,065

17.5

5,643

13.9

3,317

16.7

7,470

17.3

7,801

28.6

7,146

26.6

4,002

20.7

10,751

43.9

16.4

11.8

19.3

27.0

By Clark Fuhs

July 6, 1995

®1995 Dataquest Incorporated

t Dataquest Teleconference

1290 Ridder Park Drive San Jose. CA 95131 Phone 408-437-8000 Fax 408-437-0292

TOPIC: CAN THE WAFER FAB EQUIPMENT MARKET

EXPAND FROM $10.75B IN 1994 TO NEARLY $20B

IN 1996? BELIEVE IT!

SCOPE: The session opens with comments by analysts from the

Semiconductor Equipment, Manufacturing, and

Materials (SEMM) Worldw^ide service. An interactive discussion with Dataquest clients follows.

W H O : Discussion leaders are Clark Fuhs, Nader Pakdaman, and Calvin Chang of the SEMM service. Invitees will consist of SEMM, SPSG, and Cross-Industry clients.

WHEN: Friday, July 7,1995 at 8:30 a.m. PDT.

HOW: To confirm your attendance, call Jenny Williams (408)

437-8263, or fax Jenny at (408) 437-0292 no later than

July 6,1995.

Clients w h o respond by July 5 will receive figures (via fax) before the teleconference.

IMPORTANT: Dataquest offices will be closed July 3 and 4.

Please use attached "Fax Back" form to register early!

Page 2

#

B^taquestji^ j^ ^ j^ J.

1290 Ridder Park Drive • Sax. Jose • CA * 95131 -239« • Phone 40«-437-6000 * Fax 40*^37-0292

Formosa Plastics and Komatsu Forge Alliance in Taiwan

Komatsu Electronics Corporation and the Asia-Pacific Investment Co., which is owned by the

Formosa Plastics Group (PPG), this w^eek signed a m e m o r a n d u m to build a silicon w^afer manufacturing factory in Taiwan. The name of the new^ company w^ill be Taiw^an Konnatsu

Electronic Materials Corporation, which will have capital investment of U.S.$250 million.

Ownership will be divided 51:49 in favor of Komatsu. Construction is planned to begin by the fourth quarter of 1995, and mass production is forecast to begin by 1997.

This is a technology transfer arrangement from Komatsu. The specified final capacity of 8-inch slices is about 200,000 wafers per month, but the first phase to be completed in 1997 will be for

80,000 slices. Both companies agree that a formal contract completing the joint venture will be signed by each party no later than July. FPG Director Y.C. Wang and Komatsu Director Mr.

Nakanishi joined the meeting to sign the memorandum. Each director expressed his optimistic view for the future of this company and said details regarding the agreement w^ere still being worked out.

Following the recent successes of Taiwan's semiconductor device manufacturing companies, local investors have been eyeing long-term profits and trying to move upstream into the silicon w^afer manufacturing business. This is the second recent silicon w^afer a n n o u n c e m e n t following Taisil Electronic Materials Corporation's entry into the field. Table 1 compares the similarities and differences between these two joint venture agreements.

Dataquest Perspective

The Taiwanese semiconductor industry has been expanding rapidly in the past three years but has outpaced supportive semiconductor industries such as silicon wafers. These recent announcements are positive signs for the local industry, which is completely dependent on imported wafers. As for long-term success, both venture companies highlighted in Table 1 will have strong technical and financial resources: FPG is the biggest privately held group in

Taiwan, and the China Steel Corporation (CSC) is a government-supported company w^ith capital of more than $2.7 billion; MEMC Electronic Materials Corporation and Komatsu are global technology leaders in wafer production. The cooperation of these companies ensures their ventures will have the capital to expand and the technology to compete. Neither will have to worry about demand.

May 18, 1995 ©1995 Dataquest Incorporated

#

Dataquest Alert

Table 1

Semiconductors Asia/Pacific

Comparison b e t w e e n Taiw^an K o m a t s u Electronic Materials Corporation a n d Taisil

Electronic Materials Corporation

Companies

Investors

Technology Source

Capital ($M)

Capacity (8-Inch Wafers/Month)

Initial Capacity (8-Inch/Month)

Pilot Time

Initial Production Time

Final Capacity Reahzed

Shareholders

Location

Source: Dataquest (May 1995)

Taiwan-Komatsu

FPG, Komatsu

Komatsu (Japan)

250

200,000

80,000

Third quarter of 1996

Fourth quarter of 1996

To be announced

Komatsu (51 percent),

FPG (49 percent)

Not available

Taisil

CSC (Taiwan), MEMC

MEMC (United States)

191

125,000

70,000

Third quarter of 1996

Fourth quarter of 1996

Late 1997 to early 1998

MEMC (45 percent),

CSC (35 percent), others (20 percent)

Hsin-Chu Science Park

These two plants, taken together, could appreciably supply Taiwan's production needs for

200mm wafers. However, we believe that a portion of these plants output w^Hl be exported, resulting in a continued net deficit of 200mm wafers for Taiwan. Dataquest beheves Taiwan's semiconductor industry's appetite for silicon wafers will far exceed local production unless further investments are made. By 1997, when both companies' production becomes fully ramped, as planned, these facilities will be able to supply less than 50 percent of Taiw^an's total wafer demand. Nevertheless, this increasing high level of interest in the semiconductor industry among Taiw^an's largest conglomerates w^ill mean Taiwan has the financial wherewithal to sustain or, quite possibly, accelerate its expansion r u n rate.

By Ben F.P. Lee (Taiwan) and Daniel Heyler (Hong Kong)

t

May 18, 1995 ©1995 Dataquest Incorporated

Datanuest^ ^ ^jg y

1290 Ridder Park Drive • San Jose * CA * 95131 -239a + Phone •40^^7-5000 • Fax 405-437-0292

Midterm Update: Wafer Fab Equipment Forecast—Stronger

1995 and 1996

Based on bookings levels and capital investment patterns, particularly in DRAM in Asia and

Japan, w^e m u s t modify our December 1994 forecast fairly dramatically in the near term. This

Dataquest Alert presents the top-level u p d a t e as a guide to our clients, in advance of the next regularly scheduled forecast u pd at e in July.

Booking Levels Strong: 1995 Shows Healthy Growth

Indications are that bookings in the first half of 1995 will finish at year-over-year grow^th levels of about 35 to 40 percent (below^ a peak of more than 50 percent growth early in 1994). These strong booking levels in the first half suggest that Dataquest's earlier projection of 16 percent growth in 1995 for wafer fab equipment is much too low^. We have revised this near-term forecast to be more than 30 percent grow^th for 1995.

Reasons for the continued grow^th include the continued heavy investment in DRAM capacity, as outlined in a Dataquest Alert published a few^ w^eeks ago. The lack of availability of the critical

16Mb DRAM part at high yields that restiain the PC market from converting aw^ay from the popular 4Mb generation parts has forced continued investment in capacity. Persistent high prices for DRAMs have kept DRAM manufacturers profitable, which in turn has spurred strong spending.

It m u s t be noted that our assumptions have included the end-user markets, where semiconductor consumption is increasing stiongly. The continued stiength in markets such as

PCs and telecorrvmunication w^ill play an important role in how^ Dataquest foresees the future.

How^ever, w^e also believe that historical fluctuations in DRAM capacity investments have been the vtnderlying reason for the cyclical patterns seen in capital spending. Our analysis has shown that the over- and ttnderspending cycles correlate closely to the price per bit of DRAM devices. We believe this cyclical pattern w^ill persist and w^ill be coupled with the more stable grow^th in other segments of the industiy, in particular the investments in microprocessors and logic-oriented devices.

In a recent Dataquest Predicts new^sletter w^e asked the question, "What if the indicators... remain positive for the next six months?" Our answ^er was, "We w^ould probably n u d g e u p grow^th... for 1995, place 1996 in slightly positive territory... and bring 1997 d o w n significantly (in grow^th)." This has now^ become reality.

April 17, 1995 ®1995 Dataquest Incorporated

t ,

Dataquest Alert

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Momentum in Capacity Additions Mean Positive Growth for 1996, but

What about 1997?

We can n o longer justify a decline year for 1996, and the current forecast here of 9 percent growth m a y prove on the conservative side. The growth rate is more likely to be in the low^ teens. How^ever, as w^e mentioned in our Dataquest Alert last month, DRAM investment will inevitably decline as capacity now^ running 4Mb DRAMs converts to 16Mb parts, effectively doubling or tripling the bit capacity of the manufacturing line.

What this means is that the "pause" in the equipment will occur eventually a n d will be triggered by the conversion of the end-use market aw^ay from the 4Mb generation tow^ard the 16Mb. This wiil only h a p p e n w^hen the 1x16 configuration 16Mb DRAM becomes widely available at reasonable yields. This ramp-up, w^hich w^e expected during 1995, h a s been p u s h e d back some six to nine months, by our estimate. This means the current equipment b o o m has been extended, and the equipment slowdow^n delayed. We now^ look tow^ard 1997 as being the toughest year for the equipment industry.

The Forecast

Table 1 show^s the regional top-line w^afer fab equipment forecast as w^e now^ see it, with an initial look into 1999. Our rough estimate for the year 2000, because of the cycHcaHty of the

DRAM market and the expected capacity r a m p of 16/64Mb fabs, calls for about 35 percent grow^th, placing the wafer fab equipment market at $26 billion to $27 biUion in the year 2000.

Table 1

Wafer Fab Equipment Forecast, 1994-1999 (Millions of U.S. Dollars)

Total Wafer Fab

Equipment

Percentage Growth

1994

10,350

49.8

1995

13,531

1996

14,740

30.7 8.9

1997

13,991

1998

15,368

-5.1 9.8

1999

19,723

CAGR (%)

1994-1999

14.0

28.3

North America

Percentage Growth

Japan

Percentage Growth

Etuope

Percentage Grow^th

Asia/Padfic-ROW

Percentage Grow^th

Source: Dataquest (April 1995)

3,105

43.1

3,517

45.1

1,198

23.9

2,530

87.6

3,800

22.4

4,419

25.6

1,550

29.4

3,762

48.7

4,228

11.3

4,661

5.5

1,620

4.5

4,231

12.5

4,409

4.3

4,276

-8.3

1,547

-4.5

3,759

-11.1

4,754

7.9

4,417

3.3

1,607

3.9

4,590

22.1

6,056

27.4

5,358

21.3

2,004

24.7

6,305

37.4

14.3

8.8

10.8

20.0

April 17, 1995

®1995 Dataquest Incorporated

-w

Dataquest Alert Semiconductor Equipment, Manufacturing, and Materials Worldwide

Individual Segment Forecasts

Individual segment forecasts are not yet available in this top-level forecast. W e do expect to release an u p d a t e for individual segments consistent with the revised forecast in a few^ w^eeks.

We are w r a p p i n g u p the final market share statistics for 1994 and w o u l d like to report final

1994 segments. In the meantime, w e w^ould suggest that a first-order approximation w^ould be to use Table 3-4 from the December 26,1994 Market Trends forecast report a n d ratio the top-line numbers in each year throughout tiie segments.

Clark Fuhs and Nader Pakdaman

April 17, 1995 ®1995 Dataquest Incorporated

Dataquest^ ^ ^j^ y

1290 Ridder Park Drive • San Jose • CA • 95131-2398 • Phone 408-437-8000 • Fax 408-437-0292

Midterm Update: Wafer Fab Equipment Forecast-

Stronger 1995 and 1996

Based on bookings levels and capital investment patterns, particularly in DRAM in Asia and

Japan, we must modify our December 1994 forecast fairly dramatically in the near term. This

Dataquest Alert presents the top-level update as a guide to our clients, in advance of the next regularly scheduled forecast update in July.

Booking Levels Strong: 1995 Shows Healthy Growth

Indications are that bookings in the first half of 1995 will finish at year-over-year growth levels of about 35 to 40 percent (below a peak of more than 50 percent growth early in 1994). These strong booking levels in the first half suggest that Dataquest's earlier projection of 16 percent growth in 1995 for wafer fab equipment is much too low. We have revised this near-term forecast to be more than 30 percent growth for 1995.

Reasons for the continued growth include the continued heavy investment in DRAM capacity, as outlined in a Dataquest Alert published a few weeks ago. The lack of availability of the critical

16Mb DRAM part at high yields that restrain ttie PC market from converting away from the popular 4Mb generation parts has forced continued investment in capacity. Persistent high prices for DRAMs have kept DRAM manufacturers profitable, which in turn has spurred strong spending.

It must be noted that our assumptions have included the end-user markets, where semiconductor consumption is increasing strongly. The continued strength in markets such as

PCs and telecommtmication will play an important role in how Dataquest foresees the future.

However, we also believe that historical fluctuations in DRAM capacity investments have been the underlying reason for the cyclical patterns seen in capital spending. Our analysis has shown that the over- and underspending cycles correlate closely to the price per bit of DRAM devices. We believe this cyclical pattern will persist and wiU be coupled with the more stable growth in other segments of the industry, in particular the investments in microprocessors and logic-oriented devices.

In a recent Dataquest Predicts newsletter we asked the question, "What if the indicators... remain positive for the next six months?" Our answer was, "We would probably nudge up growth...

April 14.1995 ©1995 Dataquest Incorporated

Dataquest Alert Semiconductor Equipment, Manufacturing, and Materials Worldwide for 1995, place 1996 in slightly positive territory... and bring 1997 down significantly (in growth)." This has now become re£dity.

Momentum in Capacity Additions IVlean Positive Growtli for 1996, but Wliat about 1997?

We can no longer justify a decline year for 1996, and the current forecast here of 9 percent growth may prove on the conservative side. The growth rate is more likely to be in the low teens. However, as we mentioned in our Dataquest Alert last month, DRAM investment will inevitably decline as capacity now running 4Mb DRAMs converts to 16Mb parts, effectively doubling or tripling the bit capacity of the manufacturing line.

What this means is that the "pause" in the equipment will occur eventually and wiU be triggered by the conversion of the end-use market away from the 4Mb generation toward the 16Mb. This will only happen when the 1x16 configuration 16Mb DRAM becomes widely available at reasonable yields. This ramp-up, which we expected during 1995, has been pushed back some six to nine months, by our estimate. This means the current equipment boom has been extended, and the equipment slowdown delayed. We now look toward 1997 as being the toughest year for the equipment industry.

The Forecast

Table 1 shows the regional top-line wafer fab equipment forecast as we now see it, with an initial look into 1999. Our rough estunate for the year 2000, because of the cycUcality of the

DRAM market and the expected capacity ramp of 16/64Mb fabs, calls for about 35 percent growth, placing the wafer fab equipment market at $26 billion to $27 billion in the year 2000.

Table 1

Wafer Fab Equipment Forecast, 1994-1999 (Millions

Dollars)

Total Wafer Fab

Equipment

Percentage Growth

1994

10,350

49.8

1995

13,531

1996

14,740

30.7 8.9

1997

13,991

1998

15,368

-5.1 9.8

1999

19,723

CAGR (%)

1994-1999

14.0

28.3

North America

Percentage Growth

Japan

Percentage Growth

Europe

Percentage Growth

Asia/Pacific-ROW

Percentage Growth

Source; Dataquest (April 1995)

3,105

43.1

3,517

45.1

1,198

23.9

2,530

87.6

3,800

22.4

4,419

25.6

1,550

29.4

3,762

48.7

4,228

11.3

4,661

5.5

1,620

4.5

4,231

12.5

4,409

4.3

4,276

-8.3

1,547

-1.5

3,759

-11.1

4,754

7.9

4,417

3.3

1,607

3.9

4,590

22.1

6,056

27.4

5,358

21.3

2,004

24.7

6,305

37.4

14.3

8.8

10.8

20.0

April 14,1995 ©1995 Dataquest Incorporated

'V

Dataquest Alert Semiconductof Equipment, Manufacturing, and Materials Worldwide

A Individual Segment Forecasts

Individual segment forecasts are not yet available in this top-level forecast. We do expect to release an update for individual segments consistent with the revised forecast in a few weeks.

We are wrapping up the final market share statistics for 1994 emd wovdd like to report final

1994 segments. In the meantinne, we would suggest that a first-order approximation would be to use Table 3-4 from the December 26,1994 Market Trends forecast report and ratio tiie top-line numbers in each year throughout the segments.

Clark Fuhs and Nader Pakdaman

April 14,1995 ©1995 Dataquest Incorporated

o^'P'^ALERT

12S0 Ridder Park Drive * San Jose * CA * 99131 -239« * Phone « ) « < } 7 - d 0 0 0 * Fax 40^^37-0292

"The 1995-1996 DRAM Fab Outlook-Can PCs Absorb

That Much DRAM Capacity?"

Dataquest's Semiconductor Group held the above-entitled telebriefing on March 23,1995. This

Dataquest Alert presents the key figures and opening statement from the telebriefing.

Introduction

This is Jim H a n d y of Dataquest's Memories Worldwide service. First, let m e introduce ttie speakers, then I'U give an overview of the subject and structure of the telebriefing. In the room here at Dataquest we have Clark Fuhs, Nader Pakdaman, a n d Calvin C h a n g of the

Semiconductor Equipment, Manufacturing, and Materials Worldwide service (SEMM); Mark

Giudici and Scott Hudson from the Semiconductor Procurement Worldwide service; and Mario

Morales from our Research Operations Group. Ron Bohn and 1 represent the Memories

Worldwide service.

What is this telebriefing about?

• Dataquest performs an emalysis of DRAM supply and demand, and publishes the result in a report updated quarterly.

• There has been a big increase in capital spending for semiconductor processing equipment. Despite all the new equipment coming online, the DRAM shortage will persist through 1996.

• This teleconference will describe the reasons that the DRAM shortage will continue, despite the installation of so much semiconductor processing equipment.

First, Clark Fuhs wiU speak about semiconductor wafer fabrication plant capacity, then I will return with a statement about the DRAM market. After these statements, w e will open the phones to questions from the participants.

DRAM Supply-Side

I am Clark Fuhs, a senior analyst from the SEMM group... We track most aspects of the actual manufacturing of semiconductors worldwide... Today we wiU briefly review the dynamic arena of DRAM capacity and supply.

March 29, 1995 ©1995 Dataquest Incorporated

Dataquest Alert Memories Woridwide

The key source for our analysis today is a report Dataquest has just issued on 4Mb and 16Mb

DRAM supply and demand. We entitled it "DRAM Supply and Demand Report."

Supply-Side Analytical Methodologies

There ztre two supply-side methodologies, a n d their differences are outlined in Figure 1. Shown are three snapshots in time—at the ends of 1994,1995, a n d 1996. The left bar in each year represents results of a svirvey of DRAM suppliers, which will be referred to by Jim H a n d y later... The right hea represents an estimate based on a fab-by-fab analysis of capacity a n d committed plans.

Figure 1: Estimated Fob Monthly Run

Rates for 4 Mbit and 16 Mbit DRAMs

Trillions of bits per month

1800

1600 -I

1400

1200

1000

800 H

600

400

200 H a Est. Capacity -

DRAM Suppiier

Survey

@ Est. Capacity -

Fab-by-Fab

Analysis

End of

1994

End of

1995

End of

1996

Source; Dataquest

The capacity estimate through the fab analysis includes the following assumptions: die size • estimates, the effects of shrinks, and a gradual yield increase from today through the e n d of

1996 factored in on a company-by-company basis.

J.

We have not included any capacity associated with unannoimced capital spending increases or commitments over the next two years such as the recently announced acceleration of NEC's n e w U.S. faciUty—nor 6- to 8-inch wafer conversions.

As seen in Figure 1, these two methodologies have produced results within 3 percent of each other for the year just completed. But as w e go into the future, the figure shows an increasing divergence. The conclusion here is that suppliers are not optimistic in the near term regarding their ability to increase yields on the 1x16 configuration for the 16Mb DRAM, a n d their outlook is conservatively hedged.

i

I

i

March 29, 1995 ©1995 Dataquest Incorporated

Dataquest Alert Memories Worldwide

This leads to a question w e get fairly often: What is the capacity of the market in 4Mb

"equivalent tmits"? Ui\fortvmately, from a fab perspective, this is not a very useful w a y to view capacity—and is a metric that w e at Dataquest view may not provide an adequate picture of bit capacity. Why not? In a fab, capital spending and equipment purchases are driven by the requirement for wafers, or better yet, square inches of silicon.

Three Stages of Capital Spending

Let m e highlight here the dynamics of how capital spending and bit supply are related. There are three identifiable parts to this cycle, a n d as you wiU see, w e are in the later stages of part 2.

In the early to middle stages of a imit ramp in a specific DRAM density generation, square-inch requirements and equipment purchases are generally driven by bit demand. This is part 1 of the investment cycle, and it occurred in 1993 for the 4Mb generation.

In the later stages of ramp, as the next density generation starts to become available, w e enter part 2 of the investment cycle. During this state, capital investment is still primarily driven by bit demand. How^ever, the style of investment changes to install "convertible" capacity—in today's case equipment earmarked for 16Mb DRAM capacity, but initially rimning 4Mb parts.

Most investments in 1994 into the present have been of this nature. The later stages of part 2, v\^hich we are in today, will also tend to include dedicated capacity for the next density generation.

So today there are three types of capacity we m u s t consider—capacity dedicated to 4Mb

DRAMs, capacity dedicated to 16Mb DRAMs, and capacity that is "convertible" between the two. Why do suppliers like "convertible" capacity?

Based on die size ratios and the bit density ratio, as well as some other minor factors, a supplier can generally increase the bit capacity of a line by converting the line from 4Mb to 16Mb parts, with minimal incremental capital or equipment spending. This is accomplished because the bits per square inch are increased on the order of two or three times, meaning a supplier can double or triple the bit capacity of a line by conversion. Thus part 3 of the investment cycle, typically lasting two years, creates bit capacity primarily by conversion rather than n e w equipment purchases. Thus a "pause" in the equipment market ensues. When all the capacity is converted, w e begin part 1 of the cycle all over again.

Industry's DRAM Capacity

Back to the original question: What is the DRAM capacity of the industry? The answer is—it depends. It depends on how the "convertible" capacity is employed. The most useful w a y to view capacity is by a "window" and this is shown in Figure 2, using a fab-by-fab analysis. What

1 mean by "window" here is the area between the two lines that represent the possible capacity measured in terms of bits. Let's first relate this to Figure 1.

March 29, 1995 ©1995 Dataquest Incorporated

Dataquest Alert Memories Worldwide

Figure 2: Monthly DRAM Bit Capacity

''Window''for 4 Mbit and 16 Mbit

Source; Dataquest

2000

Trillion bits per mont

"Convertable" at 100%

1600

1200

800 -.m'

4O0

End Of

1994

End

Conveitable" at 100% 4 Mbit

of

1905

End of

1996

At the end of 1994, Figure 1 shows that the monthly r u n rate of the industry w a s slightly more than 600 trillion bits. If all the "convertible" capacity were running 4Mb DRAMs at the e n d of

1994, the r u n rate would calculate to about 550 trillion bits p e r month, shown as t h e lower limit line in Figure 2. Likewise, if all of this convertible capacity were running 16Mb parts at today's yields, the capacity would calculate to slightly less than 800 trillion bits per month. Employing this methodology into the future, w e have produced Figure 2 as a remge of capacity over time.

1 w^ould like to emphasize that the width of the window increases in 1995 a n d shrinks in 1996.

Because of the low yield a n d slow r a m p issues of the 1x16 configuration part, suppliers are being forced to a d d more 4Mb capacity of a "convertible" nature today. This will increase the growth for the front-end equipment market during 1995 weU beyond our current published forecast of 16 percent—probably not far away from 30 percent growth. M o m en t u m factors will establish 1996 as a small growth year as well (originally w e indicated a slight decline).

Our model indicates that the "pause" in the equipment market that w e have been forecasting is unavoidable, however, because installing "convertible" capacity today by definition installs future "hidden" bit capacity and will likely cause a slight decline for wafer fab equipment in

1997. This conversion stage will also reduce the growth in the consumption of silicon in a like manner.

The trigger for the pause will be the availability a n d subsequent pull of the 1x16 configuration part into the end-use market. Until this happens, silicon square inch d e m a n d will continue to be driven by bit demand closely. The severity and length of the ultimate "pause" will be determined by h o w long the current boom lasts, basically building pent-up bit supply as well as the d e m a n d for bits d u r i n g the conversion stage.

March 29, 1995 ©1995 Dataquest Incorporated

Dataquest Alert Memories Worldwide

N o w I'd like to turn it back over to Jim Handy for a review of demand.

DRAM Demand-Side

This is Jim Handy. Today's DRAM market is phenomenal. There h a s been cm undersupply for two-and-a-half years, and in our recent analysis, the February 1995 version of tiie quarterly

"DRAM Supply and Demand Report," w e have found little reason to expect the current shortage to ease through the end of the report's forecast window, the e n d of 1996.

4Mb DRAM prices have slowly risen since the third quarter of 1992, yet DRAM use in megabytes per PC has gone u p .

DRAM Pricing Alert

In the face of a strengthening yen against the dollar, Japanese suppliers of DRAMs h a v e been able to raise average selling prices measured in dollars to stabilize the yen value of their worldwide DRAM sales. By aU appearances, this trend wiU continue through the end of next year.

Our current North American contract-volume DRAM price forecast calls for firm pricing. We are aw^are that some Japan-based DRAM suppliers right now are considering an increase in the contract pricing for DRAMs. Korean and North American DRAM suppliers are waiting in the wings to see what happens. Should these changes materialize during second-quarter price negotiations, w e will make appropriate changes in our price forecast.

Let's go over some of the backgroiind causes of the current DRAM shortage:

• PC demand is strong and shows no signs of letting u p .

• Japanese DRAM suppliers, which accoxmted for nearly 50 percent of worldwide sales, were slow to react to the market and didn't increase capital spending until the fourth quarter of 1993.

• PCs have increased their per-system consumption of DRAM despite these stable/rising prices both in response to the requirements of advanced software as w^eU as to decreases in the prices of CPUs, chipsets, and other system components.

• The 16Mb DRAM has met with limited acceptance in the traditional xl and x4

' organizations. A new xl6 organization is the product of preference, and is late to market, forcing an increase in the consumption of 4Mb DRAMs unlike that seen in any other DRAM generation.

The PC market has stayed in a strong growth phase for some time now. However, the cause of this strength has varied over time. Two years ago growth was fueled by a rebound in PC shipments to the office. More recently, a big growth in h o m e computer purchases including multimedia systems has driven the market. Now the business channel once again stands poised for important growth.

March 29, 1995 ©1995 Dataquest Incorporated

Dataquest Alert

PC Forecast Revised Upward

Memories Worldwide

We w^ant to highlight today that Dataquest has just revised its w^orldwide PC unit-shipment forecast upward. Dataquest's PC shipment forecast calls for 57 million xinits to ship in 1995, and in excess of 65 miUion units in 1996. This means a strong 19 percent growth in w^orldwide

PC tmit shipments this year a n d equally impressive 18 percent growth for 1996.

Within each of these PCs w e see another j u m p in the amount of DRAM consumed. The m i n i m u m DRAM size wiU increase from the 4MB per PC that shipped in 486-based PCs in

1994 u p to 6MB w h e n the 8MB Pentium m i n i m u m is averaged with the 4MB 486 this year.

N o w let's took at Figures 3 and 4. Figure 3 shows the annual bit d e m a n d of the total data processing market zind other DRAM markets. We derive this number from our electronic equipment forecast, which includes PC and other data processing forecasts as well as forecasts for other DRAM applications such as printers, fax machines, video games, hard disk drives, and even digital set-top boxes. This is the demand side of our supply d e m a n d analysis, and is driven to a great extent by the health of the PC market.

Figure 3: DRAM Demand in Bits

Trillions of Bits

18,000

16,000

14,000

12,000

10,000

8,000

6,000

4,000

2,000

0

1994

1995 1996

Source: Dataquest

vf

March 29, 1995 ©1995 Dataquest Incorporated

6

Dataquest Alert Memories Woridwide

Figure 4: 1995-1996 DRAM Supply-

Demand Scenario (Supply as % of Demand)

0%

-2%

-4%

-6%

-8%

-10%

-12%

-14%

-16%

-18%

^

1994 1995 1996

Note: Area under 0% indicates shortage

Source: Dataquest

DRAM Shortage Will Continue

In Figure 4 w e see the difference between the supply of DRAMs and the demand based on the electronic equipment production forecasts used to generate Figure 3. This undersupply is expressed in percentage of bits, where a number less than zero indicates an undersupply of bits shipped into the market.

As long as there exists an undersupply, DRAM prices wiU either hold or go u p . Despite a recent increase in our estimates for 4Mb DRAM production in 1995, Dataquest expects the shortage to continue through 1996, mostly because of the late start of the l M x l 6 ramp-up.

Although the shortage will lessen, from nearly 20 percent in 1994 to less tiian 5 percent in 1996, it will stm be a shortage, and prices cannot be expected to fall dramatically.

So w h a t about all that added capacity, and, more importantly, why do some forecasters predict

DRAM price plunges while Dataquest holds that the undersupply will continue? This misunderstanding stems from a commonly held belief that all convertible capacity will immediately be p u t into volume production of 16Mb DRAMs. This alone would at least satisfy

DRAM demand, if not overwhelm it. Instead, a large portion of the convertible capacity is still being used to produce 4Mb DRAMs.

This begs the question: "Why don't DRAM suppliers convert aU convertible capacity quickly to

16Mb devices?" The following is our explanation.

PCs account for more than two-thirds of all DRAM consumption. There is a phenomenon called "granularity," which forces PCs to diverge from using traditional organizations of

DRAMs of xl and x4. Instead, today the great majority of 16Mb density DRAM d e m a n d is for

March 29, 1995 ©1995 Dataquest Incorporated

Dataquest Alert Memories Worldwide a 16-bit-wide organization. There are two reasons that the xl6 version of the 16Mb DRAM is not available today in the volumes needed by the PC market:

• First: DRAM meinufacturers waited to introduce the xl6 version of the 16Mb DRAM until they h a d met reasonable production yields on the more traditionally accepted xl a n d x4 versions.

• Second: Design, debug, and test of the xl6 organization present challenges never before encountered by DRAM designers and manufacturers.

The result of this is that, despite the timely ramp-up of the 16Mb DRAM, the r a m p - u p of the l M x l 6 version is about 18 months behind the market. The need for this particular part has h a d to be filled by the lMx4 version of the 4Mb density, and four times as m a n y of these parts are required to m a k e u p the difference.

The result is that, until the xl6 organization of the 16Mb DRAM ramps into high-volume production, there will be a severe shortage, and an overwhelming consumption of 4Mb

DRAMs to account for the difference.

We see strong price-ups in the spot market. However, this has not been the case in the contract market, mostly owing to the close business relationships most DRAM manufacturers try to maintain with their clients. Japanese DRAM manufacturers have pointed out to us that, while they have held the ASP for a 4Mb DRAM at 1,200 yen since mid-1992, they h a d every opportunity to raise prices. Their reluctance to raise the price to "what the market wiU bear" shows restraint in an effort to continue to satisfy their customers' needs. As I noted at the outset, we continue to carefully watch for any change in DRAM contract pricing, especially in

North America.

Conclusions

In conclusion, although Dataquest has observed a very strong response to the current DRAM shortage in the form of plant expansions, w e do not expect this n e w and existing capacity to be able to match demand through 1996. The result should be continued allocation, high spotmarket prices, and the continuance of DRAM contract-volume prices to be keyed to a fixed yen value.

(This concluded the opening statement.)

Clark Fuhs, Nader Pakdaman, and Calvin Chang of the Semiconductor Equipment, Manufacturing, and

Materials Worldwide service (SEMM)

Mark Giudici and Scott Hudson of the Semiconductor Procurement Worldwide service (SPSG)

Jim Handy and Ron Bohn of the Memories Worldwide service

March 29, 1995 ©1995 Dataquest Incorporated 8

Oataquest^ ^ ^ ^ y

1290 Ridder Park Drive • San Jose • CA • 95131-2398 • Phone 408-437-8000 • Fax 408-437-0292

Earthquake in Japan: Semiconductor Producers and Fabs

Survive witli IVIinimal impact... but Long-Term Effects Wili Be

Felt Because of the liuman Factor

Once again, Japanese industry is shaken by a major earthquake. The Dataquest Japanese and

North American Semiconductor groups have compiled a detailed summary of how the semiconductor industry has been impacted. It has been difficult to get information quickly, as telephone lines are controlled in the area from Osaka to Kobe, and it is impossible to reach anyone in the area by phone. Kyoto became accessible only yesterday.

Companies are assessing the damages they had to their facilities, but it is becoming clear that the most difficult parts of this event are the human impacts and their influence. Matsushita has dispatched several groups of people to check on about 50 employees that have not reported to the company yet, and Mitsubishi reports missing employees as well. In general there was no serious damage to semiconductor wafer fabs, as described later in this Dataquest Alert, and most facilities that went down should be up within 10 days. But we do suspect that back-end test and assembly facilities might have been more impacted.

Also, it is not yet clear how serious the distribution infrastructure in the area is damaged:

Water, electricity, and transportation could still delay recovery process. It is estimated that the port of Kobe handles about 20 percent of all of Japan's trade, and that port is not expected to be operational for a very long time.

Summary of the Affected Region

Figure 1 shows a representation of the area affected by the earthquake, along with approximate locations and sizes of the prefectures, or states, within Japan of interest to the semiconductor industry. The Hyogo Prefecture, where the epicenter was located, was of course hardest hit.

The rectangle represents the areas most affected as the fault line runs from the islands outside of Kobe to Lake Biwa in Shiga Prefecture.

The affected area has a relatively low density of major semiconductor facilities, compared with areas just east of the region. Nara and Mie prefectures, where a high concentration of Sharp and Fujitsu plants are located, were basically unaffected, both in semiconductor and flat panel display (FPD) production.

January 19,1995 ©1994 Dataquest Incorporated

Dataquest Alert

Semiconductor Equipment, ManutacturJng, and Materials Worldwide

Figure 1

Geographic Area in Japan Affected by the 1995 Quake

Source: Dataquest (January 1995)

Review of Semiconductor IVIanufacturers

Mitsubishi Electric

Kita Itami Works, Hyogo Prefecture, was the closest fab to the epicenter. Our contact in Tokyo happened to be on his way to the plant the morning the earthquake happened. He did reach the site, but was told to leave immediately. He could not spot any damage on the outside of the buildings, but he is not sure how the inside is damaged. Mitsubishi did release a statement today saying the plant was shut down for safety and damage assessment, with no indication of when operations could begin again. Water and electricity supply to the plant also is questioned.

January 19,1995

©1994 Dataquest Incorporated

Dataquest Alert Semiconductor Equipment. Manufacturing, and Materials Worldwide

This plant is more of an R&D and pilot fab, dedicated to development of next-generation devices like the 64Mb DRAM. Therefore any damage this plant had should not affect

Mitsubishi's volume production.

There is some concern about Mitsubishi's assembly and test facilities in nearby Kanebo, where discretes and ICs are assembled. Although no firm reports are available, we suspect that these facilities have been affected.

KTI Semiconductor (Kobe Steel-Texas Instruments)

No damage was reported here, and Texas Instrvtments should be commended for quick action on the news wire services. The restart of its operation is not confirmed today, but estimating from the distance and general damages in the neighborhood (very little), it will not be long—in fact, a matter of a few days—^before the equipment is requalified and operation starts again.

This was the only suspect for possible decrease in DRAM supply caused by the earthquake:

Now that it is safe, we could say there is almost no negative influence to DRAM supply in terms of fab capacity.

Matsushita Electronics (Semiconductor)

Matsushita has a fab in Osaka Prefecture, in Kadoma City, which is an R&D and pilot facility similar in scope to the Mitsubishi Kita Itami plant. The company admits that its buildings are damaged: Ceilings have fallen partially, and cracks are seen in the walls. But heaviest damages are seen in floors higher than the third, and the main office building has more visual damage than the fabs. No details of the impact to the fabs are known yet, but the company has stopped operation for the past two days. Recovery of operation was not determined as of Wednesday afternoon. Its newest plant in Tonami, Toyama Prefecture, which started operation in October, was far enough away and was not affected at all.

NEC

NEC has a plant with four fab lines in Otsu City in Shiga Prefecture, with main production in power, linear, and bipolar devices at design rules ranging from 0.8 to 2.0 microns. It reports that its diffusion equipment has been damaged. Some sources hold that it will take about four ' % weeks before NEC can resume production, while others say the damage was not serious and

-'•T-

the plant is preparing for operation already. (Resist suppliers are being told to resume delivery to the Otsu plant.) We confirmed that the plants are in operation today.

IBM

IBM's Yasu plant, in Shiga Prefecture, had no serious damage to its production facilities, but we heard that the 8mm-thick "glass-wall" (large window) df its cafeteria was shattered, so we

January 19,1995 ©1994 Dataquest Incorporated

Dataquest Alert Semiconductor Equipment, Manufacturing, and Materials Worldwide

suspect equipment could have been moved arovmd. As with the KTI fab, we suspect that a requalification process is under way and production could start again within a week or so.

Toshiba

Toshiba has two plants in the Himeji area, about 150 to 200 miles west of Kobe: Himeji Works producing color picture tubes, LCD, and electroruc parts; and Himeji Semiconductor Works producing transistors and diodes at 0.8- to 1.5-micron technology. The semiconductor p k n t was not heavily damaged, but operation was suspended temporarily to check on equipment that may have moved. TFT-LCD operation is in fact the joint venture with IBM, Display

Technology Inc., where cracks in the ceiling are reported. We understand that this plant is back in operation today.

Toshiba's Yokkaichi plant, in Mie Prefecture, the major supply base for the 4Mb and 16Mb

DRAM, was not affected at all.

Rohm

Rohm's headquarters and its main fab, located in central Kyoto City, seem to be safe from heavy damages. Rohm has three major lines that support and produce semiconductors for several fabless companies in the United States. We believe that, after a requalification process, these lines will be operational.

Ricoh

Ricoh has two fab lines in Osaka Prefecture, and is also a supply foundry to several U.S. companies in 0.8- to 1.5-micron design rule devices. We have not yet been able to contact these facilities to assess the damage, but we suspect that these plants may have been damaged more seriously than average, on the order of the Matsushita facilities nearby.

Sharp

Sharp has confirmed that it had no damage in either its Tenri plant in Nara Prefecture or

Fukuyama plant in Hiroshima Prefecture (being far enough from the epicenter, about four times the distance away to the west). Fukuyama is the plant that supplies flash memory to

Intel. Both plants are in operation.

•i

It ff

January 19,1995 ©1994 Dataquest Incorporated 4 t

Dataquest Alert Semiconductor Equipment. Manufacturing, and Materials Worldwide

Down the Food Chain: Silicon Wafer Suppliers

Sumitomo Sitix

Sumitomo Sitix has its headquarters and a major plant in Amagasaki City in the Hyogo

Prefecture, in an area hit heavily by the quake. Its operations have been suspended. The company is still assessing the damage, and no plan for recovery is reported yet. This plant primarily produced 4-, 5-, and 6-inch wafers as well as ingots for all wafer sizes, including 8inch. These ingots were then shipped to Sumitomo's plant in Imari, which was imaffected by the quake, where cutting and polishing operations are performed.

It is estimated that 3 to 4 percent of the world's silicon area goes through the Amagasaki plant.

Before the earthquake, the world supply of 4-inch and 8-inch wafers was tight (the more serious is at 4-inch), while 5-inch and 6-inch supply is adequate. We believe that the end markets will experience some additional tightness in the coming months.

The Amagasaki plant also produced polysilicon, the raw material feeding the silicon wafer manufacturing industry. We estimate that the plant produced about 5 percent of the world's supply. Polysilicon capacity worldwide should be able to pick up this slack until the factory comes back online, although some immediate tight supply may be experienced as the industry readjusts to the short-term loss of the plant.

Some may be asking the question, "Is this the Sumitomo resin shortage revisited?" The answer

is, absolutely not. The incident 18 months ago involved a plant representing 60 percent of the world's supply of a material (compared with 5 percent today) and was only one of two major suppliers (whereas there are six to eight suppliers of polysilicon). This situation can in no way be compared in scope to the resin issue of the past.

Mitsubishi IViateriai Silicon

No damage is reported in Mitsubishi Material's plant in Ikuno, which is far from the epicenter.

Other Electronic Plants

IVIatsushita (Nonsemiconductor Piants)

Matsushita's headquarters is in the heart of Osaka City, but apparently it was not damaged.

Visual damage was seen in its computer and commimications plant in Kadoma City, where shattered windows were observed by people who passed the area by train.

January 19,1995 ©1994 Dataquest Incorporated

Dataquest Alert Semiconductor Equipment, Manufacturing, and Materials Worldwide

Sanyo

Sanyo has a battery plant in Awaji Island, the small island near the epicenter, but no serious damage was reported. Current concern at the plant is centered around the safety of employees.

Some of the subcontractor suppliers in the area may have been damaged: Contrary to the sturdy buildings of the parent company, these smaller compemies are assumed to have lessquake-resistant structures.

Hosiden

There are two opposite rumors about Hosiden's small LCD operations in Osaka: One holds that there was heavy damage, and the other denies any serious impact. The company itself is denying any substantial damage, but as yet we do not have confirmation.

Dataquest Perspective

This tragic event has apparently spared the semiconductor industry to a great extent. No major shortages are expected to emerge, although slight tightness in some silicon wafer areas are probable.

Damage to the Japanese economy with the loss of the Kobe port is hard to speculate, but there should be an impact. The damage to buildings and highways does not match the assumptions for all the regulations and formulas for quake-resistant construction. Architects and city planning people are losing confidence and are afraid that thorough investigation of the damages will result in a complete review of construction standards. This will place an additional burden on the confidence and morale of the Japanese people for a time, but in the end the true Japanese spirit will prevail.

The most notable cost here is the loss of human talent. Inevitably, when the history books are written, this event will be remembered for friends and colleagues that are no longer with us, . and for the hearts they touched.

By Clark J. Fuhs, Kunio Achiwa, Yoshihiro Shimada, Naotoshi Yasuhara, Akira Minamikawa, Yoshie

Shima, and Junko Matsubara

January 19,1995 ©1994 Dataquest Incorporated

DataQuest

^ '

FILE COP

0 Not Remov

Perspective

Semiconductor Equipment^ Manufacturing, a n d

Materials W o r l d w i d e

Telebrief ing Summary

The 1995-1996 DRAM Fab Outlook-Can PCs Absorb That

Much DRAM Capacity?

AbStrSCt: Despite rampant addition of new wafer fabrication capacity, there remains a

DRAM shortage that has been with us since 1992. Datatjuest predicts the shortage to continue into the second half of 1996. In a recent telebriefing the transcript of which makes up this newsletter, key Dataquest analysts explained why massive capital investment has not caused a supply/demand balance, or even the oversupply expected by other market watchers.

By Clark Fuhs

Introduction

This is Jim Haridy of Dataquest's Memories Worldwide service. First, let me introduce the speakers, then I'U give an overview of the subject and structure of the telebrieftng. In the room here at Dataquest w e have Clark Fuhs,

Nader Pakdaman, and Calvin Chang of the Semiconductor Equipment,

Manufacturing, and Materials Worldwide service (SEMM); Mark Giudici and Scott Hudson from the Semiconductor Procurement Worldwide service; and Mario Morales from our Research Operations Group. Ron Bohn and I represent the Memories Worldwide service.

What is this telebriefing about?

• Dataquest performs an analysis of DRAM supply and demand and publishes the results in a report updated quarterly.

• There has been a big increase in capital spending for semiconductor processing equipment. Despite all the new equipment coming online, the

DRAM shortage will persist through 1996.

• This teleconference wiU describe the reasons that the DRAM shortage will continue, despite the installation of so much semiconductor processing equipment.

Dataquest

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product CodB:SEMM-WW-BR-9501

Publication Date: June 19,1995

Filing: Perspective

Semiconductor Equipment, Manufacturing, and Materials Worldwide

First, Clark Fuhs wiU speak about semiconductor Weifer fabrication plant capacity, then 1 will return with a statement about the DRAM market.

After these statements, w e will open the phones to questions frona the participants.

DRAM Supply-Side

I am Clark Fuhs, a senior analyst from the SEMM group. We track most aspects of the actual manufacturing of semiconductors worldwide. Today we will briefly review the dynamic arena of DRAM capacity and supply.

The key source for our analysis today is a report Dataquest has just issued on 4Mb and 16Mb DRAM supply and demand. We titled it "DRAM Supply and Demand Report."

Supply-Side Analytical Methodologies

There are two supply-side methodologies, and their differences are outlined in Figure 1. Shown are three snapshots in time—at the ends of 1994,

1995, and 1996. The left bar in each year represents results of a survey of

DRAM suppliers, which will be referred to by Jim Handy later. The right bar represents an estimate based on a fab-by-fab analysis of capacity and committed plans.

Figure 1

Estimated Fab M o n t h l y R u n Rates for 4 M b and 1 6 M b D R A M s

Bits per Month (Trillions)

1,800 IT

Estimated Capacity—DRAM Supplier Survey

1,600

Estimated Capacity—Fab-by-Fab Analysis

1,400

1,200-

1,000

800

600

400

200

0

End of 1994

End of 1995

Source: Dataquest (June 1995)

End of 1996

9S0E597

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated

June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

The capacity estimate through the fab analysis includes the following assumptions: die size estimates, the effects of shrinks, and a gradual yield increase from today through the end of 1996 factored in on a company-bycompany basis.

We have not included any capacity associated with unannounced capital spending increases or commitments over the next two years such as the recently announced acceleration of NEC's new U.S. facility—nor 6- to

8-inch wafer conversions.

As seen in Figure 1, these two methodologies have produced results within

3 percent of each other for the year just completed. But as we go into the future, the figure shows an increasing divergence. The conclusion here is that suppliers are not optimistic in the near term regarding their abiUty to increase yields on the 1x16 configuration for the 16Mb DRAM, and their outlook is conservatively hedged.

This leads to a question we get fairly often: What is the capacity of the market in 4Mb "equivalent units"? Unfortunately, from a fab perspective, this is not a very useful way to view capacity — and is a metric that we at

Dataquest view may not provide an adequate picture of bit capacity. Why not? In a fab, capital spending and equipment purchases are driven by the requirement for wafers, or better yet, square inches of silicon.

Three Stages of Capital Spending

Let me highlight here the dynamics of how capital spending and bit supply are related. There are three identifiable parts to this cycle, and as you will see, we are in the later stages of part 2.

In the early to middle stages of a unit ramp in a specific DRAM density generation, square-inch requirements and equipment purchases are generally driven by bit demand. This is part 1 of the investment cycle, and it occurred in 1993 for the 4Mb generation.

In the later stages of ramp, as the next density generation starts to become available, we enter part 2 of the investment cycle. During this state, capital investment is still primarily driven by bit demand. However, the style of investment changes to install "convertible" capacity —in today's case equipment earmarked for 16Mb DRAM capacity, but initially rxmning 4Mb parts.

Most investments in 1994 into the present have been of this nature. The later stages of part 2, which we are in today, wiU also tend to include dedicated capacity for the next density generation.

So today there are three types of capacity w e must consider —capacity dedicated to 4Mb DRAMs, capacity dedicated to 16Mb DRAMs, and capacity that is "convertible" between the two. Why do suppliers like "convertible" capacity?

Based on die size ratios and the bit density ratio, as well as some other minor factors, a suppHer can generally increase the bit capacity of a line by converting the line from 4Mb to 16Mb parts, with minimal incremental capital or equipment spending. This is accompUshed because the bits per square inch are increased on the order of two or three times, meeining a supplier can double or triple the bit capacity of a hne by conversion. Thus

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

part 3 of the investment cycle, typically lasting two years, creates bit capacity primarily by conversion rather than new equipment purchases. Thus a

"pause" in the equipment market ensues. When all the capacity is converted, we begin part 1 of the cycle all over again.

Industry's DRAM Capacity

Back to the original question: What is the DRAM capacity of the industry?

The answer is —it depends. It depends on how the "convertible" capacity is employed. The most useful way to view capacity is by a "window" and this is shown in Figure 2, using a fab-by-fab analysis. What I mean by "window" here is the area between the two lines that represent the possible capacity measured in terms of bits. Let's first relate this to Figure 1.

Figure 2

Monthly DRAM Bit Capacity "Window" for 4Mb and 16Mb

Bits per Month (Trillions) if.UUU -

1,800-

1,600-

1,400-

1,200-

1,000-

800-i

6 0 0 -

4 0 0 -

2 0 0 -

0 -

End of 1994

"Convertible" at

100% 16Mb

1

End of 1995

"Convertible" at

100% 4Mb

End of 1996

16Mb

Equivalent

4Mb

Equivalent

9S0ZS93

Source: Dataquest (June 1995)

At the end of 1994, Figure 1 shows that the monthly rxm rate of the industry was slightly more than 600 trilUon bits. If all the "convertible" capacity were running 4Mb DRAMs at the end of 1994, the rxm rate would calculate to about 550 trillion bits per month, shown as the lower limit Line in Figure 2.

Likewise, if all of this convertible capacity were running 16Mb parts at today's yields, the capacity would calculate to slightly less than 800 trillion bits per month. Employing this methodology into the future, we have produced Figure 2 as a range of capacity over time.

SEMM-WW-BR-9501

©1995 Dataquest Incorporated

June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

1 would like to emphasize that the width of the window increases in 1995 and shrinks in 1996. Because of the low yield and slow ramp issues of the

1x16 configuration part, suppliers are being forced to add more 4Mb capacity of a "convertible" nature today. This will increase the growth for the front-end equipment market during 1995 well beyond our current published forecast of 16 percent —probably not far away from 30 percent growth. Momentum factors will establish 1996 as a small growth year as well (originally w e indicated a slight decline).

Our model indicates that the "pause" in the equipment market that we have been forecasting is unavoidable, however, because installing "convertible" capacity today by definition installs future "hidden" bit capacity and will likely cause a slight decline for wafer fab equipment in 1997. This conversion stage will also reduce the growth in the consumption of sihcon in a Uke manner.

The trigger for the pause will be the availabihty and subsequent pull of the

1x16 configuration part into the end-use market. Until this happens, silicon square inch demand wiU continue to be driven by bit demand closely. The severity and length of the ultimate "pause" will be determined by how long the current boom lasts, basically building pent-up bit supply as well as the demand for bits during the conversion stage.

Now I'd hke to turn it back over to Jim Handy for a review of demand.

DRAM Demand-Side

This is Jim Handy. Today's DRAM market is phenomenal. There has been an undersupply for two-and-a-half years, and in our recent analysis, the

February 1995 version of the quarterly "DRAM Supply and Demand

Report," we have found little reason to expect the current shortage to ease through the end of the report's forecast window, the end of 1996.

4Mb DRAM prices have slowly risen since the third quarter of 1992, yet

DRAM use in megabytes per PC has gone u p .

DRAM Pricing Alert

In the face of a strengthening yen against the dollar, Japanese suppliers of

DRAMs have been able to raise average selling prices measured in dollars to stabilize the yen value of their worldwide DRAM sales. By aU appearances, this trend will continue through the end of next year.

Our current North American contract-volume DRAM price forecast calls for firm pricing. We are aware that some Japan-based DRAM suppliers right now are considering an increase in the contract pricing for DRAMs.

Korean and North American DRAM suppHers are waiting in the wings to see what happens. Should these changes materialize during second-quarter price negotiations, we wiU make appropriate changes in our price forecast.

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Let's go over some of the background causes of the current DRAM shortage:

• PC demand is strong and shows no signs of letting u p .

• Japanese DRAM supphers, which accounted for nearly 50 percent of worldwide sales, were slow to react to the market and didn't increase capital spending until the fourth quarter of 1993.

• PCs have increased their per-system consumption of DRAM despite these stable/rising prices both in response to the requirements of advanced software as well as to decreases in the prices of CPUs, chipsets, and other system components.

• The 16Mb DRAM has met with limited acceptance in the traditional xl and x4 organizations. A new xl6 organization is the product of preference, and is late to market, forcing an increase in the consumption of 4Mb

DRAMs unlike that seen in any other DRAM generation.

The PC market has stayed in a strong growth phase for some time now.

However, the cause of this strength has varied over time. Two years ago growth was fueled by a rebound in PC shipments to the office. More recently, a big growth in home computer purchases including multimedia systems has driven the market. Now the business channel once again stands poised for important growth.

PC Forecast Revised Upward

We want to highlight today that Dataquest has just revised its worldwide

PC unit-shipment forecast upward. Dataquest's PC shipment forecast caUs for 57 million units to ship in 1995, and in excess of 65 million tmits in 1996.

This means a strong 19 percent growth in worldwide PC unit shipments this year and em equally impressive 18 percent growth for 1996.

Within each of these PCs we see another jump in the amount of DRAM consumed. The minimum DRAM size will increase from the 4MB per PC that shipped in 486-based PCs in 1994 u p to 6MB when the 8MB Pentium minim u m is averaged with the 4MB 486 this year.

Now let's look at Figures 3 and 4. Figure 3 shows the annual bit demand of the total data processing market and other DRAM markets. We derive this number from our electronic equipment forecast, which includes PC and other data processing forecasts as well as forecasts for other DRAM appHcations such as printers, fax machines, video games, hard disk drives, and even digital set-top boxes. This is the demand side of our supply demand analysis, and is driven to a great extent by the health of the PC market.

DRAM Shortage Will Continue

In Figure 4 we see the difference between the supply of DRAMs and the

demand based on the electronic equipment production forecasts used to generate Figure 3. This undersupply is expressed in percentage of bits, where a number less than zero indicates an undersupply of bits shipped into the market.

As long as there exists an vmdersupply, DRAM prices will either hold or go up. Despite a recent increase in our estimates for 4Mb DRAM production in

1995, Dataquest expects the shortage to continue through 1996, mostly

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Figures

DRAM Demand in Bits

Number of Bits (Trillions)

1994

1995 1996

Source: Dataquest (June 1995)

Figure 4

DRAM Suppl)//Demand Scenario, 1995 to 1996 (Supply as

Percentage of Demand)

SEMM-WW-BR-9501

Note: Area under 0% indicates shortage.

Source: Dataquest (June 1995)

©1995 Dataquest Incorporated

June 19,1995

Semiconductor Equipment, IVIanufacturing, and Materials Worldwide because of the late start of the l M x l 6 ramp-up. Although the shortage will lessen, from nearly 20 percent in 1994 to less than 5 percent in 1996, it will still be a shortage, and prices cannot be expected to fall dramatically.

So what about all that added capacity, and, more importantly, why do some forecasters predict DRAM price plunges while Dataquest holds that the undersupply will continue? This misunderstanding stems from a commonly held behef that all convertible capacity will immediately be put into volume production of 16Mb DRAMs. This alone would at least satisfy

DRAM demand, if not overwhelm it. Instead, a large portion of the convertible capacity is still being used to produce 4Mb DRAMs.

This begs the question: "Why don't DRAM suppliers convert all convertible capacity quickly to 16Mb devices?" The following is our explanation.

PCs account for more than two-thirds of all DRAM consumption. There is a phenomenon called "granularity," which forces PCs to diverge from using traditional organizations of DRAMs of xl and x4. Instead, today the great majority of 16Mb density DRAM demand is for a 16-bit-wide organization.

There are two reasons that the xl6 version of the 16Mb DRAM is not available today in the volumes needed by the PC market:

• First: DRAM manufacturers waited to introduce the xl6 version of the

16Mb DRAM until they had met reasonable production yields on the more traditionally accepted xl and x4 versions.

• Second: Design, debug, and test of the xl6 organization present challenges never before encountered by DRAM designers and manufacturers.

The result of this is that, despite the timely ramp-up of the 16Mb DRAM, the ramp-up of the l M x l 6 version is about 18 months behind the market.

The need for this particular part has had to be filled by the lMx4 version of the 4Mb density, and four times as many of these parts are required to make up the dijfference.

The result is that, imtil the xl6 organization of the 16Mb DRAM ramps into high-volume production, there will be a severe shortage, and an overwhelming consumption of 4Mb DRAMs to account for the difference.

We see strong price-ups in the spot market. However, this has not been the case in the contract market, mostly owing to the close business relationships most DRAM manufacturers try to maintain with their clients.

Japanese DRAM manufacturers have pointed out to us that, while they have held the ASP for a 4Mb DRAM at 1,200 yen since mid-1992, they had every opportunity to raise prices. Their reluctance to raise the price to

"what the market will bear" shows restraint in an effort to continue to satisfy their customers' needs. As I noted at the outset, we continue to carefully watch for any change in DRAM contract pricing, especially in

North America.

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, IVIanufacturing, and IVIaterials Worldwide

Dataquest Perspective

In conclusion, although Dataquest has observed a very strong response to the current DRAM shortage in the form of plant expansions, w e do not expect this new and existing capacity to be able to match demand through

1996. The result should be continued allocation, high spot-market prices, and the continuance of DRAM contract-volume prices to be keyed to a fixed yen value.

Transcript of Telebrief ing

At this point the introductory statements ended and the forum was opened for questions. All callers remained anonymous.

Q: Can you tell me what your old PC and Pentium unit forecasts were, and can you give us a look at DRAM pricing? Where do you see the 4Mb this year during the various quarters and into next year?

Ron Bohn (RB): Our prior PC forecast worldwide for this year was approximately 54 million units; the Pentium forecast for this year was originally just over 20 million vmits. Our original worldwide PC shipments forecast for next year was in the low 60 millions. I do not know offhand what w e said for Pentium for next year. On pricing we have the contract pricing for the 4Mb. We basically showed pricing above $12, but we had anticipated perhaps going a little bit under $12 for the end of this year, and that is one of the forecasts we are looking at right now. Maybe my associate can give us a Httle indication of what the 16Mb pricing is. I do not have that offhand.

Mark Giudici (MG): In the current quarter, volume pricing for the l M b x l 6 device is approximately $54 in the North American marketplace. With the current forecast w e see it going down sUghtly to under $45 by the end of the year. However, as Ron was saying, this forecast wiU probably be revised upward to some extent as the supply and demand work through the marketplace. But this is the current forecast.

Q: So, do you expect 4Mb DRAM pricing to be over $12 by the end of this year? And in terms of spot market pricing, where do you see that in contrast to the contract pricing?

MG: Spot market prices definitely will be higher. As far as forecasting the spot market, that's like trying to find out which way the wind is blowing.

Q: How much of a premium do you see right about now?

MG: The premium is that the 4Mb level, the lMbx4 part, for example, ranges from between $15 on the low end to as high as $20 from what w e have heard, so you can see there is quite a bit of variance between contract and spot. I have not gotten any spot price points on the l M b x l 6 . 1 think the majority of those parts are going to contract users. The forecast for spot is difficult to call.

Jim Handy (JH): One thing you should keep in mind is that NEC and

Toshiba in Japan announced a 5 percent price increase to make u p for yen-to-dollar exchange ratios.

SEI\/1M-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

10 Semiconductor Equipment, IVIanufacturing, and Materials Worldwide

Q: For the various configurations of 16Mb DRAMS x 4s, x 8s, xl6s, how m a n y dice fit on the 8-uich wafer, and what are the yields for the various types at this time?

Clark Ftihs (CF): I can tell you generally the 4x4 and 1x16 average die size at present and a yield range that w e are hearing from the marketplace. The

4x4 average is about 100 sq. mm., and the 1x16 is about 5 percent to

10 percent larger than that. We have heard of yields on the 4x4 of as high as

75 percent, but we estimate cin average in the 70 percent region. We have heard a fairly wide range for the xl6, anywhere between 40 percent and the low 50s. We are pegging basically about a 45 percent to 50 percent yield on those.

Q: What is your estimate for 1995 for PC unit shipments?

JH: The 1995 total PC unit shipment number is 57 rrulHon units. This includes all levels of PCs, from deskside through transportable.

Q: And how many Pentiums? Twenty-five?

JH: More than 25.

CF: I just remembered a number on the number of dice on the 6-inch wafer.

You could fit about 150 of the 4x4-16Mb DRAMS on it; for the 8-inch wafer you multiply that by about 1.92.

Q: I have a couple of questions that relate to the 1996 analysis that you have ^ ^ here. First of all, I would like to look at the supply-side and demand-side ^ H question. On the supply side, in 1996 there seems to be a big differential ^ ^ between what you are estimating and what the suppliers are estimating, and 1 would like you to comment on that a little more. There are a lot of new suppliers coming onstream, and I'm wondering if their strong suit is in new fabs or new designs. Is it possible that the supply side that you're estimating may be a Httle high, that maybe the difficulty in bringing on new fabs has not been fuUy factored in?

On the demand side, the amount of memory per PC you are using is 4MB now, and 6MB in 1995. This seems to be about half of what I hear from some actual users. It seems to me that, if you're assuming 6MB per PC in

1995, your analysis may be a little light on the demand side, leading us to question a supply/ demand balance by the end of 1996. What supply figures are you using in your analysis of 2 percent undersupply in 1996? Is that the estimate you have generated, or is that the estimate of supply that came from the manufacturers?

RB: First of all, let me talk about the demand-side question. We are talking in ternis of megabytes per system. We were just talking about the PC/OEM level out the door. In terms of a full-upgrade scenario, w e have a much higher number. The main thing w e do not w a n t people to do is to apply this number against our PC forecast, because that would overstate the demand side.

#

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide 11

In terms of an upgrade scenario for PCs after they leave the PC/OEM, for

1995 we show an upgrade in the ballpark of under 16MB, which is much more in line with what you were talking about. Going out to 1996 the number increases more, somewhere under 20MB. So again, what w e were highlighting was the megabytes per system at the P C / O E M level coming out the door. On the demand side, I just want to highhght one item. Figure 1 shows the forecast of what DRAM suppliers expect to be shipping. The higher bar that we alluded to shows the whole point of the telebriefing.

What if the supplier really pushed to use the convertible 16Mb capacity for the production of 16Mb devices? I would Uke to turn this question over to

Clark.

CF: We did a fab-by-fab analysis. A lot of the newer fabs coming online in

1996, some of which you alluded to, have been only mildly factored into our supply-side analysis. We have been rather conservative in how those would ramp u p . The production yield on 16Mb is tracked by company in our model. The average across all companies is about 50 percent or so today, increasing by the end of 1996 to the 65 percent to 70 percent range. If yields do not pick u p that much, then of course the window for the supply side would be pinched on the top end and the overall bars for the end of

1996 in Figure 1 would come closer together. As I described in my opening statement, we believe that the primary reason for the difference is that suppliers are not optimistic in the near term about how they're able to ramp yields. We took a realistic look at the market and have assumed ramp-up will take two years, and we would believe an increase to the 65 percent to

70 percent region would be a reasonable yield at full production. That is how we came u p with the data in Figure 1.

Q: Which set of the supply-side bars in Figure 1 did you use in the supplyand-demand calculation that drives Figure 4?

JH: Let me embellish on Ron's megabyte-per-system number. 1 used minim u m system configuration to highlight the effects the Pentium will have on the market. Last year the market was nearly 100 percent 486 and the absolute minimum configuration for 486 is 4MB. This year we are expecting nearly 50 percent of the market to be Pentium, and the absolute minimum configuration on the Pentium is 8MB, so the effect wiU be that the absolute minimum averaged out over all the systems to ship is going to be 6MB.

That is a 50 percent lurch in one year. That is pretty significant, especially in the face of flat pricing. The way that w e look at DRAM pricing, and the answer that Ron was giving you, is this: We look at the overall total DRAMs that go into the PC market, and we analyze this by the different levels of PC and therefore use different megabyte numbers for each level. We also take into accotint aU the upgrades that are inserted as "after equipment" into systems that have already left the store and have been in use for a while in the office or the home, and so w e do factor those in and w e do come u p with a much higher number. It's something I do not have here to share with you.

You're right when you say that the 4MB and 6MB numbers are about half of what you've seen from other places; in fact, the numbers are about half of what you'd see if you looked at our numbers for overall DRAMs shipped into the PC channel. I hope this answers your question.

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

12 Semiconductor Equipment, IVIanufacturing, and Materials Worldwide

Q: How big is the U.S. market for DRAM SIMMs, the actual modules, and how does this affect where the demand for the DRAM chips is coming from? Is it coming from the aftermarket suppliers or from the OEMs, or both?

JH: It's kind of funny. There are two sides to the SIMMs market, and w e are going to be doing a report in June that covers the market. This is one of those markets that fall in the cracks at Dataquest. In the past, the Semiconductor Group has said, "It's not a seiniconductor so it should be covered by the Systems Group"; unfortunately, the Systems Group says, "We should watch over things that have CPUs in them." The net result is that we could have done a better job of covering this market than we actually have done.

We are working to fix that right now, and that is what the June report is going to address.

Some argue that SIMMs are half of the entire DRAM market. Some beUeve that the SIMMs market is spUt u p pretty evenly between DRAM manufacturers who produce SIMMs and aftermarket SIMM manufacturers. There's a very different complexion between the two kinds of suppliers. The aftermarket manufacturers tend to buy on the spot market at highly inflated prices and are somehow able to recoup and make a profit. Meanwhile, the

DRAM manufacturers use internal transfer costs to account for the DRAMs that go onto the SIMM and they sell in contract volume almost exclusively to OEMs for the SIMMs that go into the PCs that ship from the OEMs to the dealer. So, in terms of sales channels, the DRAM manufacturers sell SIMMs to the OEMs. Once the bpxes that are made by the OEMs go to a sales channel — a computer store or something like that —any upgrades are pretty much handled by the exclusive SIMM manufacturer, which also handles upgrades made after the equipment is installed in the office or home.

Q: Do you think the aftermarket is growing faster in terms of its demand for

DRAM chips than the OEM side?

JH: We do not think so; however, there is something to watch that is kind of a hunch right now because we have only seen one possible cycle. We have seen the u p and the down and we are expecting it to happen again. SIMMs are very appealing where there is some ambiguity as to which density

DRAM gets you the cheapest 4MB, and right now there is no question about it. The 4Mb DRAM will give you the cheapest SIMM for a 4MB

SIMM; 16Mb is not quite there yet, and a 1Mb, if you can get one, is going to be a more expensive solution. Because of that, OEMs are laying down

DRAMs right on the motherboard. As we get to the point where 4Mb and

16Mb prices for the organization of parts that people want are tracking each other, and w e expect that to be a very long period this cycle, then we'll think that SIMMs will be more popular in the OEM market, and so the

SIMM market will blossom for OEMs.

As far as selling to the aftermarket, that's more difficult to say because an awful lot of that is the perception of need on the users' behalf. We've heard an argument that the advent of Windows 95 wiU reduce the main memory size required to support software, and there is technical merit to that argument, because in the virtual memory model that is supported by Windows

95, you do not need as much memory. But w e don't buy that. We think that Windows 95 could increase demand simply because there wiU be a

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and JVIaterlals Worldwide 13

commonly held perception, right or wrong, that you do need more memory for it. Depending on an awful lot of consumer-type things, the market could grow significantly or it could just grow at today's rate.

Q: What are your perspectives on the growth of the SRAM market, and how do you see EDOs impacting that growth?

JH: That's an interesting question. SRAMs are used for caches in PCs. Our belief is that caches are bought in PCs more for their perceived performance than their actual performance. As a result, things that have been very well positioned to make the SRAM cache obsolete have not sold very well in the past, and even three years ago. We do not know how long that is going to continue, how long caches are going to be popular in PCs, but we do not believe that EDO or synchronous DRAM or any other exotic DRAM architecture will be able to dislodge the commonly held notion that more cache is better.

Q: Could you comment on the current capacity of raw 8-inch silicon wafers? Also, is there going to be enough capacity to support all the planned fab increases in the next several years?

CF: Yes, w e did a wafer s u p p l y / d e m a n d analysis for raw 200mm watfers about nine months ago. We did a fab-by-fab analysis as well as a supplierby-supplier plant analysis and compared them, and we also compared them with a top-down analysis forecast. At that time the demand caught u p with and outstripped supply by the end of 1995. Since our study, silicon wafer manufacturers committed about $1 billion of new capital for the expansion of 200mm facilities. So nine months later the crossover point where demand would exceed supply, where we would actually be put into a shortage situation that we are not into today, was also pushed out about nine months. Our outlook is that the capital will continue to come as the demand comes, because 8-inch wafer pricing is stiU pretty firm and is actually edging u p a little bit. We are confident that the supplier base will be able to adequately respond to that need. Also, there is quite a bit of flexibility in the use of test and monitor wafers, which is a pretty high percentage of 200mm usage. That can be scaled down quite dramatically to make u p any shortages or spot shortages that could exist.

Going back further in the food chain, we've heard that there is some concern about the polysilicon market —the raw bulk silicon supply that the wafer meinufacturers buy from. In 1992 we actually tracked, on a rough basis, the utilization of those plants. We really do not cover that market in any more detail than that. In 1992 the capacity utilization was in the low to mid seventies, and it's probably in the mid eighties right now. There are two suppliers that have been expanding pretty aggressively over the last two years, and these two basically supply just under 50 percent of the market.

Again, our outlook is for fairly high utilizations going forward for these plants, but they will be just about balanced with a little bit more supply than demand. They will be able to stay ahead of the power curve. There is enough commitment to the investments, enough commitment to the market. We feel confident that the suppliers will be able to respond.

SEMI\/1-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

14 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Q: Looking at 1995,1996, and 1997, how much of the capital spending would support 16Mb/64Mb processing of 0.35 micron and better?

CF: I've got that —what mix of equipment is being shipped, what kind of technology that can support—but not with me. Suffice it to say the bulk of equipment shipped in 1994 was for 0.5 micron and the bulk that wiU ship in

1998 will be 0.35 micron production. As far as the exact numbers for the years in between, you would have to give m e a call later.

Q: It is no secret that the DRAM vendors themselves said that the xl6 configuration is the one the market wants. In your 1996 supply-and-demand analysis, what assumptions have you made about their abihty to improve the xl6 yields?

CF: I think I already stated that for 1994 w e basically used an average yield figure of about 50 percent or so; that includes a mix of 4x4 and 1x16. The outlook for the end of 1996 is that the average yield would be in the 65 percent to 70 percent range, and I think that's fairly representative of the assumptions we made for the xl6 yield ramp.

JH: Let me just add something to that. Those assumptions are based on the pale bars in Figure 1. You can see we are getting a little more conservative estimates from the DRAM manufacturers themselves, and that's represented by the lighter-colored bars. We do get a lot of input from DRAM manufacturers as to what they expect their capacity to do, and that is what that represents there. So we take aU this into consideration when looking at supply and demand.

Q: Could you please provide more detail on the 4Mb DRAM pricing, by quarter; what you had predicted and what your estimate is as you look again at that pricing?

RB: First of all, our price outlook: We are definitely going to hold off imtil we see what happens with the contractual negotiations going on right now.

Let me go over the 4Mb DRAM pubUshed price. This is for the first-year

North American buyers. I will just highlight the lMbx4. For the first quarter of 1995 we had it at $12.66, second quarter around $12.50, third quarter at

$12.25, and fourth quarter just around $12.00. If you look back six months ago, aU these numbers for the end of this year are somewhat higher, so each quarter we see more and more factors that play into the hands of DRAM suppUers. The Kobe quake at the beginning of the year had no real impact on supply and demand, but psychologically it helped the suppHers. And the yen-to-dollar exchange rate we are highlighting today definitely is a fundamental factor, and it causes Japan-based suppHers to start looking at their profit margin and their DRAM revenue. The Japan-based suppHers really use a ¥1,200 price for the 4Mb DRAM in their forecast, in their bank loan applications and other functional analyses. It's of fundamental importance to them.

Q: What is the premium for EDO and 3.3 volts?

JH: We have already heard from one supplier that is not charging any premiums for the EDO whatsoever. The way that EDO is viewed both by suppliers and by users is that it's a zero cost option, as it should be, because it's

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide 15

going to be widely supplied and it has near-zero die area penalty and so doesn't reaUy cost anything for the DRAM manufacturers to crank out. It's a very popular option because it gives added speed at no extra cost. The early initiators of 4Mb DRAM EDO today ask for a 5 percent premium, which should disappear sometime next year. Because of that, Dataquest is forecasting that EDO DRAMs will have virtually displaced the standard fast page-mode DRAM by the end of 1997.

As for the 3 volts, I have always thought this was an odd kind of thing. The reason we have 3 volts is that DRAM manufacturers couldn't see a way that they could get down to submicron geometries without asking the world to go down to 3 volts with them. Several years ago they took this to the standards committees and asked for things to change. Now that DRAM manufacturers have a 5-volt interface on a part that works 3 volts internally, which they did not think was a reasonable thing to do back when the standards were being created, they are shipping DRAMs that are pure

3-volt DRAMs at higher prices than they were charging for 5-volt DRAMs with a 3-volt core. That switchover probably will not happen until there is an oversupply, because there is an existing market. It is kind of a feat of daring to go out there and sell a 3-volt part for less than any of your competitors. And actually 3-volt-only operation offers some benefits to the user.

Because of that, DRAM manufacturers are taking advantage of the situation and are charging a premium. 1 would believe that the premium is somewhere around 15 percent, but Ron might be able to provide more information on that. We could talk offline if you need a precise number.

Q: You show a 12 percent shortfall in s u p p l y / d e m a n d . Is that evenly broken u p between 4Mb DRAM and 16Mb DRAM?

RB: We show a shortage for the 4Mb DRAM in 1995. Just over 10 percent for the 4Mb DRAM in 1995; 1996, if you go for the full year, may be almost balanced. 1 will get to that in a second. For the 16Mb DRAM, a 13 percent shortfall for 1995 and about a 5 percent shortfall for 1996. Looking at the

4Mb DRAM, on a quarter-by-quarter basis, in the first half of 1996 we show a shortage of 5 percent to 10 percent, so we are looking at the fourth quarter of 1996 to see whether there wiU be a s u p p l y / d e m a n d balance or a kind of excess-supply scenario. We definitely view that as a hedge factor and are focusing a lot on it. Basically, the 16Mb DRAM has a clear 5 percent shortfall in 1996; and in a very aggregate level, maybe the 4Mb DRAM is close to a supply/demand balance, if you roll in the fourth quarter of 1996.

Q: When do you think you will be able to republish your estimated prices on the 4Mb DRAM? My understanding is that those prices will go u p . Is that your feeling?

RB: The first thing w e will do in our DQ Monday Pricing will be to provide guidance on that. Our next published forecast, the one that I cited here, is officially scheduled to come out in the May time frame. But independent of this, the DRAM memory forecast is being revised in the near term, so maybe sometime during April we will have a little more guidance on that.

That is a little more on the worldwide perspective, but it should fit into all the scenarios that emerged based u p o n the contract negotiations.

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

16 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Q: You mentioned that the pricing was North America pricing. How does this stack u p regionally?

RB: In the pricing shown there, North American buyers are the preferred customers. This is, in most cases, the lowest pricing that we will really see.

There may be some spot-market differentials or preferred pricing within a transfer price, but basically this is on the low side.

Q: 1 have a question about the yields. You are looking at 40 percent to

50 percent for the 16Mb DRAM, and 65 percent to 70 percent in 1996. Do you have best-case and worst-case scenarios, and what would that imply to the equipment market demands? Is the boom in equipment spending this year in anticipation of the low end of the yield forecast or the high end?

CF: Well, the boom in equipment this year is more in relation to what kind of capacity people have to add right now. The 1x16 yields are not high enough at present or the price per bit on a cost level low enough to translate to the pricing. So it's actually more economical to make 4Mb DRAM right now, and that's just on a "per-bit" basis. That requires a little bit more silicon, and that is why capacity was added this year. There is going to be more added convertible capacity that is initially going to be running 4Mb

DRAM this year. That's basically the bottom line for 1995. As the 16Mb

DRAM 1x16 yields increase and get to a level that is cost-effective for the supplier, on a price-per-bit basis, that will be part of what triggers a pull of the 1x16 configuration part into the end-use market, and that should trigger the conversion of 4Mb DRAM capacity to 16Mb DRAM capacity. The dynamics would then cause the equipment market to pause, because bit capacity can be added without new equipment. So that's the course of events that will take place, as we see it.

If 16Mb yields remain low, and the market never converts, the equipment boom will never go away. If you take the other side to the ultimate limit and yields shoot u p to 80 percent tomorrow, you could see a pretty bad second-half 1995. We are basically watching the yield issues in the 1x16

DRAM. That's the key.

Q: Where is Korea on the yield curve here?

CF: Korea has basically been producing the 4x4 part. As of the second half of 1994 going into 1995, the two main suppliers of the 1x16 part were NEC and Toshiba. Ron and Jim might have some additional comments on that.

RB: The main comment on that is that Toshiba clearly is pushing ahead with the l M b x l 6 . We heard that Hyundai will be ramping up in the second quarter and that Samsung and NEC were also ahead of the pack. A lot of the other suppliers w e talked to about the 4Mb DRAM tell us that they are interested in extending the life cycle for the 4Mb DRAM density, that they wished they had ramped up the l M b x l 6 already, and that it is more and more tomorrow's story. So the l M b x l 6 ramp-up scenario is quite conservative, and we're not getting any reports that cause us to expect a big acceleration in the l M b x l 6 ramp-up in the near term.

SEMM-WW-BR-9501 ©1995 Dataquest incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide 17

CF: One additional comment: The 4x4 yields are high enough to warrant a lower price per bit. In fact, the 4Mbx4 DRAM can be purchased at a lower price per bit than the 4MB DRAM. The problem is that it is not the right part for the PC market and therefore demand is not being puUed, even with that lower price. That's why we have the scenario w e do today.

Q: In the demand side, has any consideration been given in the Latin

American situation?

JH: The PC Group is not represented in the teleconference right now, and someone from that group would probably be able to talk about that. In general, when the PC Group analyzes demand, it analyzes for existing markets more than hypothetical markets. Latin America is just starting to emerge— as soon will be India, China, Eastern Europe, and even Japan — and these are markets that we do not know an awful lot about but that look poised to increase PC consumption. If anything, I would tend to beheve that our PC Group is conservative about forecasting what's going into those areas. Fortunately w e are not predicting some glut of DRAMs, because if those markets open up, the shortage could be worse than it is now.

Analyst Bios

Ronald A. Bohn, Senior Industry Analyst, Memories Worldwide

Mr. Bohn is a Senior Industry Analyst for Dataquest's Semiconductor

Memories Worldwide service. He is responsible for research and analysis in semiconductor memory pricing, supplier, and product technology trends including DRAMs and flash ICs. His responsibiHty includes strategic planning, competitive analysis, and consulting projects. He works with securities companies, banks, and other members of the financial community on semiconductor trends and also tracks world trade, intellectual property, and related legal trends for their impact on the electronics industry. At

Dataquest he has forecast pricing of more than 100 semiconductor products. Mr. Bohn has written a series of reports on benchmarking and has assessed semiconductor life cycles from a component engineering perspective. This research served as a basis for Dataquest's PC "teardown" cost analysis. At Dataquest, he has also served as the analyst tracking semiconductor trends in the interactive CD-ROM player and PCMCIA markets.

Prior to joining Dataquest in the mid 1980s, Mr. Bohn assessed worldwide electronic markets on a macro- and microeconomic basis for a market research company. He served as International Market Research Manager for the Korea Trade Center in the United States and has financial, legal, and government experience. Mr. Bohn received a B.A. degree from Cornell University, an M.B.A. degree from the University of California at Berkeley, and a J.D. degree from the Hastings College of Law.

Clark J. Fuhs, Senior Industry Analyst, Semiconductor Equipment,

Manufacturing, and Materials Service

Mr. Fuhs is a Senior Industry Analyst for Dataquest's Semiconductor

Equipment, Manufacturing, and Materials service in the Semiconductors group. He is responsible for research and analysis of semiconductor materials and trends in IC manufacturing techniques along with forecasting capital spending and the wafer fab equipment market. Prior to joining

Dataquest, Mr. Fuhs was Strategic Marketing Manager for Genus Inc., a mantiiacturer of advanced chemical vapor deposition (CVD) and high

SEMM-WW-BR-9501 ©1995 Dataquest Incorporated June 19,1995

18 Semiconductor Equipment, Manufacturing, and Materials Worldwide energy ion implantation equipment. During his 10 years at Genus, he held positions of Product Manager, several responsibilities in Product Marketing, and Process Engineer in the metal CVD group. In his most recent position, Mr. Fuhs was responsible for correlating process techniques with demand for equipment and materials. He has been involved with the Modular Equipment Standards Committee of SEMI, a trade organization, as chairman of a task force, authoring a standard. His experience also includes

Chevron Oil, where he was a Process Engineer in the Richmond, California, refinery responsible for the hydrogen manufacturing plant. Mr. Fuhs earned a B.S. degree in Chemical Engineering from Purdue University in

West Lafayette, Indiana, and received an M.B.A. degree from the

University of California at Berkeley.

Mark Giudici, Director and Principal Analyst, Semiconductor

Procurement Service

Mr. Giudici is the Director and Principal Analyst of Dataquest's Semiconductor Procurement service. He is responsible for tracking and analyzing emerging semiconductor procurement issues and trends. He also covers regional semiconductor prices and cost modeling issues including product/supplier analysis on MPU and ASIC markets. In addition, he has participated in various custom research projects involving procurement needs, contract manufacturing, system teardown analysis, and regional price differentials. Prior to joining Dataquest, Mr. Giudici spent eight years in both the computer and semiconductor industries, where he held a variety of financial eind marketing positions. Most recently, he was a Product Marketing Engineer with American Microsystems, where he was responsible for cost modeling and marketing semicustom and foundry-custom semiconductor components. Mr. Giudici received his B.S. degree in Business

Administration from the California State University, Chico and his M.B.A. in Business Management from the University of Oregon.

Jim Handy, Director and Principal Analyst, Semiconductor

Memories Group

Mr. Handy is Director and Principal Analyst for Dataquest's Semiconductor

Memories group. He is responsible for the forecasting and analysis of memory products and markets. Previously, he was strategic marketing manager for static RAMs at Integrated Device Technology (IDT). Before IDT, he was product marketing manager of memory and microcomputer-based products at Intel Corporation, National Semiconductor Corporation, and

Siemens Corporation. Mr. Handy earned his M.B.A. degree at the

University of Phoenix and holds a B.S.E.E. degree from Georgia Tech. He is the author of "The Cache Memory Book" (Academic Press, 1993), he is a frequent speaker, and his work has been widely pubHshed in the trade press.

Clark Fuhs, Senior Industry Analyst (408) 437-8375

Internet address [email protected]

Via fax (408) 437-0292

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in

U 3 I 3 Q 1 J C S L part to other parties shall be made upon the written and express consent of Dataquest

, ©1995 Dataquest Incorporated — Reproduction Prohibited m m aoompeuiyot ^ '^ '

tmUm ThcOun&Biadsticct Corporation Dataquest is a registered trademark of A.C. Nielsen Company

t

Perspective

S e m i c o n d u c t o r E q u i p m e n t , Manufacturing, a n d Materials W o r l d w i d e

Dataquest Predicts

F I I _ E C O R V

The Cyclicality of the Equipment Market:

Do Not Remove

What Makes It Tick?

Abstr3Ct: We believe that the forces driving the cyclicality of the equipment marltet have

not changed, and that DRAM density crossover and looming price declines mean the equipment market should start a downturn within the next year.

The wafer fab equipment market has cycled through boom and bust times. With 1994 being an apparent boom year, several questions present themselves: Is there a looming downturn?

Why or why not? Are we really changing to a market that is more stable and not subject to the boom-to-bust cycles? In this article, Dataquest describes a new way to view the equipment market—perhaps a way that adds insight to the cyclicality of the business.

By Clark J. Fuhs

Setting tlie Stage

The year 1994 started great, and it just kept getting better— in somewhat scary proportions. Wafer fab equipment spending in 1994 expanded 50 percent worldwide, driven by an explosive Asia/ Pacific region (88 percent growth), a DRAM-sensitive Japan (45 percent growth), and continued heavy expenditure in North America (43 percent growth). This follows a

35 percent growth year worldwide in 1993. Although the outlook for semiconductors in general remains bright, there is cause for concern that the recent buying binge in wafer fab equipment of the last two years, to carry over into at least the first half of 1995, has been too much, and that a normal contraction of the equipment industry is unavoidable, albeit at a historically modest level.

Capital spending in 1994 exploded. The major reason for this is the surprisingly persistent growth in PC unit shipments, and a correction of the underinvestment that had been a part of semiconductor manufacturing since

OataQuest

Program:Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Cod8:SEMM-WW-PD-9501

Publication Data: February 20,1995

Filing: Perspective

Semiconductor Equipment, Manufacturing, and Materials Worldwide

1991. Major DRAM expansion accelerated in the second half of 1993 and is expected to continue through the first half of 1995. From what we can see now, there is plenty of equipment that could be brought to bear on 16Mb

DRAM capacity by midyear 1995 and online to answer demand through

1996. A marked downturn in the DRAM investment cycle will be triggered by the approach of DRAM price-per-bit crossover of the 16Mb (in the lMbxl6 configuration), making available some capacity as the 4Mb generation ramps down in the second half of 1995.

Wafer fab equipment spending in 1994 had peak growth of 50 percent worldwide (see Table 1) and is expected to grow 16 percent in 1995 (in a range of 15 to 22 percent worlwide) based on backlog and bookings momentum from the 1994 surge. Segment growth in 1994 into 1995 is being led by DRAM-sensitive equipment, with steppers, high-current and highvoltage implant, wafer inspection, and polysilicon etch exhibiting growth significantly stronger than the market. This mix will change to favor multilevel metal logic capacity as the year progresses. The upper end of the range could easily be reached if DRAM pricing remains firmer than we expect through the first half of 1995.

On a regional basis. North American investment will be driven by stable growth in microcomponents and logic devices, as semiconductors continue to pervade everyday life. Japan has concentrated on ramping memories to preserve its market share against the Koreans and was the second-fastestgrowing region in 1994, investing in next-generation memory capacity.

Healthy but subdued growth is anticipated in 1995, but a struggling economy and lack of a clear production strategy beyond the DRAM will keep capital investment muted once this DRAM ramp is satisfied. Globalization strategies will benefit both European and Asia/Pacific investment, the latter being the fastest-growing region of the next five years. The emergence of the dedicated foundry will create new opportunities in the Asia/Pacific region over the next several years. Projects being started in 1994 wiU continue to attract investment, but the last two years saw an investment spike that we beheve will be difficult to repeat in growth, but will be sustained as foundry projects are equipped.

Table 1

Worldwide Wafer Fab Equipment Market^ Forecast by Region

(Millions of U.S. Dollars, Calendar Year)

CAGR (%)

1993-1998

14.7 Total Wafer Fab Equipment

Percentage Change

North America

Percentage Change

Japan

Percentage Change

Europe

Percentage Change

Asia/Pacific-ROW

Percentage Change

Source: Dataquest (January 1995)

1993

6,910

34.9

2,171

37.1

2,424

14.9

967

53.5

1,348

68.3

1994

10,350

49.8

3,105

43.0

3,517

45.1

1,198

23.9

2,530

87.7

1995

12,031

16.2

3,530

13.7

3,943

12.1

1,396

16.5

3,162

25.0

1996

11,203

-6.9

3,542

0.3

3,541

-10.2

1,352

-3.2

2,768

-12.5

1997

11,530

2.9

3,752

5.9

3,562

0.6

1,322

-2.2

2,894

4.6

1998

13,731

19.1

4,505

20.1

3,932

10.4

1,545

16.9

3,749

29.5

15.7

10.2

9.8

22.7

SEMM-WW-PD-9501 ©1995 Dataquest Incorporated February 20,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Wafer Fab Equipment Market Cycles Are Real

The cyclicality of the market will be covered in detail later in this article, but it should be noted here that the wafer fab equipment market has experienced more severe cycles than has the semiconductor industry.

Figure 1 compares the equipment and semiconductor markets. The revenue of both markets are plotted on separate logarithmic scales, and superimposed, in order to directly compare one with the other. As can be seen, the semiconductor chip market has a trend upward — not a straight line, but close to it. However, because the industry cycles through over- and undercapacity situations, capital and equipment spending is more volatile. The wafer fab equipment market can be viewed as "a cyclical market arotmd a trend." As mentioned, the reasons for this will be explored later in this article.

Based on cycHcal trends, we would expect 1999 and 2000 to be good years. Our preliminary estimate for the size of the market in the year 2000 is

$21 billion to $22 billion.

Forecast Summary

Our forecast for capital spending and wafer fab equipment sales during the next five years assumes the explosive growth in 1994 will carry over into

1995, with a modest decline in equipment spending in 1996 and a relatively flat 1997 before a resumption of double-digit growth in 1998. Although we continue to beUeve that the cycHcal nature of investment in semiconductor capacity wHl diminish, it will take a couple more boom-and-bust cycles before the underlying semiconductor growth is large enough to dampen the memory component of the cycle. Equipment companies concentrating on or diversified in multilevel metal logic technologies should fare best in the coming slowdown.

Figure 1

Wafer Fab Equipment Market Cyclical around Semiconductor Trend

Log Scales

- - - Semiconductor Production

^ ^ ^ ^ Equipment

^^<-r~- - - - -

^ .

^ ^

,^r"=^^

0

1 1 1 1 1 1 1 1 1 1

1

1 1

1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998

Source: Dataquest (February 1995)

9500619

SEMM-WW-PD-9501

©1995 Dataquest Incorporated February 20,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Semiconductor memtifacturing is a global business. Production of semiconductors is constantly shifting among regions, as new capital money is flowing toward areas of relative lower capital cost and higher growth areas of consumption. Shifts in semiconductor production mean that equipment and material suppliers will absolutely need a global presence in every sense of the w o r d to remain competitive in the market. Product supply and support can no longer concentrate on local trends, as all major semiconductor companies have made it clear they are investing on a worldwide basis.

Dataquest Predicts

We believe that the forces driving the cyclicality of the equipnent market have not

changed, and that DRAM density crossover and looming price declines mean the equipment market should start a downturn within the next year.

The wafer fab equipment market has cycled through boom and bust times.

With the steUar growth of 1994 being within a whisker of the growth experienced in 1988, and 1995 shaping u p to be a growth year similar to 1989, questions about the cyclicahty of the wafer fab equipment industry have resurfaced. Is this cycle different?

Although the driving forces for the semiconductor industry itself have changed over the last several years, and the long-term growth rates of the

IC industry have stabilized at the 15 percent annual range (higher than most people thought a few years ago), w e believe that the key driving forces for the cyclicality of the equipment market have not changed.

We have been developing and refining various models during the last several months that give us a better picture of the magnitude and timing of these cycles, as well as the triggers and indicators to watch. Our view is that the dynamics of the DRAM market hold the key, and we will explore these dynamics in the remainder of this article.

Trend Modeling Toois and indicators

We have developed and are refining three models or indicators that will help us understand the magnitude and timing of investment trends in the wafer fab equipment markets. These three models have very different purposes and give an indication of different crosscurrents in the equipment market—although they generally move together within the same cycHcal trend. We will detail one of these indicators, but a review of all three is appropriate here.

It should be interesting to note w h e n reviewing these models that none of these models is particularly complex. Each reHes on fairly simple input that is well understood. We believe that, if the input is simple, the output can be logically interpreted without much error and without the temptation to read more into it.

Net Cumuiative Investment IViodel

The net cumulative investment (NCI) model tracks the investment of wafer fab equipment relative to semiconductor revenue or production. The model's purpose is to identify the trends and values of the absolute level of business for the wafer fab equipment market (that is, quantitative). It provides an annual snapshot of the market, and is more of an indicator of

SEMM-WW-PD-9501 ©1995 Dataquest Incorporated February 20,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

future trends in the aggregate balance of semiconductor capacity. This model is Hmited in its abihty to show any specific timing of an upturn or downturn, but does give a general one- to two-year indication of a turning point.

The basic assumption is that the industry invests a stable percentage of revenue toward wafer fab equipment over the long term. Of course, there are factors that can fundamentally change this level, but these tend to be longterm in nature and do not affect conclusions reached in the near term (two years). If there has been a trend in the last 10 years, it is that the percentage of revenue invested in wafer fab equipment has actually declined slightly.

The model was recently refined to take into account fundamental changes in the long-term demand for semiconductors, which have affected the absolute levels of the equipment market slightly but have not changed the conclusion regarding the cychcahty of the market.

What is this model showing today? The details of the model are described in our most recent forecast report (SEMM-WW-MT-9402, dated December

26,1994) and wiU not be shown in detail here. However, the model indicates that the semiconductor industry will be nearly $2.3 billion overinvested in wafer fab equipment (or about 19 percent) by the end of 1995

(assuming a 16 percent growth in wafer fab equipment this year). The last peak in 1989 was at about the 17 percent level, so the current level is within the window for overcapacity to emerge in some areas of the semiconductor market, and is also consistent with the equipment market growing slower than the semiconductor market for 1996 and 1997.

DRAM Quarterly Price Statistic

We have discovered an indicator that appears to have a direct correlation to the cycUcaUty of the equipment industry. This statistic tracks the relative average DRAM price. This statistic is not the price of a particular DRAM, and it is not a price-per-bit calculation. It most closely resembles the relative profitability of the DRAM manufacturer and indicates how much the buyers of DRAM are pulling dem«tnd for the current mainstream part. If this statistic is rising, demand is accelerating and supply is tight. If this statistic is falling, demand is either faUing or making the transition away from the mainstream part. If this statistic is stable, either supply and demand are in bjilance or the DRAM manufacturer is pricing the parts relatively close to cost.

This statistic and how it relates to the equipment industry is the major focus of this article. This statistic has continued to rise from the third quarter of

1992 through the fourth quarter of 1994—indicating that a significant dechne in the equipment market is not expected in the next six to nine months.

^ Wafer Fab Equipment Leading Indicator

We are refining a monthly leading indicator for wafer fab equipment. The purpose of this indicator is to provide a three- to five-month advance

"warning" of major turning points in equipment bookings. We expect to pubhsh this indicator in more detail within the next few months, as refinements are made to the model.

Table 2 summarizes the various models and their characteristics.

SEMM-WW-PD-9501 ©1995 Dataquest Incorporated February 20,1995

Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide

Table 2

Stmunary of D a t a q u e s t Indicator M o d e l s for Wafer Fab Eqttipment Market

Dataquest Model

Net Cumulative

Investment

DRAM Pricing Statistic

Frequency

Armual

Quarterly

Wafer Fab Equipment

Leading Indicator

Monthly

Source: Dataquest (February 1995)

Measures

Equipment spending relative to semiconductor revenue

Quantitative?

Yes

DRAM pricing and profitability trends

No

Acceleration profile of booking trends

Moderately

Indicates

Timing?

1 to 2 years

2 to 3 quarters

3 to 5 months

A preliminary curve for this indicator through the end of 1994 does not show any significant downturn in bookings through the first half of 1995, in a broad range of 25 to 40 percent year-over-year growth. We would expect the DRAM price statistic to show weakness before this indicator, however.

Segmenting the Equipment IVIaricet: Key to Gyclicaiity

Historical data (detailed in our forecast report and earlier in this article) clearly shows the equipment market to be cycHcal around the trend of semiconductor revenue—but why? Here we will propose a way to view and segment the equipment market that can be used to characterize its cyclicality.

We beUeve that the market for equipment can be cleanly segmented into two markets —equipment sold for DRAM capacity, and all other capacity

(primarily logic and ASIC oriented). This would be a convenient market split for several other reasons. First, the technology road maps for DRAM and logic/ASIC are distinctly for some areas of process equipment, as recently outlined in our Focus Report (SEMM-WW-FR-9401, dated November 28,1994). Second, semiconductor companies can be segmented both by regional ownership and product strategies along the lines of DRAM, microcomponents, and ASIC/other logic. Last, there is some segmentation along these lines already inherent in how equipment company strategies are developed. For example, most equipment segments that are critical for the

DRAM are dominated by Japanese suppliers, and most critical for logic are dominated by U.S. suppliers.

From the perspective of capital spending, this segmentation is also useful.

Logic and ASIC device demand (and capital spending for capacity) is driven by the pervasiveness of semiconductors into everyday life, and partially on PC unit demand into the worldwide communications infrastructure. This investment tends to exhibit stable growth and depend more on world economies and global trends in the use of semiconductors, and therefore is event-driven and has only small cyclic fluctuations. Large U.S. companies such as Motorola and Intel, as well as the emerging dedicated foundry companies, are the key drivers of this spending and production.

Investment in DRAM capacity over time does depend on actual demand for bits, with respect to the level of investment, but the timing of that investment has more to do with the pricing of DRAMs and the resulting profitability of DRAM manufacturers. DRAM pricing, large volume demand,

SEI\/IIVI-WW-PD-9501

©1995 Dataquest Incorporated

February 20,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

and density crossovers tend to make this investment extremely cyclical and characterized by large equipment orders. Companies such as Micron Technology, Korean companies, most large Japanese companies, and to lesser extent Texas Instruments have a cycUc pattern to capital spending. A case in point is pictured in Figure 2, showing the capital spending trends of Micron

Technology —a very well rtm DRAM company.

Supporting Data for a DRAM-Logic Segmentation

As 1994 unfolded as a classic DRAM investment rarap year, we saw that the best-performing segments either had a direct tie to DRAM technology (such as high-voltage implant) or depended on new DRAM capacity spending to drive large unit consumption (steppers and high-current implant). The latter group also tends to exhibit compound annual growth rates (CAGR) either at or below that for the overall wafer fab equipment market. Further,

Dataquest estimates that 40 percent of the wafer fab equipment sold in 1994 went into DRAM capacity. However, DRAMs accounted for only 21 percent of semiconductor revenue in 1994—clearly the amoimt of investment in capacity in 1994 outstripped what the demand would logically dictate.

This is the characteristic of the cyclicahty, however, and there are years

(both past and in the future) where the investment was/will be below demand.

We have called these segments "DRAM-sensitive" because their market sizes have tended to follow the cyclic nature of DRAM capital spending.

One can imagine equipment segments, such as sputtering, that tend to be more closely tied to the logic investment patterns. These "logic-sensitive" segments we believe are less cyclic, when compared to the overall equipment market.

Figure 2

Capital Spending of Micron Technology, 1984-1995

Millions of U.S. Dollars

600

500

400

300

200

100

M^-^^-^^^^V-^^^^^LJ

• ' • • • ' - ' • • I I I • 1 — I I • — — 1

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

Est. Est.

Source: Dataquest (February 1995) ssooaso

SEMM-WW-PD-9501 ©1995 Dataquest Incorporated February 20,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

To test our theory, we did an analysis that took four good equipment years

(1988-1989,1993-1994) and compared them with four "bad" equipment years (1986,1990-1992) to determine any trends.

Table 3 shows a comparison of the overall wafer fab equipment market with the underlying semiconductor market trend (for the year prior). An average "bad" year in semiconductors was 1.5 percent growth, while an average good year saw nearly 26.0 percent growth. The wafer fab equipment market is more volatile/cyclic than its underlying trend market by about 14 to 15 percentage points, both u p and down. This top-Kne analysis determined that the years chosen would characterize the equipment market well.

Table 3

A v e r a g e "Good" a n d "Bad" Years for Wafer Fab E q u i p m e n t a n d

S e m i c o n d u c t o r s (Percent)

Strong Periods

(1988-1989,1993-1994)

Up 41.1

Up 25.7

Source: Dataquest (February 1995)

Weak Periods

(1986,1990-1992)

Wafer Fab Equipment Down 12.2

Semiconductor Production Up 1.5

(for Year Prior)

It is also interesting to note that, for 1995 and 1996, Dataquest is forecasting

14 and 10 percent growth in semiconductors, respectively. The wafer fab equipment forecast is minus 7 percent and plus 3 percent growth for 1996 and 1997, respectively — an average of 14 points lower for the two years combined. This is interesting trivia, and just a reflection of how the forecast model is now tied to the net cumulative investment model.

Table 4 summarizes how major equipment segments have fared in the good years compared to the bad. Although it is not perfect, the general trend holds that the less-cyclic segments tend to be those that are logic-sensitive and the more volatile are those that are DRAM-sensitive.

Tying the Equipment Marlcet Cyclicality to a DRAIVI Statistic

Earher in this article we described various models. Now that w e have shown that the wafer fab equipment market is cyclic and that cychcality depends on DRAM investment patterns, we will describe and tie the equipment market to a specific indicator.

To review, a statistic that tracks the relative average DRAM price appears to have a direct correlation to the cyclicality of the equipment industry. This statistic is not the price of a particular DRAM, and it is not a price-per-bit calculation. It most closely resembles the relative prof itabihty of the DRAM manufacturer and indicates how much the buyers of DRAM are pulling demand for the current mainstream part. The statistic is shown in Figure 3.

This statistic indicates a very regular pattern, and w e have normalized the average selling price (ASP) within each cycle in order to show this regularity. Price peaks have occurred every five years, and are coincidental with a market transfer to the next density level for the DRAM. The 1989 peak

SEMM-WW-PD-9501 ©1995 Dataquest Incorporated

February 20,1995

I

I

Semiconductor Equipment. Manufacturing, and Materials Worldwide

Table 4

Average "Good" and "Bad" Years for Wafer Fab Equipment Segments (Percent)

High-Current Implatit

Medium-Current Implant

Steppers

Diffusion •

Total Wafer Fab

Dry Etch

Critical Dimension

CVD

Track

Sputtering

Epitaxial Silicon

Auto Wet Stations

Wafer Inspection

RTF

Dry Strip

Weak Periods

(1986,1990-1992)

Down 43.0

Down 34.9

Down 16.9

Down 16.7

Down 12.2

Down 10.0

Down 7.0

Down 3.9

Down 2.8

Down 0.2

Up 3.8

Up 5.5

Up 6.7

Up 7.8

Up 7.9

Strong Periods

(1988-1989,1993-1994)

Up 69.6

Up 51.8

Up 57.6

Up 50.0

Up 41.1

Up 50.5

Up 28.6

Up 47.8

Up 43.3

Up 34.3

Up 38.4

Up 49.4

Up 70.3

Up 21.7

Up 38.9

Note: This table represents cyclical trends, therefore segments and periods were excluded when an underlying major trend exists.

Example: High-voltage implant is a known DRAM technology but grew more than 300 percent in 1986 because it was an emerging technology. Long-term growth rates for wafer inspection and auto wet stations make these segments perform better than the market historically for all periods analyzed.

Source: Dataquest (February 1995)

Figures

Normalized DRAM Price Statistic, 1977-1994

Normalized Quarterly DRAM ASP

4 . 5 -

0.0

— I 1 1 1 1 ! 1 1 1 1 1 1 1 1 1 1 1

1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

Source: Dataquest (February 1995) ssQoazi

SEMM-WW-PD-9501

©1995 Dataquest Incorporated February 20,1995

10 Semiconductor Equipment, Manufacturing, and Materials Worldwide

represents the beginning of the end for the 1Mb DRAM, and the start for the 4Mb DRAM. To better tmderstand the dynamic, let's go through a cycle.

Each cycle consists of about two years of stability, followed by seven to nine quarters of price rise, followed by three to five quarters of price decline.

During the price rise, the demand for the mainstream part catches and exceeds supply, which is currently happening at the 4Mb level. Typically, the next-generation DRAM is close to production availability and capacity additions are made in advance of the demand during this time for the nextgeneration part (this is happening at the 16Mb level). For there to be a price rise, a clear mainstream part has to be in the market, which occurred near the end of 1992 for the 4Mb DRAM.

At the price peak, the next generation of DRAM has become popular enough to start to puU demand away from the mainstream generation, thereby applying price pressure to the mainstream product. This does not correspond to a price-per-bit crossover, and has preceded the price-per-bit crossover in the past by several quarters. In general, the DRAM buyer has

"pulled" the next generation into the market. Once this process starts, several dynamics occur. First, as demand falls off for the mainstream part, companies with a weaker position in the new part tend to lead a price cut to keep the market, putting pressure on profitabihty and thus available capital for equipment purchases. Second, equipment being used to produce the mainstream part is transferred to the more advanced part to conserve capital and add production capacity for the new part.

With profitability down somewhat, the price stabilizes. As demand ramps further for the new part, shrink designs and yield increases are used to increase capacity instead of capital outlays. This continues until demand outstrips supply by more than about 20 percent, putting u p w a r d pressure on prices. Prices and profitability rise, capital is again freed, and the cycle begins again.

In summary, if this statistic is rising, demand is accelerating and supply is tight. If this statistic is falling, demand is either falling or making the transition away from the mainstream part. If this statistic is stable, either supply and demand are in balance or the DRAM manufacturer is pricing the parts relatively close to cost.

How is this tied to the equipment market? As can be seen from Figure 4, where w e have overlaid the price statistic with historical boom and bust periods, any time the price statistic rises an equipment boom occurs, and within one to three quarters of the price peak a period of tough times ensues. In effect, this statistic cein be viewed as a leading indicator for the equipment industry. This indicator started to rise in the third quarter of

1992. If Dataquest had used this indicator at that time, w e would have forecast a start of a boom in 1993. This indicator has continued to rise through the fourth quarter of 1994, indicating that a significant dechne in the equipment market is not expected in the next six to nine months.

f

SEMM-WW-PD-9501 ©1995 Dataquest Incorporated February 20,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Figure 4

Wafer Fab Eqtiipment Market Cycles D r i v e n b y D R A M Pricing

Boom Time

Bust Time

11

1977

1979

1997

Source: Dataquest (February 1995) esooeaa

Dataquest Perspective

The wafer fab equipment market has cycled through boom and bust times.

With the stellar growth of 1994 being within a whisker of the growth experienced in 1988, and 1995 shaping u p to be a growth year similar to 1989, w e have become concerned about the possibiHty of tough times ahead. We beheve that the key driving forces for the cyclicality of the equipment market have not changed.

We have been developing and refining various models over the last several months that give us a better picture of the magnitude and timing of these cycles, as well as the triggers and indicators to watch. Our view is that the dynamics of the DRAM market hold the key, and in particular DRAM price trends and density crossover periods.

All of our analysis indicates that the first half of 1995 wiU remain strong in the wafer fab equipment market, but the second half may hold some softening. The current DRAM cycle is already into its sixth year, and the priceper-bit crossover of the 16Mb DRAM is imminent (currently about the third quarter of 1995).

Although the severity of this downturn will not be as serious as the past, because the end semiconductor market is strong, the cycUcaHty of the equipment market still cannot be avoided this time aroimd. Our current forecast is consistent with the DRAM price statistic peaking during the first quarter of 1995, with a price-per-bit crossover for the 4Mb to 16Mb DRAM complete during the third quarter of 1995.

SEMM-WW-PD-9501

©1995 Dataquest Incorporated

February 20,1995

12 Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide

What if the indicators presented here remain positive for the next six months? How would our forecast be modified in July? We would probably nudge up growth in the wafer fab equipment market for 1995 to the low- to midtwenties, place 1996 in slightly positive territory (nearly flat), and bring down 1997 significantly.

For More Information...

Clark J. Fuhs, Senior Industry Analyst (408) 437-8375

Internet address [email protected]

Via fax (408) 437-0292

The content of dus report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in

L 1 3 l 3 Q I J I ^ S l p u t to other parties shall be made upon the written and express consent of Dataquest

_ , ©1995DataquestIncorporated—ReproductionProhibited n m acompanyof ^ '^ '^

MMB ihcDun&BndsticetCorpoiatnn Dataquest is a registered trademark of A.C. Nieben Company

DataQuest

I

I

Market

Analysis

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Market Analysis

1994 Stepper IVIarl(et: Reflection of the Growing Strength in

Semiconductor Production

Abstract: in this document we will review the 1994 lithography stepper market and its

dynamics. We will compare 1994 regional activity and company market share with 1993.

The purpose of the document is to set the stage for Dataquest's latest stepper fore cast and its underlying assumptions to the end of the century. The stepper forecast and analysis will be published soon in another Market Analysis.

By Nader Pakdaman

Introduction

The stepper market continues to be a very sensitive indicator of the semiconductor industry's status and its capital spending trends. The evolution and market growth of stepper technologies also serve as a barometer of trends in semiconductor production. Although the semiconductor market grew nearly 30 percent in revenue from 1993 to 1994, the overall front-end equipment market grew by over 56 percent in this time frame. The stepper market accelerated even further with revenue growth of over 80 percent in

1994.

The demand for steppers has gained further momentum in 1995. We recently reported the stepper capacity constraints and backlogs stretching into 1997 (SEMM-WW-MA-9504). The strength of the stepper market was caused not only by increases in unit shipments but also by higher average selling prices (ASPs) of the systems. The production capacity buildup that began in mid-1993 has shown no signs of slowing down, as manufacturers are benefiting from the increased demand in semiconductor applications.

This demand is manifest not only in numbers but also for higherperformance ICs. The faster rate of shrinks and emphasis on productivity has driven the higher ASPs of the steppers.

DataQuest

Program: Semiconductor Equipment, IVianufacturing, and iVIaterials Worldwide

Product Code: SEMM-WW-MA-9505

Publication Date: November 13, 1995

Filing: IVIarket Analysis

Semiconductor Equipment, Manufacturing, and IViaterials Worldwide t t

Regional Markets and Analysis

Figure 1 shows the regional market splits, stepper unit shipments from each region, and company market share by revenue. Hitachi's departure from optical lithography changes the 1994 market share of the current five stepper suppliers. In terms of regional unit consumption, we estimated that

251 steppers were shipped into North America. This is 30 percent of the total units, similar to the unit share of this region in 1993. Manufacturers in

Japan consiimed 264 units, near 30 percent growth over 1993. These vinits comprised 32 percent of the total worldwide steppers shipped in 1994, below the 36 percent of 532 units we estimated for 1993. The strong growth was primarily a result of the underinvestment in manufacturing capacity by Japanese companies in 1992, for which they have begtm to strongly compensate since 1993. But the lower share of the total market exemplifies the strong and vmdeniable investment trend in the other two markets, Europe and Asia/Pacific. The estimated 6 percent decrease in Japan was compensated equally by these two regions.

Figure 1

1994 Stepper Market and Ownership

Asia/Pacific (14%)

Markets

Ownership

European Companies (13%)

U.S. Companies (9%)

Total = 826 Units

SVG Lithography (2%)

Ultratech Stepper (3%)

Revenue

ASM Lithography (15%)

Total = 826 Units

Total = $1.84 Billion

Source: Dataquest (November 1995)

SEIVIIVl-WW-MA-9505 ©1995 Dataquest Incorporated

95061SJ i\lovember13,1995

Sgmieonductor Equipment, Manufacturing, and Materials Worldwide

Europe and Asia/Pacific are experiencing massive investments by local and foreign capital. The strength of spending in contract manufacturing in

Asia/Pacific has been coupled by large investments for DRAM manufacturing, especially in Korea and Taiwan. Along with heavy spending by local manufacturers, large investments of foreign capital and technology have continued to pour into the Asia/Pacific region. We expect local consumption and export markets for semiconductor-related products to continue to attract more capital in both Asia/Pacific and Europe. In Europe, all major local manufacturers have committed to multibillion-dollar manufacturing projects in their local market. The U.S., Japanese, and Korean companies have followed suit with their expansions or new plans and joint ventures in this region.

Table 1 shows Dataquest's estimates of regional revenue and unit growth for each of the stepper suppliers. Europe constituted the largest growth rates over 1993 in both revenue (129 percent) and units [95 percent) than any other region of the world. The growth in revenue for stepper manufacturers in the European market was followed by our estimates of less than

110 percent in Asia/Pacific. North America and Japan followed strongly with 84 and 53.4 percent growths in terms of revenue, respectively.

Company Ranking and Analysis

In terms of vmits, market share by regional base of company did not change significantly from 1993 in 1994. However, from a revenue perspective, we saw a different picture in 1994, primarily because of Nikon's increased market share. We estimate that Nikon held 49 percent of the total market of

$1.02 billion in 1993. Nikon increased its market share to 55 percent of the

Table 1

1994 Regional Unit and Revenue Growth Rates by Company (Percent)

Japan

Europe Asia/Pacific North America

Revenue

ASM Lithography

Canon

Hitachi

Nikon

SVG Lithography

Ultratech Stepper

Total Growth

Units

ASM Lithography

Canon

Hitachi

Nikon

SVG Lithography

Ultratech Stepper

Total Growth

Source: Dataquest (November 1995)

42.7

123.4

-100.0

122.1

36.0

25.5

83.5

- 38.5

90.9

-100.0

74.1

-16.7

33.3

57.9

0

65.1

-100.0

53.8

0

134.9

53.4

0

45.6

-100.0

24.1

0

175.0

29.4

166.7

52.6

-100.0

127.3

0

71.4

94.9

167.4

81.1

-100.0

181.9

0

30.9

128.7

47.4

39.6

-100.0

203.1

10.5

309.0

109.1

38.1

25.0

-100.0

129.2

0

300.0

78.2

Worldwide

55.1

52.5

-100.0

62.6

-14.3

86.1

55.3

60.6

75.0

-100.0

106.1

31.7

76.3

82.4

SEMM-WW-MA-9505

©1995 Dataquest Incorporated November 13,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

$1.84 billion in 1994. In spite of strong growth in revenue by every player in the market, except for Ultratech Stepper, every stepper supplier lost market share in 1994 to Nikon.

Table 1 showed that revenue grew faster than units in every region. This difference was the strongest in Europe, followed closely by Asia/Pacific.

Therefore it is seen that advanced system installations vfitix higher ASPs in these regions reflect the higher rate of capital spending in the high end of the technology spectrum, compared with previous years. The larger and more mature advanced manufacturing markets in Japan and Norlii

America showed a similar trend, although with a smaller difference than the two previous regions.

Table 2 lists the stepper companies by their ranking in unit shipments. TTie order did not change, except for Hitachi's exit. I-line steppers' share of the technology mix grew over g-line and deep-UV systems, with 68 percent growth from 1993. G-line system shipments grew by 25 percent, and deep-

UV shipments fell by 28 percent.

The i-line systems dominated the market with nearly 83 percent of the shipped units. The increase in ASPs was primarily by this technology.

Canon begun the large volume shipments of its flagship high-resolution

FPA-3000-i4 system in early 1995, tikerefore its ASPs lagged those of the other two players (Nikon and ASM Lithography, or ASML) in the advanced high-resolution i-line market. A considerable portion of ASML's and

Nikon's shipments in 1994 comprised their respective PAS 5500/100 sCTies and the NSR-200511 series; steppers. These systems carry a considerably larger price tag than the models they have replaced. Because of more upgrades and newer models capable of higher throughputs in 1995 introduced by all three of these players, we should expect the ASPs of advanced i-line systems to increase even more. We will discuss this important trend further in the forecast document.

As shown in Table 1, Nikon and Canon both had revenue growth in North

America of over 120 percent, although Canon's unit growtih of over 90 percent was higher than Nikon's 74 percent growth. In Japein, Canon's revenue grew more than Nikon's 54 percent growth. Canon's unit growth rate was

Table 2

1994 Worldwide Stepper Company Ranking (Unit Shipments)

Nikon

Canon

ASM Lithography

Ultratech Stepper

SVG Lithography

Hitachi

Worldwide Total Market

Percentage Technology Mix

Source: Dataquest (November 1995)

Units

431

215

107

67

6

0

826

Share (%)

52.2

26.0

13.0

8.1

0.7

0

100.0

Change from

1993 (%)

63

53

55

86

-14

-100

55 g-Line

41

24

8

51

0

0

124

15 i-Line

384

191

95

16

0

0

686

83

Deep-UV

6

0

4

0

6

0

16

2

SEMM-WW-MA-9505 ©1995 Dataquest incorporated

November 13,1995

#

Sen^conductor Equipment, Manufacturing, and Materials Worldwide nearly twice that of Nikon's shipments into Japan. We surmise that, although Nikon was shipping relatively higher-priced steppers in 1994 to these two regions. Canon was shipping units at a higher rate. Because the product mixes and technical road map of these two leaders of the stepper market have historically been relatively similar, should we expect their

ASPs to converge? In the short term, capacity and backlog will determine the answer. In the long term, however, the ASP-versus-imits issue of all the suppliers in the market will be settled more or less in the deep-UV arena, and the dynamics of its adoption in volume manufacturing. Deep-UV lithography will be a major topic in the forecast document.

ASML has had consistent strong growth in the past several years in both revenue and uruts, without the benefits of sales in the Japanese market.

ASML's strong presence in Asia/Pacific (in particular in Taiwan), along with a growing large installed base in North America, has allowed it to better weather the cyclic nature of the stepper market. In 1995, as seen in

Table 1, ASML had similar strong growth rates in North America and Asia/

Pacific, in both units and revenue. However, its worldwide growth rates were higher than the figures in these two regions, because of its over

165 percent growth in both revenue and units in Europe, ASML's smallest market. Dataquest's recent updated front-end equipment forecast shows over 20 percent average growth in Europe to the end of the century. This is a market that should play a critical role in the worldwide competition for market share.

Ultratech Stepper further grew its leading position in the g-line market in

1994. Ultratech grew its g-Iine unit shipments by over 50 percent in 1994.

This is an arena that Ultratech aptly calls the scanner replacement market.

Dataquest estimates that over 55 percent of all square inches of silicon processed in 1994 were for devices that had critical geometries of more than one micron. Although the other players in the market are phasing out their g-line systems (except for special markets or applications) in favor of their high-resolution systems, Ultratech has strengthened its focus on this important market. Many manufacturers, among them producers of microcontrollers, discrete^ and analog devices, are well aware of the benefits and advances of steppers over its precedent technology, scanner systems. These benefits include high throughput, larger field sizes, higher resolution along with better overlay, and reliability—in short, higher productivity.

Ultratech's i-line shipments grew from our estimated three uruts in 1993 to

16 units in 1994. Its 2244i i-line stepper is designed as the high-throughput companion to high-resolution steppers of other stepper suppliers in mixand-match lithography technology. Ultratech has led the charge in ttiis growing market, and in 1994 introduced a second model of its highthroughput stepper line. The other three players in the i-line market also introduced their respective versions and models of high-throughput systems in 1994. For the 1995 market and future years, Dataquest will begin to estimate the number of steppers used as the high-throughput systems in mix-and-match scenarios.

We estimate that SVG Lithography (SVGL) shipped one fewer Micrascan unit in 1994 than it did in 1993. However tiie ASPs of the systems shipped grew by over 30 percent, and this resulted in an overall growth in revenue for SVGL. SVGL has built a backlog that signifies the onset of deep-UV lithography's merge with mainstream processing. SVGL has continued to make investments to increase capacity and also develop the Micrascan IE, a

0.25-micron version of its step-and-scan steppers. The introduction and

SEMM-WW-MA-9505 ©1995 Dataquest Incorporated November 13,1995

Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide capacity availability of the Micrascan in is a significant event for SVGL, because Nikon's recent shipment of its first NSR-201 0.25-micron deep-UV scanning system into IBM's East FishkiU, New York, facility has raised the bar in deep-UV stepper technology.

SVGL should not be solely evaluated on the basis of market share, but should also be viewed from the prism of the deep-UV stepper technology and market dynamics. SVGL had a watershed year in 1994, and the events did not concern the immediate market. The negotiations with Canon over the licensing of SVGL's deep-UV step-and-scan Micrascan technology was finally terminated after more than a year of discussions without an agreement. SVGL in turn found other partners to address its capacity and R&D capital needs. SEMATECH, along with several U.S. semiconductor manufacturers, committed nearly $100 million of capital to further develop the

Micrascan technology and build the infrastructure for sales and support of this critical lithography technology. Our forecast document will discuss in detail our assumptions on the evolution of the deep-UV market.

+>

Oi

Dataquest Conclusions

Tl

0)

The stepper market in 1994 spearheaded the strong growth in the front-end equipment market. The strength was in both units and ASPs. One other very important facet of the 1994 market was the growing difference in the resolution capabilities of the high-resolution steppers and the critical geometries of ttie advanced ICs in volume manufacturing. Most of the highresolution steppers shipped in 1994 were geared to process sub-0.5-micron devices. Manufacturers historically have leveraged the added flexibility in manufacturing by processing the critical geometries of the ICs at resolutions lower than the capability of the steppers. This so-called resolution gap increased noticeably in 1994, as critical device geometries processed in advanced manufacturing were more or less in the 0.5-micron regime and the mass of the steppers shipped throughout 1994 increasingly tended toward the 0.35-micron capabilities. We have observed the same trend continuing in 1995 and will further discuss the ramifications of this technology gap in the stepper forecast document.

en

<B in fH o tc) in U

u

o

ft

u

o o

CPi O

1 S H

§ O 4J

I -^ m i^l'^

E • -p »

S M «s I c n S P '-'

I

I

CO

H

O

E-i

H

In summary, we saw the following trends:

• ASPs grew faster than units in every region.

• Europe and Asia/Pacific led the market in growth of revenue and imits, which accompanied strong growth in North America and Japan.

• The company rankings did not change, while Nikon and Ultratech increased their market share.

• Mix-and-match continued its ascent from a concept to a practical solution in addressing rising costs and the increasing need for productivity.

For More Information...

Nader Pakdaman, Senior Industry Analyst (408) 468-8417

Internet address [email protected]

Via fax (408) 954-1780

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in

O S I l d Q U C S t part to other parties shall be made upon the written and express consent of Dataquest.

©1995 Dataquest Incorporated—Reproduction Prohibited

BB ilieOunlLBradstiectCorpoiation Dataquest is a registered trademark of A.C. Nielsen Company

Market

Analysis

Semiconductor E q u i p m e n t , Manufacturing, and Materials W o r l d w i d e

Market Analysis

Lithography Stepper Capacity: Bottienecic of

Semiconductor Production

AbstrSCt' Semiconductor manufacturing's unprecedented and sustained growth of the past

several years has put a significant strain on the infrastructure of the markets that support it. In manufacturing semiconductors, lithography defines capacity and technical capability.

The capacity constraints of the suppliers in this critical lithography stepper market are discussed in this article by looking at each stepper supplier.

By Nader Pakdaman

Stepper Suppliers Step Up Capacity

Nikon recently announced expansion of its stepper production facility in

Kumagaya, Japan. The company will increase output at its main plant to nearly 1,000 steppers per year from a current capacity of about 650 units.

The primary emphasis will be on the high-end i-line systems (NSR-

2205illD and beyond) and the newly announced scanning stepper

(NSR-S201A deep-UV system). Full capacity will be reached gradually through 1996.

At about the same time, IJltratech Stepper and ASM Lithography (ASML) announced that they are holding preliminary discussions to form a strategic alliance. Ultratech and ASML will offer their clients the option of mixing and matching their respective noncritical and critical resolution steppers.

The primary goal of the d iscussions is to focus on developing faster responses to capacity constraints.

Dataquest Analysis

The news releases on both of these announcements were short but significant. Without steppers, one cannot build a fab and expect to pattern the

DataQuest

Program: Semiconductor Equipment, IVIanufacturing, and Materials Worldwide

Product Code: SEMM-WW-lVlA-9504

Publication Date: October 23,1995

Filing: IVlarket Analysis

Semiconductor Equipment, Manufacturing, and Materials Worldwide

circuits on the silicon. Concerns about infrastructure constraints in the semiconductor industry are mounting. Could the current undercapacity in manufacturing curtail the semiconductor industry's goal of surpassing

$300 billion by the end of the century? These concerns are balanced by warnings of historical overinvestment patterns that have taken the industry through some very hard times. There is merit to both of these arguments.

As usual/ the reality will probably lie somewhere in between and be more complex than either.

In a recent update of the semiconductor revenue forecast, Dataquest increased its forecast of average growth rates for the industry to over

20 percent until the end of the decade. Growth rates h a d been just over

15 percent for the first half of the decade. We believe this increase is a fundamental indicator of the growing prevalence of semiconductors in our lives. From set-top boxes to PCs and communications applications, semiconductors are changing all facets of our lives at home and work.

The other fundamental aspect of the industry is the cyclical nature of the markets and investment patterns that have produced cycles of relatively stronger and slower years. As the industry migrates to a new device generation, the cycle begins with investment to address undercapacity. Then, as device shrinks take hold and yield increases, demand is met and a period of overcapacity may occur. This cyclical pattern is seen very clearly in the

DRAM market.

In short, we are experiencing a fundamentally stronger demand for semiconductors that the industry is addressing by unprecedented increases in capacity. However, the central cyclical drivers still exist in the market. In

Dataquest's view, the issue of over- and undercapacity will not prevail uniformly in all markets and for all ICs. The Dataquest forecasts for the semiconductor market and the directly related manufacturing equipment and materials market reflect both the growth in demand and the cyclical nature of the market.

Since 1993, megafab announcements have become daily news. Capital spending figures for fabs ranging from $1 billion to $2 billion are now taken as ordinary events. The technology and capacity of these fabs are defined by the number and capabilities of their lithography steppers.

The lithography tools supply-demand equation has been producing backlogs extending into the first quarter of 1997. This may sound fantastic, but if these backlogs are not addressed in a timely fashion by stepper suppliers, it could spell disaster. As both technology and users' needs evolve, these extended delivery times could translate into changing specifications.

Changes in the market over such an extended period could even spur cancellations. A more balanced supply-demand picture with realistic backlogs would make planning and investment for stepper manufacturers and semiconductor manufacturers a more reasonable and less chaotic exercise.

All of the major stepper manufacturers (ASML, Canon, Nikon, SVG Lithography, and Ultratech Stepper) are facing historic capacity demands. Inability to address the problem could spell loss of market share in a segment that accounts for over 15 percent of the total front-end equipment market, on average. Dataquest estimated the stepper market at $1.8 billion in 1994.

SEMM-WW-MA-9504 ©1995 Dataquest Incorporated October 23,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Current estimates call for this market to reach $5 billion by the end of the decade.

Nikon has led the stepper market with shares about and above 50 percent in the last several years. Nikon's current capacity is approximately 650 stepper units a year. For 1995, this capacity should preserve the company's position in a market that we estimate will surpass 1,000 steppers for the first time, with over 1,200 steppers shipped to manufacturers. Nikon plans to sustain and even gain market share by quickly increasing capacity in a market that is highly cycHcal. The peak and trough years of the stepper market are very steep. But w e estimate that the growing demand in production will not be met by capacity until 1997, at least, and that the slowdowns that follow will not be as severe as those of previous cycles. Nikon's lens capacity, or the better availability of glass, is even more critical. It takes over a year to anneal the quartz material used in the optical train of the steppers. We must assume that Nikon has been planning this increase for many months and that its glass capacity will match its 1,000-stepper goal.

Nikon's Japanese counterpart. Canon, has been increasing capacity since early 1994. By our estimate. Canon is now running at a capacity of over

400 steppers per year. We expect Canon to increase its capacity to preserve and perhaps increase its share of about 25 percent of the stepper market.

Canon is on a very aggressive product introduction path as it tries to reestablish itself as a leader for 0.25-micron deep-UV lithography scanning tools.

The Ultratech and ASML announcement is also directly related to the capacity issue. ASML has been working hard with its optical shop, the lens supplier Zeiss of Germany, to increase capacity for its systems. Sophisticated tools and major investments are required to process the glass and manufacture and test the optical train for all stepper manufacturers. Based on announcements made by ASML and Zeiss, Dataquest estimates their lens capacity at fewer than 180 units for 1995. This figure could easily grow in the next several years.

ASML's primary strategy is to compete with Canon and Nikon on the highresolution stepper front as the industry moves to 0.35-micron and smaller geometries. However, resolution is not the only measure of steppers and equipment. Cost-effectiveness and productivity must match technical capabilities. Because of this, stepper manufacturers have introduced mix-and-match lithography as a means of reducing cost eind increasing throughput.

Ultratech has led the way in matching its high-throughput systems in a mix with high-resolution steppers on the fab floor. Canon, ASML, and Nikon have all followed suit with their high-throughput systems. By our estimates, Ultratech shipped nearly 70 steppers for semiconductor manufacturing in 1994. The company has been growing strongly in all regions of the world with its i-hne steppers for mix and match. Unlike other players in the market, which have product lead times of 18 months or more, Ultratech enjoys turneiround times of six to nine months. This is primarily because of the relatively simple optical design of the Ultratech steppers.

SEMM-WW-MA-9504 ©1995 Dataquest Incorporated October 23,1995

Semiconductor Equipment, Manutacturing, and Materials Worldwide

The agreement between Ultratech and ASML translates into a marketing strategy that would allow semiconductor maniif acturers to choose between two mix-and-match scenarios. ASML and Ultratech will each propose two mix-and-match packages to the customer. ASML will offer one package pairing its high-resolution stepper with its own high-throughput system and one with the Ultratech companion. Ultratech will offer its system partnered with ASML's equivalent to the high-resolution stepper (which the prospective client may have already chosen) or with the non-ASML stepper.

Discussions between the companies are still in the prehminary stages. Both will benefit from this agreement even if cooperation goes no further than the outlines of their announcement. Most Ultratech high-throughput steppers are matched with Canon and Nikon steppers because of the massive installed base of these two vendors. Ultratech will be well positioned to enjoy the benefits of ASML's client base and to further soHdify its position in the critical mix-and-match arena. In turn, ASML will get much-needed capacity relief for its high-resolution steppers.

It will be interesting to see what other synergy may exist between these two players. In 1994, ASML and Ultratech together held over 18 percent of a market that totaled more than $1.8 biUion. Each has been growing over

60 to 70 percent per year in the past several years. Ultratech's presence in

Japan may compensate for ASML's lack of visibility in this critical market.

From a manufacturing point of view, both companies rely heavily on their external suppliers in strategic OEM agreements. Could this similarity and flexibility in manufacturing strategy and this partnership in mix and match lead to cooperation and exchange of technology?

We cannot close this discussion without mentioning the other critical player in the market, SVG Lithography (SVGL). For SVGL, more than any other company in the market, capacity does not merely translate into market share but may determine its future. SVGL's deep-UV Micrascan has led hthography's foray into sub-0.5 micron deep-UV scanning lithography.

Dataquest estimates that SVGL holds over 25 percent of the installed base of deep-UV steppers, trailing only Nikon's near-40 percent share of installed deep-UV systems.

Nikon's shipment earlier this year of the first NSR-S201A 0.25-micron deep-

UV scanning system to IBM's East Fishkill, New York, plant has the market awaiting SVGL's next stepper—the Micrascan III. However, semiconductor maniifacturers are also interested in seeing how SVGL will build the infrastructure needed to support the deep-UV market into the next century.

SVGL has benefited from the approval and financial backing of several

U.S. semiconductor manufacturers and SEMATECH. Nobody doubts that

SVGL's position in technology is formidable, but annual unit shipments now in the low teens would have to increase to levels of deep-UV stepper capacity comparable to SVGL's competitors — ASML, Canon, and Nikon.

Semiconductor revenue and the ensuing growth in capital spending have translated into long backlogs for equipment suppliers, particularly for stepper manufacturers. If stepper undersupply is gating the supply of chips, the market's ability to spur demand by producing in volume and at lower cost will be greatly inhibited. We believe that by mid-1996, even at the current

SEMM-WW-MA-9504 ©1995 Dataquest Incorporated October 23,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

high rate of demand, increases in stepper capacity will relieve the pressure on the market. Through adoption of mix-and-match lithography strategies and/or faster shrinks of IC designs, semiconductor manufacturers will address the growing demand for their products.

SEMM-WW-MA-9504 ©1995 Dataquest Incorporated October 23,1995

Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide

For More Information...

Nader Pakdaman, Senior Industry Analyst (408) 468-8417

Internet address [email protected]

Via fax (408) 954-1780

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in

U3TaQUCSl part to other parties shall be made upon the written and express consent of DataquesL

_ _ ^ _ , ©1995 Dataquest Incorporated —Reproduction Prohibited m m aoomponyof i r r

l i O ThcDun&BradsticetCorporation Dataquest is a registered trademark of A.C. Nielsen Company

DataQuest

Year-End 1995 Forecast:

Capital Spending, Wafer Fab

Equipment, and Siiicon

Market Trends

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

ProduGt Code: SEMM-WW-MT-g502

Pnblieation Date: January 22,1996

Filing: Market Analysis

Year-End 1995 Forecast:

Capital Spending, Wafer Fab

Equipment, and Silicon

Market Trends

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MT-9502

Publication Date: January 22,1996

Filing: Market Analysis

Year-End 1995 Forecast: Capital Spending, Wafer Fab Equipment, and Silicon

Table of Contents

Page

1. Executive Summary 1

Year-End 1995: Backlog Heaven ... Needed for 1997 1

What Are the Trends? 2

Dataquest Perspective 4

2. Semiconductor Capital Spending Forecast 7

Chapter Highlights 7

Capital Spending Tables 8

And the Spending Binge Continues 8

When Will the Spending Boom End? 11

Is There Excessive Spending and Impending Doom? 12

Before 1985: An Immature Manufacturing Industry 12

1985 Through 1992: Becoming Manufacturing Smart 12

1993 through 2001: Growth 13

Beyond 2001: Maturing Manufacturing Infrastructure 13

The North American Market Continues to Exhibit Strategic

Strength 14

Japan: DRAM Capacity Additions Drive Spending, but a Strong

Yen Subdues 15

Europe Sustains Presence as a Growth Market 16

Asia/Pacific Is Madly Investing in Two Distinct Ways 16

Who's Investing Where? 18

Dataquest Perspective 20

3. Wafer Fab Equipment Forecast 21

Chapter Highlights 21

Annual Investment Themes for 1994-2000 26

When Will Capacity Expand to Meet Demand? An Update of the Over- or Underinvestment Model 26

When Will Capacity Expand to Meet Demand? A Fundamental

Set of Analyses 29

Highlights of Key Equipment Segment Markets and Forecasts 30

Steppers and Track 30

Etch and Clean: Dry Etch and Chemical Mecharucal .

Polishing 31

Deposition: CVD, PVD, and Silicon Epitaxy 32

Thermal Nondeposition Processes: Diffusion and RTP 33

Ion Implantation 33

Segments and Tools of Emerging Importance 34

Dataquest Perspective 35

4. Silicon Wafer Forecast 37

Silicon Forecast Tables 37

The 200mm Wafer Ramps U p 45

What about 300mm Wafers? 46

Highlights of the North American Silicon Wafer Market and

Forecast 47

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table of Contents (Continued)

Page

Highlights of the Japanese Silicon Wafer Market and Forecast 47

Highlights of the European Silicon Wafer Market and Forecast.... 48

Highlights of the Asia/Pacific-ROW Silicon Wafer Market and

Forecast 48

Silicon Wafer Revenue Forecast 48

Dataquest Perspective 49

5. Semiconductor Consumption Forecast 51

Semiconductor Consuniption 51

6. Semiconductor Production Forecast 53

Historical Semiconductor Production 53

Captive Semiconductor Production 53

The Move toward Asia Continues, Europe Growth Rests

Temporarily 55

Semiconductor Production Trends: Accelerating Shift to Asia/

Pacific 56

Dataquest Perspective 58

Appendix A—Economic Assumptions, Fourth Quarter 1995 59

Appendix B—Exchange Rates 65

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Year-End 1995 Forecast: Capital Spending, Wafer Fab Equipment, and Silicon

List of Figures

Figure Page

3-1 Net Cumulative Over- and Underinvestment of Semiconductor

Wafer Fab Equipment 28

3-2 Net Cumulative Over- and Underinvestment of Semiconductor

Wafer Fab Equipment 28

SEMM-WW-MT-9502 ©1996 Dataqu est January 22,1996

W Semiconductor Equipment, Manufacturing, and Materials Worldwide

List of Tables ^ . ^ « ^ ^ « ^ . ^ ^ . ^ ^ . ^ ^ ^ « ^ . ^ . ^ _

Table Page

2-1 Semiconductor Capital Spending—Top 20 Spenders, Comparison of 1994 and Projected 1995 Worldwide Capital Spending 9

2-2 Worldwide Capital Spending by Region — Historical, Includes

Merchant and Captive Semiconductor Companies 10

2-3 Worldwide Capital Spending b y Region — Forecast, 1994-2000,

Includes Merchant and Captive Semiconductor Companies 10

2-4 Regional Investment Patterns of Semiconductor Manufacturers in 1994 19

2-5 Regional Investment Patterns of Semiconductor Manufacturers in 1995 19

3-1 Worldwide Wafer Fab Equipment Market, by Region—

Historical, 1988-1994 23

3-2 Worldwide Wafer Fab Equipment Market, by Region—Forecast,

1994-2000 23

3-3 Wafer Fab Equipment Revenue by Equipment Segment—

Historical, 1988-1994 .....24

3-4 Wafer Fab Equipment Revenue by Equipment Segment—

Forecast, 1994-2000 25

3-5 Annual Driving Forces and Investment Themes for Wafer Fab

Equipment, 1995-2000 26

4-1 Forecast of Captive and Merchant Silicon and Merchant

Epitaxial Wafers by Region 37

4-2 Forecast Growth Rates of Captive and Merchant Silicon and

Merchant Epitaxial Wafers by Region 38

4-3 Forecast of Captive and Merchant Silicon Wafers by Region 39

4-4 Forecast Growth Rates of Captive a n d Merchant Silicon by Region 39

4-5 Forecast of Merchant Prime and Test/Monitor Wafers by Region 40

4-6 Worldwide Wafer Size Distribution Forecast, 1993-2000 41

4-7 North American Wafer Size Distribution Forecast, 1993-2000 42

4-8 Japan Wafer Size Distribution Forecast, 1993-2000 43

4-9 European Wafer Size Distribution Forecast, 1993-2000 44

4-10 Asia/Pacific-ROW Wafer Size Distribution Forecast, 1993-2000 45

4-11 Worldwide Merchant Silicon Wafer Revenue Forecast,

1992-2000 49

5-1 Worldwide Semiconductor Consumption by Region—

Historical 51

5-2 Worldwide Semiconductor Consumption by Region—Forecast 52

6-1 Worldwide Semiconductor Production by Region—Historical 54

6-2 Worldwide Semiconductor Production by Region—Forecast 57

A-1 Gross Domestic Product/Gross National Product Growth Rates:

Outlook as of December 15,1995—Constant Prices and

Exchange Rates, Local Currencies 60

A-2 Consumer Price Index Growth Rates: Outlook as of December

15,1995 60

B-1 Exchange Rates per U.S. Dollar 65

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Chapter 1

Executive Summary,

Year-End 1995: Backlog Heaven ... Needed for 1997

The year 1995 was one of those that the industry almost wishes will not be repeated for a while, because the infrastructure of supply is being strained.

Backlogs in the equipment industry exploded in 1995, with reports that stepper lead times are over one year. Companies are having trouble filling all their open positions, and silicon supplies are extremely tight. Yet the engines of demand for semiconductors—^PCs and communications/networking equipment—continue. A monthly leading indicator we have been developing went crazy in April and has not rested since.

A couple of trends do concern us, however. The number of planned fabs announced in the forecast horizon has increased dramatically from about

85 one year ago (counting 1994 and after) to over 130 (counting 1995 and after). The density of new fabs, about 35 per year, has not changed much, but the announcement horizon has lengthened. A year ago new fabs were announced to come on line in 18 to 21 months. Today, with the huge backlogs in equipment companies, announcements are made for fabs 24 to 30 months ahead. Funding to place an order for a stepper is generally done

before ground is broken on a new fab, just to get in line. There are already about 25 fabs armounced to come on line in 1998—a visibility that is imcharacteristic.

Are these fabs real? We believe so, because in order to achieve a $330 billion semiconductor market by 2000, the industry needs to maintain a run rate of 30 to 35 fabs per year. Our concern is that the 1996 and 1997 run rates are slightly ahead of this figure, and some of the 1997 and 1998 fabs may be delayed by six to 12 months, softening the quality of backlogs for

1997.

Although we think the recent talk on DRAM pricing has been overblown

(we see continued shortages through 1996), we do believe we are past the

"pinch point" in the supply/demand imbalance, and availability of DRAM is beginning to improve. We are now viewing 1997 as a year for a slight oversupply of DRAM (based on silicon area capacity), so capital investment in 1997 is likely to cool in this area.

Does this speU doom for 1997? No, the foimdry industry supply will get tighter in 1996, leading to increased spending for logic-oriented capacity over DRAM starting late in 1996 and continuing into 1997. The foundry market will not see relief in capacity until the middle of 1998, by our silicon area-based capacity analysis.

SEMM-WW-MT-9502 ©1996 Dataquest

Semiconductor Equipment, Manufacturing, and Materials Worldwide

What Are the Trends?

On the heels of booming growth in 1994, global semiconductor capital spending grew another 72 percent during 1995 to $38 billion. Anticipated continued tight capacity and a strong semiconductor market in 1995 mean continued growth into 1996, now forecast at about 30 percent.

North America is showing consistent strength, with a 69 percent growth in

1995. Worldwide demand for desktop connectivity products and telecommunications equipment continues to fuel the investment strategies in

U.S.-manufactured semiconductor products, heavily weighted toward logic and ASIC capacity. North American capital spending is expected to remain strong in 1996 and to moderate in 1997 as these investments are absorbed, but we expect the North American region to grow at fasterthan-market rates as foreign multinationals and foundry companies invest in capacity in the United States.

Japanese companies are continuing to invest in semiconductor capacity to preserve their market share position in memories, although the strenglii of the yen during the middle of the year temporarily put a lid on spending enthusiasm. Japan as a region kept pace with the world in investment in

1994 but lagged the market in 1995 as Japanese comparues invested more outside Japan. Healthy, but subdued, growth of 47 percent in spending within Japan occurred during 1995—about 34 percent on a yen basis. Lagging investment patterns in Japan are expected to continue throughout the decade. Japanese companies, however, grew spending during 1995 at about 61 percent worldwide, for a total of $12.2 billion, second only to

North American companies' spending of $13.6 billion and well ahead of

Asian companies' $9.1 billion.

Dataquest has been bullish on the prospects in Europe, and remained so in

1995, although the region came in slightly under expected spending. European companies are a large part of this expansion, aided by strong domestic economies, and major projects by the multinational manufacturers are also contributing. We still see Europe as a significant growth region for spending through the decade.

Following very strong capital investment growth of 77 percent in 1994, the

Asia/Pacific-Rest of World (ROW) region grew an astronomical 104 percent in 1995, as Korean DRAM expansion accelerated (further), foundry expansion in Taiwan, Singapore, and others continued to grow, and new

DRAM players entered the scene in Taiwan. These new projects started in

1995 will continue to consume capital funds in 1996. We are expecting about 31 percent growth in 1996. Asia/Pacific-ROW will continue to be one of the fastest-growing regions through this decade.

Dataquest believes that the relatively large capacity expansion of 1993 to

1995 (three-year growth of 265 percent) has now exceeded the three-year growth recorded in the 1987-to-1989 expansion. It should be noted, however, that the two periods are different in two key respects. First, the current period is experiencing accelerated long-term growth for the underlying semiconductor industry, driven by a productivity-related

PC boom. The PC boom is expected to continue, so we are not overly alarmed about the magnitude of this cycle. Second, the manufacturing

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Executive Summary infrastructure is more efficient today, and there is a diminishing return in productivity and yield improvements. This has led to a higher natural capital investment ratio required today than in the 1987-to-1989 period, closer to 22 percent of revenue, on average, being a standard (versus 18 percent in the late 1980s).

However, we also believe that, in 1996, spending will decelerate, starting in the second half of the year, causing a relatively flat spending pattern through 1997 and 1998. Although we continue to believe that Sie cyclical nature of investment in semiconductor capacity will diminish, the PC boom must continue to drive the underlying semiconductor growth strongly enough to dampen the memory component of the cycle. After a flat two-year period, investments should pick up again in 1999.

Wafer fab equipment spending grew 66 percent in 1995 and is expected to grow 36 percent worldwide in 1996, driven by massive spending in the

Asia/Pacific region and strong investments in the United States. DRAMsensitive Japanese investment and an expanding European production base complete the strong picture for 1996. Segment growth in 1994 and

1995 is being led by DRAM or capital spending-sensitive equipment, with steppers, implant, wafer inspection, and factory automation exhibiting significantly stronger-than-market growth. New technology segments, such as chemical mecharucal polishing (CMP), high-voltage implant, and rapid thermal processing (RTP) were tiie fastest-growing segments. We expect no major segment declines in 1996, as capacity additions are broadbased and worldwide. Record backlog levels going into 1996 will be a buffer against a market decline in 1997, but we do expect softness from the

DRAM companies to become evident by the end of 1996.

After strong expansion years from 1993 through 1996, equipment purchases in 1997 should decline markedly, followed by a slight decline in

1998. Investment in DRAM capacity will be curtailed as producers elect to convert their 4Mb DRAM capacity to 16Mb, which adds bit capacity through the instant increase in bits per square inch. Also, many Japanese

DRAM facilities now running 150mm wafers wiU convert to 200mm wafers, further delaying the need for new equipment. DRAM-sensitive equipment technologies or capital-intensive segments such as steppers, implantation, diffusion, and polysiUcon etch wiU be affected more than logic-sensitive technologies such as sputtering, epitaxial reactors, CMP,

RTP, nontube chemical vapor deposition (CVD), or metal etch. The next expansion should begin by 1999, driven by 0.35-micron to 0.4-micron capacity expansion.

We have factored in an infrastructure investment in equipment for late

1997 through 1999, which will affect the forecast size of 4ie markets positively. This additional investment will be for initial equipment to fill a couple of 300mm fabs for running silicon by 1999. However, we believe that a significant 300mm equipment market will wait until well after 2000.

Yield enhancement is the trend of the time. Any system technology that can be priced relatively low and that has a direct impact on 5deld will gain immediate acceptance in volume. Areas emerging today as particularly important are in cleaning technology, photostabilizers, and process control metrology.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Equipment, Manufactiring. andMaterials Worldwide

The silicon wafer market, driven by a stronger long-term picture for semiconductors in general, will grow faster over the next six years than in the recent past. As the industry moves to a 200mm baseline, the outlook for silicon wafer manufacturers becomes brighter. The current sellers' market in silicon, which we expect to last for several years, vdll enable the industry to become healthy—a necessity for the semiconductor industry. Silicon manufacturers have answered the call for 200mm capacity, and the semiconductor market has again responded with a cry for more. We believe that the ramp plans of silicon manufacturers in 200mm have been strategically and smartly measured because the memory of the overcapacity of

1985 is still fresh. While there will be activity with 300mm wafers, this is expected to be focused on R&D and of low volume tintil after the turn of the decade.

Dataquest Perspective

Our forecast for capital spending and wafer fab equipment sales during the next six years assumes that ^ e explosive growth of 1995 will carry over into 1996. These sales are being driven by the PC market, with telecommimications and networking spurring demand for semiconductor chips across a broad spectrum—and with continued tight capacity is convincing companies to expand. Our outlook for the future includes moderated growth in equipment spending in 1997 and a slight decline in 1998 before a resumption of double-digit growth in 1999.

The semiconductor industry is a global manufacturing business. Production of semiconductors is constantly shifting among regions as new capital flows toward areas of relatively lower capital cost and higher consumption growth. Where the PC goes, so go semiconductors. This is true from the perspective of the business forecast as well as the production line.

Europe and Asia/Pacific, with very large capital spending upticks over the last several years, will continue to gain share in world production over the next few years.

The shifts and currents in semiconductor production trends mean that equipment and material suppliers will absolutely need a global presence, in every sense of the word, to remain competitive in the market. Product supply and support can no longer concentrate on local trends, because all major semiconductor companies have made it clear that they are investing worldwide. Silicon plants are now being strategically placed, such as Shinetsu Handotai's (SEH's) Malaysia plant and its recently announced joint venture in Taiwan, Komatsu's joint venture with Formosa Plastics in

Taiwan, and MEMC Electronic Materials' joint ventures in both Korea

(Posco-Huls) and Taiwan (Taisil).

Taiwan is clearly the new major production growth area. We would expect

Malaysia and Thailand to be the next major growth countries in three to five years. Evidence of this includes recent joint-venture fab annovmcements by Texas Instruments and others.

Further, the concept of contract manufacturing in semiconductors is clearly here to stay. Equipment and material suppliers could find themselves selling their technical products to an international team from several companies, including the manufacturer and the designer.

SEI\/IM-WW-l\/IT-9502 ©1996 Dataquest January 22,1996

Executive Summary

However, the emergence of the dedicated foundry company, taking ownership of the process and manufacturing flow, will tend to centralize this activity. Dataquest has started a research program in Semiconductor Contract Manufacturing, with a major report expected to be released by the end of January 1996. This report will explore the key trends in contract manufacturing and foundries, including technology trends and supply/ demand balance through the decade.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Chapter 2

Semiconductor Capital Spending Forecast

This chapter presents data on worldwide semiconductor capital spending by region. Capital spending in a region includes spending by all semiconductor producers with plants in that region. Components of capital spending are property, plant, and equipment expenditure for front- and backend semiconductor operations.

Chapter Highlights

This chapter wUl discuss the following highlights:

• On the heels of booming growth in 1994, global semiconductor capital spending grew another 72 percent during 1995 to $38 billion.

Anticipated continued tight capacity and a strong semiconductor market in 1995 mean continued growth into 1996, now forecast at about

30 percent.

• North America is showing consistent strength, with a 69 percent growth in 1995. Worldwide demand for desktop connectivity products and telecommunications equipment continues to fuel the investment strategies in U.S.-manufactured semiconductor products, heavily weighted toward logic and ASIC capacity. North American capital spending is expected to remain strong in 1996 and to moderate in 1997 as these investments are absorbed, but we expect the North American region to grow at faster-than-market rates as foreign multinationals and foundry companies invest in capacity in the United States.

• Japanese companies are continuing to invest in semiconductor capacity to preserve their market share position in memories, although the strengtii of the yen during the middle of the year temporarily put a lid on spending enthusiasm. Japan as a region kept pace with the world in investment in 1994 but lagged the market in 1995 as Japanese companies invested more outside Japan. Healthy, but subdued, growth of

47 percent in spending within Japan occurred during 1995—about

34 percent on a yen basis. Lagging investment patterns in Japan are expected to continue throughout the decade.

• Japanese companies, however, grew spending during 1995 at about

61 percent worldwide, for a total of $12.2 billion, second only to North

American companies' spending of $13.6 billion and well ahead of Asian companies' $9.1 billion.

• Dataquest has been bullish on the prospects in Europe, and remained so in 1995, although the region came in slightly under expected spending.

European companies are a large part of this expansion^ aided by strong domestic economies, and major projects by the multinational manufacturers are also contributing. We still see Europe as a significant growth region for spending through the decade.

• Following very strong capital investment growth of T7 percent in 1994, the Asia/Pacific-Rest of World (ROW) region grew an astronomical

104 percent in 1995, as Korean DRAM expansion accelerated (further), foundry expansion in Taiwan, Singapore, and others continued to grow, and new DRAM players entered tiKe scene in Taiwan (Powerchip

SEMM-WW-l\/IT-9502 ©1996 Dataquest 7

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Semiconductor, Vanguard, and Nan Ya Technology). These new projects started in 1995 will continue to consume capital funds in 1996. We are expecting about 31 percent growth in 1996. Asia/Pacific-ROW will continue to be one of the fastest-growing regions through this decade.

Earlier this year, we stated that Asia/Pacific would actually exceed

North American spending in this decade. During 1995, Asia/Pacific did surpass Japan, becoming the second-largest region for capital investment. However, we believe a renewed interest in setting up capacity in the United States, particularly by Asian companies, coupled with the low cost of capital will keep North American spending slightly ahead of

Asia/Pacific spending through this decade. Eventually, as more countries join the semiconductor manuJfacturing "club" (such as Malaysia and Thailand), Asia/Pacific will become the top region in investment.

Capital Spending Tabies

A list of the projected top 20 semiconductor capital spending companies in

1995 is presented in Table 2-1. Capital spending details by region are provided in two tables in this chapter: Table 2-2 shows historical semiconductor capital spending by region for 1987 through 1994, and Table 2-3 shows the capital spending forecast by region for 1994 through 2000. Yearly exchange rate variations can have a significant effect on the interpretation of the 1987-through-1995 data. For more information about the exchange rates used and their effects, see Appendix B.

And tlie Spending Binge Continues...

After a three-year rest, the growth cycle began again in a big way in 1993.

After a 23 percent growth in semiconductor capital spending in 1993, acceleration growth of 54 percent followed in 1994, and growth now peaks with a 72 percent increase worldwide during 1995, based on our most recent capital spending survey.

The continued growth in PC unit sales, with increased growth in telecommuiucations and networking products, has created tremendous demand for a variety of semiconductor components. The wafer fab capacity crunch has continued into all regions and most semiconductor products, most notably DRAMs and advanced ASICs. The capacity shortage has given rise to sharp acceleration in capital spending in all areas, with the strongest growth occurring in DRAM expansion in Asia/Pacific and Japan.

The big three Korean companies increased spending an unbelievable

124 percent to a combined total of $5.6 billion in 1995. A mostly new crowd of Taiwanese companies is now entering the DRAM manufacturing business, spending over $1 billion collectively in 1995. Japanese suppliers of memory are increasing investment, as well, for a collective increase of

• nearly 61 percent, to $12.2 billion, a larger dollar increase than the Korean companies. Intel and Motorola still head the list for 1995 as microprocessor and microcontroller demand continues to be strong. Equipping new and acquired facilities (in the case of Motorola) will continue to drive spending for these companies. The big three Korean companies, with their increase in DRAM capacity spending, now occupy the No. 3, No. 5, and No. 9 spots, and fully 13 companies are now part of the $1 billion spending club, with five near or over $2 billion. NEC, Fujitsu, Hitachi, Toshiba, and IBM

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Capital Spending Forecast

Table 2-1

Semiconductor Capital Spending—^Top 20 Spenders^ Comparison of 1994 and Projected

1995 Worldwide Capital Spending (Millions of U.S. Dollars)

16

17

18

19

20

9

10

11

12

13

14

15

1995

Rank

1

2

6

7

8

3

4

5

8

16

18

15

7

11

14

12

10

19

13

20

23

1994

Rank

1

2

9

3

5

6

4

Company

Intel

Motorola

LG Semicon

NEC

Samsung

Hitachi

Fujitsu

Toshiba

Hyimdai

IBM Microelectronics

Mitsubishi

Texas Instruments

Siemens AG

Micron Technology

Matsushita

SGS-Thomson

Philips

Advanced Micro Devices

Sanyo

National Semiconductor

Total Top 20 Companies

Total Worldwide Capital Spending

Top 20 Companies Percentage of Total

Source: Dataquest (December 1995)

Projected

1995

3,538.0

2,375.0

2,125.0

2,028.7

1,975.0

1,719.7

1,606.8

1,558.5

1,500.0

1,200.0

1,128.5

1,100.0

1,060.0

960.0

854.5

850.0

750.0

745.0

672.6

597.0

28,344.3

37,993.8

74.6

933.1

700.0

525.0

675.3

825.0

410.0

387.0

513.2

780.0

385.0

1994

2,419.0

1,640.0

800.0

1,117.3

1,000.0

969.9

1,072.6

625.0

356.1

325.0

16,458.5

22,036.5

74.7

Change(%)

46.3

44.8

165.6

81.6

97.5

77.3

49.8

67.0

114.3

128.6

67.1

33.3

158.5

148.1

66.5

9.0

94.8

19.2

88.9

83.7

72.2

72.4

SEMM-WW-MT-9502

©1996 Dataquest January 22,1996

10

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 2-2

Semiconductor Companies (Millions of U.

S. Dollars)

N o r t h America

Growth (%)

Japan

Growth (%)

Japan (Billions of Yen)

Growth (%)

Eiirope

1987

2,622

25.9

2,458

33.2

356

15.4

885

Growth (%)

Asia/Padfic-ROW

Growth (%)

Worldwide

Growth (%)

15.7

540

23.6

6,505

26.8

Source: Dataquest (December 1995)

1988

3349

27.7

4,495

82.9

584

64.0

960

8.5

1,033

91.3

9,837

51.2

1989

3,794

13.3

5,360

19.2

740

26.7

1,186

23.5

1,865

80.5

12,205

24.1

1990

4,217

11.1

5,596

4.4

806

8.9

1,560

31.5

1,542

-17.3

12,915

5.8

1991

3,895

-7.6

5,702

1.9

768

-4.7

1,248

-20.0

2,300

49.2

13,145

1.8

1992

4,135

6.2

3,958

-30.6

500

-34.9

1,188

-4.8

2,318

0.8

11,599

-11.8

1993

4,943

19.5

4,413

11.5

491

-1.9

1,729

45.5

3,238

39.7

14,323

23.5

1994

7,182

45.3

6,654

50.8

677

38.0

2,481

43.5

5,720

76.7

22,037

53.9

CAGR (%)

1989-1994

13.6

4.4

-1.8

15.9

25.1

12.5

Table 2-3

Worldwide Capital Spending by Region — Forecast,

1994-2000, Includes Merchant and

Captive Semiconductor Companies (Millions of U.S.

Dollars)

CAGR (%)

1994-2000

23.7

N o r t h America

Growth (%)

Japan

Growth (%)

1994

7,182

45.3

6,654

50.8

Japan (BiUions of Yen)

Growth (%)

Evirope

Growth (%)

677

38.0

2,481

43.5

Asia/Padfic-ROW

Growth (%)

Worldwide

5,720

76.7

22,037

Growth (%)

Source: Dataquest (December 1995)

53.9

1995

12,169

69.4

9,777

46.9

910

34.3

4,384

76.7

11,665

103.9

37,994

72.4

1996

16,579

36.2

12,102

23.8

1,192

31.1

5,293

20.7

15,308

31.2

49,281

29.7

1997

18,021

8.7

11,766

-2.8

1,159

-2.8

5,521

4.3

15,458

1.0

50,765

3.0

1998

17,954

-0.4

11,320

-3.8

1,115

-3.8

5,240

-5.1

14,927

-3.4

49,441

-2.6

1999

20,645

15.0

12,542

10.8

1,236

10.8

5,968

13.9

15,623

4.7

54,778

10.8

2000

25,722

24.6

15,986

27.5

1,575

27.5

7,402

24.0

20361

30.3

69,472

26.8

15.7

15.1

20.0

23.6

21.1

SEMIVI-WW-MT-9502

©1996 Dataquest January 22,1996

Semiconductor Capital Spending Forecast 11

Microelectronics all make this top 10, as the memory capacity keeps rolling in. In fact, the top spenders from No. 3 through No. 15 are all heavily concentrated in DRAM spending. With the general health of the industry, smaller semiconductor companies in all regions are participating in the capital spending boom in 1995, keeping the concentration of capital spending by the top 20 flat at 75 percent.

Taiwan Semiconductor Manufacturing Co. (TSMC) is on the cusp of debuting on the top 20 list for 1995 with an estimated $585 million spent on capacity. Foundry capacity expansion has now evolved into a major trend. In fact, foundry spending in Asia/Pacific will more than double in

1995 to an estimated $1.8 billion. This industry has been transformed into a dedicated, bona fide business and is no longer a specialized way to use excess capacity.

When Will the Spending Boom End?

Our longer-term forecast projects that significant growth in capital spending will spill into 1996 from sheer momentum, with a moderated growth of 30 percent, concentrated toward filling new fabs with equipment.

Thirty new 200mm fabs are planned to come on line in 1996, as well as nine 150mm and smaller fabs, for a total of 39. This compares with a total of 30 for 1995 (18 of which are 200mm). Dataquest believes that, as the end of 1996 approaches, capital spending wiU begin to decelerate as the capacity additions of 1994 through 1996 are ramped. From what we can see, there is plenty of equipment that could be brought to bear on 16Mb

DRAM capacity by the end of 1996 and to be on line to answer demand through 1997.

Overall semiconductor product demand is expected to remain strong through 1996, with long-term sustained growth through 2000 at a compound aimual growth rate (CAGR) of 20.1 percent (see Chapter 5). We expect that microcomponent capacity will continue to ramp through 1997 before pausing in 1998, with the next major investment cycle picking up in

1999. Our model does not currentiy include significantiy more 16Mb

DRAM capacity expansion (over capacity being put in place in 1995 and

1996) until 1999. In the two "pause" years of 1997 and 1998, we believe

DRAM manufacturers will concentrate on converting capacity from

4Mb toward 16Mb, which increases bits per square inch processed, and then concentrate on shrinks to squeeze out value per square inch before a capital cycle starts again. Further, in Japan, we expect that many 4Mb/

16Mb fabs now running 150mm wafers will convert to 200mm wafers, further gaining efficiency and productivity from the capital investment made.

Through 2000, we project a five-year worldwide capital spending CAGR of 21.1 percent, slightly ahead of semiconductor consumption growth. We believe that capital spending may be influenced positively in 1997 and

1998 with facility construction and purchase of equipment for the world's

. first 300mm wafer fab. We have included this infrastructure investment in our model.

Over a year ago, we introduced a model that quantifies the over- and underinvestment picture for wafer fab equipment and semiconductor capacity. Although activity of the last several years has created and sustained a net underinvestment, not fully corrected to overinvestment in

1994, we expect about a 35 percent overinvestment by the end of 1996 (see

Chapter 3 and Figures 3-1 and 3-2).

SEI\/lM-WW-l\/IT-9502 ©1996 Dataquest January 22,1996

12 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Is There Excessive Spending and impending Doom?

With capital spending as a percentage of semiconductor revenue likely to exceed 25 percent in 1996, one question is being asked often. Does this spending level point to impending danger for equipment suppliers and a spending crash? We think not.

The industry normally cycles through over- and undercapacity, because there is an inherent lag time between capital investment decisions and productive capacity. This cycle will never disappear. But the levels of spending we are experiencing today have not been seen since 1984, and many people are pointing to the 1985 downturn as an example of what is imminent today.

We think the industry is structured differently today and that there is a valid reason why investment as a percentage of revenue is increasing.

We would split the market into four periods: before 1985,1985 to 1992,

1993 to 2001 (?), and 2002 and after Let's describe the conditions and trends in the production market and manufacturing infrastructure during these periods.

Before 1985: An Immature Manufacturing Industry

Characteristics of this period are:

• The semiconductor nianufacturing infrastructure was fairly immature, characterized by large integrated systems companies, mostly in the

United States.

• Manufacturing technology was favored over efficient use of capital, and device performance was favored over yield.

• Capital equipment manufacturers did not assume complete ownership of processes and system performance.

• All of these led to a capital-spending ratio between 26 percent and

30 percent of revenue.

Two things happened that changed the structure. First, the semiconductor downturn in 1985 ended up being extremely bloody. Second, Japanese producers emerged that introduced an element of true manufacturing efficiency into the infrastructure.

1985 Through 1992: Becoming IVIanufacturing Smart

The Japanese manufacturing ethic and the losses incurred during the 1985 downturn introduced the need for the industry to become more efficient in the manufacturing infrastructure. The following things happened:

• Manufacturing productivity and yield became a focus.

• SEMATECH was formed, in part as a result of the need to coordinate this effort in the United States.

• Equipment companies were expected to take ownership of the process and system performance parameters, and they accepted.

• Equipment performance and productivity increased substantially.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Capital Spending Forecast 13

• Fab factory automation migrated from robotics to computer-controlled systems, and statistical process control became commonplace.

• The emergence of Korean companies' manufacturing power provided momentum for this transition.

By 1992, manufacturers had increased yields and fab productivity was up dramatically. During this period, it was natural for the industry to have to spend less on capital equipment, since the return per dollar spent was very high. Capital spending decreased to an artificially low 18 percent of revenue, on average.

Once yields achieved high levels and system productivity approached the point at which it was impractical to continue (to do so would have meant decreased equipment utilization because throughput could exceed practical run rates), ttie return from these activities diminished. The industry then entered the next period.

1993 through 2001: Growth

Once the industry had built an efficient manufacturing infrastructure, it became ready for the emergence of the semiconductor as the enabling technology in many electronic systems. The industry entered into an era of prosperity. Semiconductors have become the productivity engine for the world's business, implementing communications systems and the power of the PC to improve worker efficiency. We are experiencing the following characteristics:

• Unit growth in semiconductors has required manufacturers to invest in capacity for growth.

• Profitability has attracted new entrants with a focus on manufacturing.

• Dedicated contract manufacturing has emerged as a new manufacturing model for the industry, made possible by equipment efficiencies and the need to separate manufacturing from device innovation.

When the industry reached a point of diminishing returns from yield and equipment productivity, the capital spending ratio could no longer be maintained at artificially low levels. We believe that the current equilibrium level is about 22 percent of semiconductor revenue.

Further, fueled by the growth in the device market, the current boom has been unprecedented in the industry in terms of length and levels of growth in capital spending. This has resulted from the industry's adjustment to a new, higher spending level and the increased imit demand.

Beyond 2001: Maturing Manufacturing Infrastructure

We believe that the emergence of dedicated contract manufacturing means that the industry's manufacturing infrastructure will evolve to a point at which the foundry becomes an integral part of the manufacturing environment, as been the case with electronic equipment in general (Solectron is an example).

The business model of the foundry requires high equipment utilization, and we expect that, in the next five years, this will influence capital efficiency and decrease the capital spending ratio to perhaps 20 percent or

SEI\/ll\/l-WW-MT-9502 . ©1996 Dataquest January 22,1996

14 Semiconductor Equipment, Manufacturing, and Materials Worldwide

21 percent of revenue. This will not become evident until contract manufacturing increases in scope above the estimated 9 percent of semiconductor production today, to levels that could approach 35 percent to 45 percent by the year 2010.

The North American l\1arlcet Continues to Exhibit Strategic Strength

Capital spending in North America grew at an accelerated 69 percent in

1995, witii most of the investment growth coming from U.S. companies connected with ASIC and logic products. We expect capital spending to decelerate gradually through 1998, v/ith acceleration resuming in 1999, resulting in a CAGR of 23.7 percent for 1994 through 2000. This is among the fastest-growing regions for investment, driven by recent low cost of capital and the need for foreign multinational and foundry manufacturers to be closer to their customers.

This relatively strong growth in capital spending has been driven by growth in PCs, telecommvmications, and networking. These products have seen increasing use as tools to increase productivity in the workplace.

These electronic products, with their increased semiconductor content, have created enormous demand for microprocessors, microcontrollers,

SRAM, programmable logic and memory, standard logic, and peripherals controllers. U.S. companies dominate many of these market segments.

These segments combined are expected to maintain fairly stable growth rates over the next few years, witii PC growth slowing (but still maintaining a CAGR of 17 percent) and networking and telecommunications expanding. The near-term market for PCs has reaccelerated from Intel's new, aggressive Pentium pricing strategy, which has hastened the conversion to ttie Pentium.

New products and services such as personal cormnimicators, interactive television, and video-on-demand provide the potential for enormous growth in semiconductor sales, especially for the highly integrated complex logic and signal processing chips that will be the core engines of future systems.

Although the strategic strength of the core logic products makes for a healthy and flourishing semiconductor production environment, it also means less volatility in capital spending. In the boom years of 1994 and

1995, the North American region grew at somewhat lower-than-market rates. This will enable the North American capital spending to grow at faster-than-market rates in slower years like 1997 and 1998. We believe companies will invest strategically in capacity to preserve competitive advantage, and cutbacks are more likely to occur in the smaller companies rather than the first-tier manufacturers.

Capital investments in North America for 1995 came from equipping new fabs by both the majors and smaller companies. The major projects include

Intel's Fab 11.2 in New Mexico and expected orders for Fab 12 in Arizona,

Advanced Micro Devices' (AMD) Fab 25 in Austin, Texas, Motorola's

MOS-13 in Texas and the continued ramp of MOS-12 in Arizona, Cypress'

Fab 4 in Minnesota, SGS-Thomson's new Arizona facility, IBM's expansion in Burlington, Vermont (yes, IBM is back), and purchases for Texas Instruments' DMOS-5 fab in Dallas, Texas. Smaller companies are also spurring

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Capital Spending Forecast 15 growth this year, with Integrated Device Technology, VLSI Technology,

Zilog, Atmel, International Rectifier, American Microsystems Inc. (AMI), and National Semiconductor all bringing on new capacity. In 1996, equipment installation into these fabs will continue, plus 10 new fabs under construction, such as those of Sony and VLSI Technology in San Antonio,

Texas, TwinStar in Texas, and Motorola's MOS-21 in Arizona. Samsung has stated its intention of investing $1 bilUon to build a fab in the United

States. We have factored this and other foreign company fabs, such as

TSMC's joint venture foimdry and Toshiba's fab with IBM, into our 1997 to

1998 forecast.

Japan: DRAI\/I Capacity Additions Drive Spending, but a Strong Yen Subdues

Japan's 47 percent increase in capital spending in 1995 is only a 34 percent increase on a yen basis, and Japanese companies look to invest outside

Japan to optimize buying power. We are forecasting a subdued 24 percent growth in capital spending for 1996, factoring in a slight decline in 1997 as

&ie nussion will have been accomplished in DRAMs in the near term.

Some of the Japanese electronics giants are experiencing good profit growth, driven by semiconductor operations. The demand for world memory capacity presented an opportunity to increase profits from semiconductors. Investments by Japanese companies grew by nearly 61 percent in 1995, with an increased amount going overseas. However, as long as the Japanese economy is imder pressure, Japanese companies will feel a

"patriotic" dedication to invest in Japan, and we see no company spending more than 35 percent of committed investnient outside Japan. Witii the strength of the yen, multinationals are reducing their investment proportion inside Japan as well.

Although new facilities by Japanese companies will come on line outside Japan throughout the rest of this decade, DRAM investments inside Japan are really the only driving force today. Beyond 1995, investment increases in Japan wiU need to come from growing the domestic economy. Dataquest believes an economic recovery in Japan started in

1994, but with slow acceleration. The degree at which companies will invest will be affected by the strength of this recovery. We are forecasting a below-average CAGR of 15.7 percent in Japan for 1994 through 2000.

One bright spot is that a PC boom could emerge in Japan over the next year or two, spawned by the networking infrastructure that is cxirrently being built. This would breathe new life in the Japanese semiconductor market, and our forecast would brighten a bit. We do not think that even a

PC boom, however, would create a forecast better than several percentage points below the world average. The fundamentals of Japanese production capacity are still too heavily concentrated in DRAMs, with no clear future direction emerging yet, which keeps us from being more optimistic about capital activities in Japan.

SEMIVI-WW-l\/IT-9502 ©1996 Dataquest January 22.1996

j6 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Europe Sustains Presence as a Growth Market

After a higher-than-expected growth bubble of 46 percent in 1993,

European spending "moderated" to a slower-than-market growth in 1994 as multinationals (such as Intel) substantially completed the majority of their expansions in 1993. The growth of 43 percent in 1994 is nonetheless extremely healthy, fueled primarily by European companies themselves— the ever-present SGS-Thomson, Philips expanding in Nijmegen, Netherlands, and Ericsson equipping its expansion.

Europe continues to attract the capital in 1995, growing an impressive

76 percent. Large multinationals are still present, with Motorola upgrading the Scotland fab bought from Cttgital, the new IBM-Philips venture in

Germany, Analog Devices expanding in Ireland, Texas Instruments expanding again in Italy, and the IBM-Siemens fab in France continuing to ramp 16Mb DRAMs. The key expansion is Siemens' new fab in Dresden,

Germany, the key driver pulling Siemens into the top 10 in capital spending worldwide. The continued commitment of NEC and other Japanese companies to Europe has been the most recent boost to investment momentum, focused on land and facilities in 1995, bringing the growth in capital spending significantiy above the growth in wafer fab equipment.

We are looking for continued growth in 1996 of 21 percent as production continues to ramp from these investments and eight new fabs, most notably by GEC-Plessey, NEC, and Temic.

Europe is to be viewed as a strategic location for production to take better advantage of European and 16Mb DRAM growth in the future, driven by the PC production boom (see Chapter 6), while avoiding import tariffs.

Samsung has announced a fab to come on line in Europe during 1998, but is still undecided about the exact location.

With a stronger multinational presence starting again in 1995 as economies pick up and with recent trends for PC production and foimdry providers

(such as Newport WaferFab arid Tower Semiconductor), we now expect

Europe to be an average investment region in the long term, with a sixyear CAGR of 20 percent.

Asia/Pacific Is Madly Investing in IWo Distinct Ways

The often-erratic but sustained semiconductor capital spending growth in the Asia/Pacific region continued at the explosive rate of 77 percent in

1994. And those who thought this market could not accelerate from that level should think again—^for 1995 is the year in which the Asia/Pacific region became the second-largest exj^ansion region in the world, surpassing Japan in terms of dollars spent (tliat is, spending within the region, not by comparvies based in Asia). The region saw 104 percent growth in 1995, and we expect moderated growth of about 31 percent in 1996 as several new fab projects continue to be built and equipped and the number of new projects grows. Longer term, we expect Asia/Pacific's growth in capital spending to be among the most aggressive of any region. Dataquest forecasts a CAGR of 23.6 percent for 1994 through 2000.

Spending in 1995 is focused primarily in two areas: DRAMs and foundry capacity. The Korean conglomerates plan to continue their relentless

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Capital Spending Forecast 17

DRAM capacity expansion in 1995. Hyimdai is installing equipment for its new E-3 project, the third phase of a 200mm wafer, 16Mb DRAM fab started in 1992. LG Semicon (formerly Goldstar) is expanding its equivalent C2 line, in accordance with its agreement with Hitachi for the 16Mb

DRAM ramps. LG Semicon is also bringing online the G-FAB for non-

DRAM memory products. Samsung is continuing to spend, ramping Line

6, also a 200mm, 16Mb DRAM facility

The real story in 1995 is the new Taiwan players. Vanguard International will be bringing on its new DRAM fab later this year, and Powerchip and

Nan Ya have broken groimd on DRAM fabs coming online in 1996. All of these are targeted at 16Mb DRAM running 200mm wafers. Current players such as TI/Acer and MOSel-Vitelic are also increasing their spending with new projects.

Taiwan chip companies TSMC, Macronix, Winbond, and United Microelectronics Corporation (UMC), along with Chartered Semiconductor in

Singapore, have undertaken major projects in expanding foundry capacity, and a new foundry player has emerged in Thedland—SubMicron Technology. SubMicron has received funding of $1.6 billion for two separate fab lines, the first to come on line in 1996. Funding has also been allocated to establish a technology park in or near Bangkok, similar to Hsinchu in Taiwan. We have seen the start of this Bangkok park become reality as a new joint venture that includes Texas Instruments has just announced commitments for a new fab there. The combined spending of these companies for foundry (which excludes spending associated wiiJi their own products in the case of UMC, Macronix, and Winbond) increased from about $900 million in 1994 to about $1.8 billion in 1995, continuing at significantly higher levels into 1996 and 1997. The driving reason is the changing face of contract manufacturing in semiconductors. Gone are the days where excess fab capacity could support the foundry business of fabless companies (as well as integrated device manufacturers, or IDMs—companies with fabs).

Dataquest estimates that only about 32 percent of the contracted manufacturing of semiconductors originates from fabless companies. The remainder is from IDMs that wish to place lower manufacturing-value-added products away from their own facilities in order to maximize resources and cost, to reduce investment risks using foimdries as an extension of their own capacity, or to use the more advanced technology of some foundries as a growth strategy. The last few years have seen the dedicated foundry flourish, mostly in Asia/Pacific. It is still believed that the largest concentration of foundry capacity in the world, however, is in Japan, with companies like Rohm, Seiko-Epson, Sharp, and other large integrated companies.

However, the appetite for leading-edge foundries has caused another transformation to occur. With the cost of capital increasing and investment at a higher level for leading-edge equipment, foundry companies such as

Chartered have established longer-term contracts with design companies, often with capital infusions for production equipment. Many joint ventures have been announced in the last several months, and we expect this trend to continue throughout 1996.

SEi\/ll\/l-WW-i\yiT-9502 ©1996 Dataquest January 22,1996

_18 Semiconductor Equipment, Manufacturing, and Materials Worldwide

The foundry industry is now a strategic industry rather than simply a tactical one. With this transformation nearly complete, we are starting to see dedicated investment to build new foundry capacity.

In addition to the established semiconductor-producing countries, huge long-term opportunities exist in developing countries like China and tiie

Unified States (formerly the Soviet Union). Ultimate demand for semiconductor products in those countries could approach demand in super-consumer countries like the United States and Japan. China, in particular, generates a gross domestic product comparable to that of Japan, if evaluated on the basis of purcliasing power parity. U.S., European, and Japanese telecommunications companies are working with the Chinese government to install telephone exchjinge equipment and digital lines.

Several hurdles must be overcome before either China or Russia becomes a viable market for advanced front-end semiconductor manufacturing.

Technology export restrictions must be eased to allow the construction of relatively advanced fabrication facilities. Foreign suppliers must establish local sales and service centers and define their market access. Financing capability must be established by the host countries. Solidification of international trade relationships tiuough participation in the General Agreement on Tariffs and Trade (GATT) must also be established. China's internal political structure poses a potential barrier to maintaining its status as a most favored nation with the United States. It will likely take a few years to sort out these issues. Dataquest assumes that semiconductor investment in China could begin to expand in 1997 (NEC is leading the investment charge today), accelerating into the later half of the decade.

Dataquest believes that the next countries to experience huge front-end semiconductor production growth will be Malaysia and Thailand. The latter's plans for a new science park we have already mentioned, and a new fab has been announced in Malaysia (Sarawak), financed by a group of investors, to start production in 1997.

Who's investing Wliere?

Dataquest's recent capital spending survey shows how money is being spent. Table 2-4 summarizes how companies based in different regions are spending their money abroad for 1994, and Table 2-5 svunmarizes this for

1995. About 78 percent of money spent goes into the domestic economy worldwide, a ratio that held steady for 1994 and 1995.

Asia/Pacific companies have historically invested domestically, but diversification began in 1994 and continued in 1995. About 96 percent of Asia/

Pacific companies' spending was domestic in 1994, and this proportion held steady in 1995. We would expect this ratio to decrease significantly over the next two or three years. Europeans have been the most aggressive capital exporters, historically, making only 63 percent of their investments within Europe. This figure has grown slightly to 64 percent in 1995 and should expand in 1996 as the domestic economies have resurged.

Japanese companies are very close to the worldwide average, with about

80 percent domestic investment in 1994, dropping to 75 percent in 1995.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Capital Spending Forecast

19

Table 2-4

Regional Investment Patterns of Semiconductor Manufacturers in 1994

(Millions of U.S. Dollars)

North American Companies

Japanese Companies

European Companies

Asia/Padfic-ROW Companies

All Comparues

Growth from 1993 (%)

Source: Dataquest (January 1996)

Worldwide

8,628.3

7,587.4

1,866.0

3,954.7

22,036.5

53.8

North

America

6,223.7

545.7

277.6

135.0

7,182.0

45.3

Japan

566.3

6,087.5

0.4

0

6,654.3

50.8

Europe

957.4

311.5

1,182.7

30.0

2,481.5

43.5

Asia/

Pacific-

ROW

880.9

643.7

405.3

3,789.7

5,719.6

76.6

Percentage of World

Spending

39.2

34.4

8.5

17.9

100.0

Table 2-5

Regional Investment Patterns of Semiconductor Manufacturers in 1995

(Millions of U.S. Dollars)

North American Comparues

Japanese Companies

European Companies

Asia/Padfic-ROW Companies

All Companies

Growth from 1993 (%)

Source: Dataquest (January 1996)

Worldwide

13,616.6

12,194.2

3,112.0

9,071.0

37,993.8

72.4

North

America

9,963.6

1,390.7

466.9

347.5

12,168.7

69.4

Japan

588.4

9,187.8

0.6

0

9,776.8

46.9

Europe

1,525.0

835.3

1,984.1

39.5

4,383.8

76.7

Asia/

Pacific-

ROW

1,539.6

780.4

660.4

8,684.0

11,664.5

103.9

Percentage of World

Spending

35.8

32.1

8.2

23.9

100.0

North American companies are also domestic spenders, with about 72 and

73 percent of investment staying at home for 1994 and 1995, respectively.

The North America and Japan regions are net investors, while the Europe and Asia/Pacific regions are net beneficiaries of that investment. This parallels the status of these regions as net exporters and net importers of semiconductors, respectively.

Although all regions are spending in Asia/Pacific and all mtiltinational regions are investing in Europe, only North American companies have the strategic vision to invest in Japan. Japanese companies are also investing on a worldwide basis. We believe this is one of the key elements necessary for a semiconductor company to be competitive on a global basis.

SEMI\/!-WW-MT-9502

©1996 Dataquest January 22,1996

20 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Dataquest Perspective

Capital spending in 1994 exploded and accelerated in 1995. The major reason for this is the surprisingly persistent growth in xmit PC shipments, with the aggressiveness of Intel's Pentium pricing strategy. Major DRAM expansion accelerated in the second half of 1993 and wiU continue throughout 1996. From what we can see now, there is plenty of equipment that could be brought to bear on 16Mb DRAM capacity by the end of 1996 and brought on line to answer demand through 1997. A marked downturn in the DRAM investment cycle will be triggered by the lMbxl6 configuration of the 16Mb DRAM, achieving yield in the area of 60 percent to

65 percent, which is expected to occur sometime in 1996.

Desktop connectivity products, telecommunications, and the PC market will lead to stable growth in microcomponents and logic devices, giving strategic strength to the North America region. Japanese companies will concentrate on ramping memories in order to hold their market share against Korean and Taiwanese companies. A struggling economy will keep capital investment muted once the DRAM ramp is satisfied. Globalization strategies will benefit both European and Asia/Pacific investment, with Asia/Pacific being the fastest-growing region over the next five years.

Dataquest believes that the relatively strong capacity expansion phase of

1993 to 1995 (with three-year growth of 265 percent) has now exceeded the three-year growth recorded in the 1987 to 1989 expansion. It should be noted, however, that these periods are different in two key respects. The current accelerated long-term growth for the underlying semiconductor industry is driven by a productivity-related PC boom. This PC boom is expected to continue, so we are not overly alarmed about the magnitude of this cycle. The momentum of investments will make 1996 a year of healthy growth in capital spending. Second, the manufacturing infrastructure is more efficient today, and there is a diminishing return in productivity and yield improvements. This has led to a higher riatural capital investment ratio being required today than in 1987 to 1989.

However, we also believe that spending will decelerate starting in the second half of 1996, causing a somewhat flat spending pattern through 1997 and 1998. Although we continue to believe that the cyclical nature of investment in semiconductor capacity will diminish, the PC boom must continue to drive the underlying semiconductor growth strongly enough to dampen the memory component of the cycle. After a flat two-year period, investments should pick up again in 1999.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Chapter 3

Wafer Fab Equipment Forecast.

This chapter presents data on worldwide spending by region for wafer fabrication equipment. Wafer fab equipment spending in a region includes spending by all semiconductor producers with plants in that region.

Included are all classifications of equipment for front-end semiconductor operations.

Chapter Highlights

This chapter will discuss the following highlights:

• Wafer fab equipment spending growth in 1995 exceeded 1994, with over

66 percent growth worldwide, compared to 56 percent growth in 1994.

• The growth was driven across all regions, led by Asia/Pacific (80 percent growth). Continued investment in North America and a DRAMsensitive Japan (68 and 61 percent growth, respectively) drove the volume for this third year of tiie boom. Europe, with an emphasis on land and facilities spending in 1995, "lagged" the market in consumption of wafer fab equipment (54 percent growth).

• Momentum and backlogs from 1995 and the 39 new fabs coming on line in 1996 will keep growth intact, and Dataquest is now forecasting a healthy 36 percent growth in wafer fab equipment shipments in 1996.

• Segment growth, beginning in 1994 and continuing in 1995, is being led by DRAM and capital spending-sensitive equipment, with steppers, implant, wafer inspection, and factory automation equipment exhibiting significantly stronger-than-market growth. We expect no major segment declines in 1996, as capacity additions wiU continue to be broadbased and worldwide.

• New equipment technologies grew exceptionally well in 1995, led by high-voltage implant, CMP, RTP, and nontube reactor CVD. The first three of these segments at least doubled, with the CVD segment growing 89 percent.

• Our model, which measures the net cumulative under- or overinvestment, indicates that, by the end of 1995, the semiconductor manufacturing world will be overinvested in wafer fab equipment by $5.8 billion, or 32.7 percent of the market. This is above the peaks exhibited in 1984 and 1989, so excess capacity should emerge by tiie end of 1996, probably in the DRAM market, where capacity has been added recently. This is supported by fundamental silicon square-inch area analysis completely recently.

• Capacity constraints remain in the foundry market into 1998, leading to strong logic capacity investment in 1997.

• After strong expansion years in 1993 through 1996, eqviipment purchases in 1997 should slow markedly, followed by a slight decline in 1998. The next expansion should kick in by 1999, driven by a

0.35-micron to 0.4-micron capacity expansion. The worldwide wafer fab equipment market is forecast to grow at a 21.4 percent CAGR between

1994 and 2000, slightly above the semiconductor market growth.

SEMM-WW-MT-9502 ©1996 Dataquest 21

22 Semiconductor Equipment. IVtanufactjring, and Materials Worldwide

• We have factored in an infrastructure investment in equipment for 1997 through 1999, which will affect the forecast size of the markets positively. This additional investment will be for initial equipment to fill some 300mm fabs for rimning silicon by 1999. The bulk of this "300mm equipment bubble" occurs in 1998. However, we believe that a significant 300mm equipment market will not be seen until 2001 or 2002.

This chapter presents historical and forecast data on the worldwide wafer fabrication equipment market, by region and by key equipment segment.

In this year-end forecast for wafer fab eqtiipment, we have chosen to focus our forecast of equipment categories on the following specific segments and issues:

• The annual investment theme for 1995 to 2000

• Steppers and automatic photoresist processing equipment (track) in lithography

• Dry etch and CMP in etch and clean

• Silicon epitaxy, CVD, and physical vapor deposition (PVD) in deposition

• Diffusion and RTP

• Ion implantation (medium-current, high-current, and high-voltage)

• Segments of emerging importance

These segments of the equipment market not only represent the majority of all wafer fab equipment expenditure in the world today, but also embody the key technological capability for advanced device production.

Highlights of some of the factors affecting individual equipment segment forecasts also are presented.

Equipment spending in a region refers to spending by all companies— both domestic and foreign—within the region. We note also that yearly exchange rate variations can have a significant effect on 1988-through-

1995 data appearing in the tables in this chapter. Appendix B details the exchange rates used in this document.

Table 3-1 provides historical market data, by geographic region, for 1988 through 1994. Table 3-2 shows forecast market data, by geographic region, for 1994 through 2000. Table 3-3 presents historical data for key equipment segments for 1988 through 1994. Table 3-4 shows forecast data for key equipment segment for 1994 through 2000.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Wafer Fab Equipment Forecast

23

Table 3-1

Worldwide Wafer Fab Equipment Market, by Region—Historical, 1988-1994

(Millions of Dollars)

North America

Chaiige (%)

Japan

Change (%)

Europe

Change (%)

Asia/Pacific

1988

1,534

38.9

2,270

77.8

662

25.9

519

1989

1,657

8.0

2,813

23.9

721

8.9

820

1990

1,589

-4.1

2,992

6.4

764

6.0

522

Change (%)

Total Wafer Fab Equipment

126.6

4,984

58.0

6,011

-36.3

5,867

Change (%)

58.9 20.6 -2.4

Note: Some columns do not add to totals shown because of rounding.

Source: Dataquest (January 1996)

1991

1,524

-4.1

3,007

0.5

641

-16.1

843

61.5

6,014

2.5

1992

1,570

3.0

2,096

-30.3

634

-1.1

789

-6.4

5,089

-15.4

1993

2,129

35.6

2,460

17.4

978

54.3

1,309

65.9

6,876

35.1

1994

3,141

CAGR (%)

1988-1994

12.7

47.5

3,668

49.1

1,385

41.6

2,562

95.7

10,755

56.4

8.3

13.1

30.5

13.7

Table 3-2

Worldwide Wafer Fab Equipment Market, by Region—Forecast, 1994-2000

(Millions of Dollars)

North America

1994

3,141

1995

5,262

1996

7,512

Change (%)

Japan

Change (%)

Europe

47.5

3,668

49.1

1,385

41.6

67.5

5,901

60.9

2,130

53.8

42.8

7,094

20.2

2,733

28.3

Change (%)

Asia/Pacific

Change (%)

Total Wafer Fab Equipment

Change (%)

2,562

95.7

10,755

56.4

4,618

80.2

17,911

66.5

6,998

51.5

24,338

35.9

Note: Some columns do not add to totals shown because of rounding.

Source: Dataquest (January 1996)

1997

7,969

6.1

6,445

-9.1

2,590

-5.3

7,705

10.1

24,709

1.5

1998

7,855

-1.4

6,082

-5.6

2,503

-3.3

6,852

-11.1

23,293

-5.7

1999

9,281

18.2

7,073

16.3

2,927

16.9

7,261

6.0

26,542

14.0

2000

11,723

26.3

9,055

28.0

3,655

24.9

9,931

36.8

34,365

29.5

CAGR (%)

1994-2000

24.5

16.3

17.6

25.3

21.4

SEMM-WW-MT-9502

©1996 Dataquest

January 22,1996

24

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 3-3

Wafer Fab Equipment Revenue by Equipment Segment-

(Millions of Dollars)

-Historical, 1988-1994

Equipment Segment

Worldwide Fab Equipment

Change (%)

Steppers

Track

Maskmaking Lithography

Other Lithography^

Total Lithography/Track

Automated Wet Stations

Other Clean Process

Dry Etch

Dry Strip

Chemical Mechanical

Polishing

Total Etch and Clean

Tube CVD

Nontube Reactor CVD

Sputtering

Silicon Epitaxy

Other Deposition^

Total Deposition

Difhision

RTF

Total Diffusion/RTP

Medium-Current Implant

High-Current Implant

High-Voltage Implant

Total Ion Implantation

Total Process ControP

1988

4,984

58.9

921

253

62

245

1,481

144

133

533

100

NS

911

186

275

260

86

165

972

296

22

318

118

241

18

377

608

1989

6,011

20.6

1,181

322

69

192

1,764

243

134

670

121

NS

1,168

220

388

320

75

170

1,173

332

25

357

131

301

25

457

676

1990

5,867

-2.4

1,052

317

47

195

1,612

268

132

690

118

NS

1,208

259

457

359

68

153

1,296

325

33

358

114

250

7

370

605

1991

6,014

2.5

979

364

48

158

1,549

291

143

717

119

11

1,281

268

474

425

89

147

1,403

326

46

372

108

228

18

353

643

1992

5,089

-15.4

646

353

53

95

1,147

286

103

682

123

20

1,212

213

437

446

84

119

1,300

246

36

283

83

164

16

263

544

1993

6,876

35.1

1,014

507

52

110

1,683

285

198

1,096

138

44

1,761

283

585

584

83

115

1,650

342

45

388

108

233

18

359

634

1994

10,755

56.4

CAGR (%)

1988-1994

13.7

1,838

711

12.2

18.8

78

112

2,739

519

233

1,555

202

65

2,573

446

903

1,025

100

101

2,575

492

80

572

234

391

29

654

1,017

Factory Automation

Other Equipment

Total FA/Other

Equipment

Total Wafer Fab

Equipment

130

187

317

4,984

195

222

417

6,011

216

202

418

5,867

227

185

412

6,014

194

146

340

5,089

250

151

401

6,876

NS = Not surveyed

NA = Not applicable

'Includes contact/proximity, projection aligners, direct-write, and X-ray lithography

^Includes evaporation, MOCVD, and MBE

^Includes optical CD, CD SEM, wafer inspection, thin film measurement, and other process-control equipment

Note: Some columns do not add to totals shown because of rounding.

Source: Dataquest (January 1996)

412

213

625

10,755

3.8

-12.2

10.8

23.8

9.8

19.5

12.4

NA

18.9

15.7

21.9

25.7

2.7

-7.8

17.6

8.9

23.6

10.3

12.1

8.4

8.1

9.6

8.9

21.2

2.2

12.0

13.7

SEMM-WW-MT-9502

©1996 Dataquest

January 22,1996

Wafer Fab Equipment Forecast 25

Table 3-4

Wafer Fab Equipment Revenue by Equipment Segment—Forecast, 1994-2000

(Millions of Dollars)

Equipment Segment

Worldwide Fab Equipment

Change (%)

Steppers

Track

Maskmaking Lithography

Other Lithography^

Total Lithography/Track

Automated Wet Stations

Other Clean Process

Dry Etch

Dry Strip

Chemical Mechanical

Polishing

Total Etch and Clean

Tube CVD

Nontube Reactor CVD

Sputtering

Silicon Epitaxy

Other Deposition^

Total Deposition

Diffusion

RTP

Total Diffusion/RTP

Medium-Current Implant

High-Ctirrent Implant

High-Voltage Implant

Total Ion Implantation

Total Process Control^

Factory Automation

Other Equipment

Total FA/Other

Equipment

Total Wafer Fab

Equipment

1994

10,755

56.4

1,838

711

78

112

2,739

519

233

1,555

202

65

2,573

446

903

1,025

100

101

2,575

492

80

572

234

391

29

654

1,017

412

213

625

10,755

1995

17,911

66.5

3,274

1,214

104

110

4,701

824

387

2,507

336

152

4,205

623

1,703

1,562

157

120

4,164

674

168

842

424

670

122

1,216

1,742

740

300

1,040

17,911

1996

24,337

35.9

4,515

1,662

153

124

6,454

1,144

543

3,339

443

268

5,736

803

2,371

2,040

222

146

5,581

893

260

1,154

535

961

178

1,674

2,356

1,020

363

1,382

24337

1997

24,709

1.5

3,874

1,643

155

120

5,793

1,251

580

3,527

445

338

6,140

863

2,584

2,184

258

138

6,025

975

310

1,286

485

880

148

1,513

2,464

1,088

400

1,488

24,709

1998

23,293

-5.7

?>,3y7

1,468

161

113

5,079

1,239

555

3,422

432

366

6,014

814

2,511

2,131

271

123

5,851

897

364

1,260

397

755

146

1,298

2,356

1,022

413

1,435

23,293

1999

26,542

13.9

4,122

1,842

181

122

6,267

1,340

571

3,769

499

465

6,644

956

2,800

2,346

300

122

6,524

1,048

393

1,441

411

823

186

1,420

2,636

1,160

451

1,611

26,542

1 includes contact/proximity, projection aligners, direct-write, and X-ray lithography

^Includes evaporation, MOCVD, and MBE

^Includes optical CD, CD SEM, wafer inspection, thin film measurement, and other process-control equipment

Note: Some columns do not add to totals shown because of rounding.

Source: Dataquest (January 1996)

2000

34,365

29.5

5,605

2,474

227

134

8,440

CAGR (%)

1994-2000

21.4

20.4

23.1

19.6

3.0

20.6

1,632

691

4,811

622

684

21.1

19.8

20.7

20.6

48.1

8,440

1,220

3,653

2,986

351

134

8,344

1,323

471

1,794

543

1,168

258

1,969

3,330

1,509

540

2,048

34,365

21.6

17.9

34.5

21.0

15.1

20.0

43.7

20.2

21.9

24.1

16.8

21.9

21.9

18.2

26.2

19.5

23.2

4.9

21.4

SEMM-WW-MT-9502 ©1996 Dataquest

January 22,1996

26

Semiconductor Equipment, Manufacturing, and l\/lateriais Worldwide

Annual Investment Themes for 1994-2000

Behind our equipment and segment forecasts are assumptions about how semiconductor producers will perform and invest. These are summarized in Table 3-5 for 1995 through 2000. The following areas are considered: the availability of profits for reinvestment, memory versus logic growth, technology shifts, and brick and mortar versus equipment purchases.

Table 3-5

A n n u a l D r i v i n g Forces a n d I n v e s t m e n t T h e m e s for Wafer Fab E q u i p m e n t , 1995-2000

Logic Semiconductor

Unit Growth*

Investment in Logic

Capacity*

Memory Semiconductor

Unit Growth*

1995

Solid

Solid

SoUd

Investment in Memory

Capacity*

Front-End Equipment versus Facilities Loading of Capital

Strong

Facilities

Primary Technologies

Invested

0.35-0.6 micron

'Scale: Strong > Solid > Moderate > Weak > Dead

Source: Dataquest (December 1995)

1996

SoHd

SoHd

Moderate

Moderate

Equipment

0.35-0.5 micron

1997

Moderate

Solid

Weak

Weak

Balanced

0.35-0.5 nrucron

1998

Moderate

Moderate

Moderate

Weak

Facilities

0.3-0.5 micron

1999

Moderate

Moderate to Weak

Solid

Moderate

Equipment Equipment

0.25-0.4 micron

2000

Solid

Solid

Strong

Strong

0.25-0.4 micron

When Will Capacity Expand to Meet Demand? An Update of the Overor Underinvestment Model

In our forecasts last year, we introduced a model that provided a measure of the net cumulative over- or underinvestment in wafer fab equipment to support capacity needs. Since equipment purchases precede actual capacity on line by a number of months or quarters, this model could be viewed as a gross "leading indicator" of capacity shortages and excesses. The results of this model are closer to a one-and-one-half-year to three-year indicator of turning points in the equipment industry. The methodology of the net cumulative investment (NCI) model is linked to our longer-range forecast model.

Our methodology starts with the following key assumptions and baselines:

• Long-term growth rates for semiconductors and wafer fab equipment are correlated. In other words, semiconductor revenue and profits are needed before money can be spent on equipment, and vice versa.

• Also, net cumulative investment equals zero over time, which means that in a noncyclical environnient where annual growth rates are constant, investment and capacity are at equilibrium at all times. Of course, the semiconductor industry cycles through over- and underinvestment.

SEMIVi-WW-IViT-9502 ©1996 Dataquest

January 22,1996

Wafer Fab Equipment Forecast 27

• The output is a tangible number, and is in dollars of over- or underinvestment at year-end. However, the more useful output of the model divides this gross dollar number by the wafer fab equipment market size. The result is a figure for percentage of market tihat is repeatable in level from cycle to cycle.

• To take into consideration the long-term growth of the semiconductor and equipment industries, the model has a factor allowing the fundamentals of the industry to change over time.

A net positive or negative investment is calculated relative to the longterm growth baseline annually and then added to the prior year. The calculation results in a dollar value net cumulative over- or underinvestment and has correlated well with historical patterns.

Figures 3-1 and 3-2 show the most recent results of the model, little changed from our midyear forecast update. In absolute dollar terms, by the end of 1995, the industry will be overinvested by $5.8 billion, or

32.7 percent of the wafer fab equipment market, exceeding levels witnessed during the 1984 and 1989 peaks. These levels are being driven by two basic factors. First, PC unit demand is continuing to grow annually in the high teens. About one-third of the seniiconductor industry—and over one-half of the capital spending on new capacity—is geared to supporting this demand.

Second, the DRAM market has not yet converted to run the more siliconefficient 16Mb DRAM, placing this investment cycle about seven years behind the last cycle. DRAM bit demand generally runs at 50 percent annually, and DRAM manufacturing has depended on increasing unit densities (increasing the bits per square inch) to meet this demand.

Shrinks of existing generations alone bring only 15 percent to 25 percent armual bits per square inch efficiencies. Converting a fab running 4Mb

DRAMs to 16Mb DRAMs would increase bits per square inch by two to three times. Since low yields are holding back the economic conversion of the market, top-line bit demand is translating to square-inch demand

(silicon) and the equipment to run it. Since the equipment being installed is fully "convertible" to run 16Mb DRAMs, we can tiiink of these fabs building "pent-up supply" in bits. Once 16Mb DRAM jdelds (for the

1x16 configuration) exceed 60 percent to 65 percent, it will be more economical to n m these lines, and DRAM prices will erode. The cvirrent view is that this will not occxir vmtil well into 1996.

We have factored into the model an investment in a couple of 300mm fabs starting in 1997 through 1999, with the bulk in 1998. This is considered an equipment "bubble demand" because the equipment will be shipped into a nonproductive fab (so that no semiconductor revenue will be generated initially).

V\^th our forecast for momentum-style growth in 1996 and two pause years in 1997 and 1998, the model indicates a reacceleration of equipment spending starting in 1999.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

26

Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide

Figure 3-1

Net Cumulative Over- and Underinvestment of Semiconductor Wafer Fab Equipment

Millions of Dollars

10,000-

8,000

6,000 H

4,000

2,000

0

' ^ ' s g ' ^ ' ' ^ -

^ ^ .

r ^

M

n m

-2,000

-4,000

-6,000

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1 9 9 4 * 1 9 9 5 * 1 9 9 6 * 1 9 9 7 * 1 9 9 8 * 1 9 9 9 * 2 0 0 0 *

"Forecast

Source: Dataquest (December 1995)

Figure 3-2

Net Cumulative Over- and Underinvestment of Semiconductor Wafer Fab Equipment

Percentage of Wafer Fab Equipment Market

4011

0

30

20

M

10

i

^

p

m n. n

-10

-20

-30-"

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995* 1996* 1997* 1998* 1999* 2000*

"Forecast 9507314

Source: Dataquest (December 1995)

©1996 Dataquest

SEi\/IM-WW-l\/IT-9502 January 22,1996

Wafer Fab Equipment Forecast 29

When Will Capacity Expand to Meet Demand? A Fundamental Set of Analyses

The NCI model is only a tool to indicate possible future turning points and should not be relied on to forecast capacity supply versus demand absolutely. A more fundamental, basic approach is required for that—looking at square inches of silicon capacity. Calculating how many wafers can be processed and in what type of process is much more enlightening in measuring capacity than focusing on a particular device t5rpe. Capacity is very fluid: a stepper does not care whether the picture it takes is for a DRAM,

SRAM, or logic device. But there are limits in transferring capacity. For example, logic processes have specialized process techniques that are not found in DRAMs, and vice versa. SRAMs can use a DRAM equipment mix or a logic-oriented process scheme; the latter tends to have faster access times.

There are two major markets that can be isolated in order to understand basic capacity supply and demand: the DRAM market and the foundry market. The latter is particularly interesting for two reasons. First, foundry capacity has tended to be more heavily logic- and ASIC-process oriented, giving us a second perspective on capacity versus supply. Second, the major customer base for foundries is the fabless company, whose products tend to be placed within PC logic and graphics chipsets. Since PCs now account for about one-third of all semiconductor use and around twothirds of all DRAM consumption and are the main engine driving the current semiconductor boom, looking at PC-related capacity issues is important for understanding potential future equipment market turning points.

The details of these analyses are provided in other Dataquest dociiments, including Dataquest's Quarterly DRAM Supply/Demand Report and a recent comprehensive study on the foundry market. A summary of the basic results and impacts will be given here.

Supply/demand trends usually show a cycle between oversupply and imdersupply. Investment in capacity tends to be in reaction to tiiese situations, and there is inevitably overshoot in both directions. Analysis is based on square inches of silicon, not on revenue, bit demand, or xmit demand. If demand for silicon area exceeds supply, the market is technically in undersupply. We will refer to the maximtim undersupply point as the market "pinch point."

The following are the basic conclusions and impacts:

• The DRAM market's pinch point occurred in 1995, but DRAM expected to be in undersupply through 1996. The market is expected to reach oversupply by early 1997, and is expected to last 12 to 18 months.

• The foundry market is now in undersupply and is expected to get tighter in 1996. Next year will represent the capacity pinch-point in foundry, and the market is expected to remain in undercapacity through early 1998. The oversupply part of the cycle is expected to be short-lived and to peak in 1999.

• The year 1996 will bring continued tight supply and strong investment in capacity. DRAM prices will edge down early in the year , but only because the upward pressure is no longer present. The market will

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

30 Semiconductor Equipment, Manufacturing, and Materials Worldwide begin to transition to the 16Mb density later in the year, and prices will continue to decline gradually into 1997. Foundry prices are expected to edge upward during the year.

• In 1997, there will be crosscurrents: weaker DRAM capacity investment, but continued very strong logic capacity growth. As DRAM goes into oversupply, investment is likely to dry up quickly. But the continued tight supply of foundry and logic capacity is expected to draw continued investment in the United States and Asia in 1997. We expect DRAM investment to decline some 10 percent to 20 percent, while logic investment increases 20 to 30 percent.

• The toughest year will be 1998. Even though DRAM capacity is likely to reach peak oversupply early in the year, reinvestment will not likely occur imtil 1999. As the foundry market approaches oversupply, logic investment will cool.

• Another crosscurrent year will be 1999. DRAM investment is likely to pick up again, while fovmdry investment will lag. U.S. logic capacity will likely lead foimdry investment out of the "pause." By 2000, the next equipment boom will be imder way and is to last through 2001 or 2002.

Highlights of Key Equipment Segment Maricets and Forecasts

Steppers and Track

From 1989, the peak year of stepper shipments at more than 950 units, the market tumbled to fewer than 400 tools in 1992 before recovering. In the

DRAM-sensitive ramp now occurring, the industry has experienced its first 1,000-stepper year. In fact, we believe that in excess of 1,240 steppers shipped in 1995. Shifts in the product mix toward higher-priced i-line systems and wide field lenses have also driven up average selling prices

(ASPs). This, along with the strong yen, has yielded a revenue increase of

78 percent on a dollar basis. This faster-than-market growth is expected to continue into 1996 as semiconductor companies invest in the bottleneck equipment sets and backlogs are at record levels.

Stepper revenue is forecast to grow at a 20.4 percent CAGR, slightly below the market average for 1994 through 2000. Our forecast for stepper unit growth over the five-year forecast horizon remains modest but higher

5ian we have seen in the past, about a 7 percent CAGR between 1994 and 2000.

Wiih. the adoption of phase shift mask technology, off-axis illumination techniques as well as conventional i-line tools with variable numerical aperture (NA), i-line is clearly a viable technology down to the

0.30-micron regime and will continue to dominate the overall stepper technology mix through 2000. Excimer/deep-ultraviolet (deep-UV) steppers will begin to represent a more significant portion of the product mix from 1997 onward for use in below-0.30-micron devices and ICs with large chip areas, such as advanced microprocessors, that require large field size capability. Dataquest believes that field size pressures accompanied by shrinking geometries will drive the industry toward step-and-scan (or step-and-stitch) technologies for the majority of excimer/deep-UV shipments, beginning in 1997.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Wafer Fab Equipment Forecast 31

Track equipment is forecast to grow at a 23.1 percent CAGR between 1994 and 2000, slightly ahead of the industry growth of 21.4 percent. Although we believe that the rapid shift in the product mix toward higher-priced systems has been recently completed, we do expect another product shift to occur in the track market and we expect higher ASPs associated with the ramp of deep-UV steppers, which require more sophisticated environmental control systems.

Etch and Clean: Dry Etch and Chemical Mechanical Polishing

Dataquest began covering the chemical mechanical polishing (CMP) market in 1993, a year in which sales increased 121 percent to over $44 million.

Dataquest includes the post-CMP dean system, usually sold in conjunction with a CMP tool, as part of the cleaning segment, not in the CMP segment. It was expected that 1994 would be another high growth year, but the market was disappointing, growing at a slower-than-market average to approach $65 mUlion (or 49 percent growth). We believe that this resulted from the time that was required to evaluate the technology and gain production experience with it before widespread adoption occurred.

In 1995, that adoption has started, with CMP systems growing 135 percent to over $152 million. Even though the application appears to be limited to devices with at least three levels of metal, which tends to exclude the

DRAM market, acceptance of the technology into the foundry market has been the key. Until recently, most foundries we have talked with had not planned on putting CMP into their standard 0.5-micron process flow.

Based on the customer demand, however, and because more robust equipment has been introduced, we now believe that many foundries will offer

CMP at 0.5-micron.

These systems are used to remove material from the surface of the wafer, resulting in a flat surface over the entire wafer. This global planarization, primarily of dielectric layers, is required to achieve higji yields in devices where three or more levels of metal are used. Today's advanced logic and

ASIC devices are fueling this market growth. Dataquest believes that this technology and market will become a major part of semiconductor manufacturing in the long run.

If we were to create a forecast based purely on technology driving the market, we would not slow the CMP market forecast in 1998. However, we believe that in time we will see some holding back of capital investment, and history has shown us that even advanced, emerging technologies are rarely spared in a capital slowdown. Nonetheless, CMP is our fastestgrowing segment with a 48.1 percent CAGR for 1994 through 2000.

Dry etch systems continue to exhibit strong revenue growth, with a CAGR of 19.5 percent forecast for 1994 through 2000. Urut shipments are expected to grow as greater multilevel interconnect process capacity is brought online, increasing the need for dielectric and metal capacity. Relatively strong ASP growth will lend additional momentum to dry etch revenue growth as new high-density plasma systems for 0.35-micron applications enter the market and multichamber cluster tools continue to increase their presence. The success of CMP will hold etch below market growth, however, particularly in metal etch as stringer removal becomes a nonissue.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

32 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Clean process equipment is forecast to show revenue growth of 20.7 percent annually from 1994 through 2000, basically on par with overall equipment market growth. This market is more heavily dependent on the brickand-mortar part of the capital investment dollar, as wafer cleaning is a fundamental in all parts of the fab. Automated wet stations dominate this segment, now accounting for nearly 70 percent of the clean process systems. Emerging areas of cleaning tools include (in order of size) spray processors, post-CMP clean, and vapor phase cleaning. As manufacturers look for novel ways to increase yields, we would expect these and other new techniques to evolve and increase in importance.

Deposition: CVD, PVD, and Silicon Epitaxy

CVD equipment sales are predicted to grow at a 23.9 percent annualized rate from 1994 through 2000, above overall equipment growth. The steady growth in multilevel interconnect capacity will continue to generate demand for dielectric and metal CVD systems. ASP growth will also contribute to revenue growth, as more highly integrated systems with improved productivity and particle control appear in the marketplace.

Advanced dielectric deposition systems utilizing high-density plasma

(HDP) sources will be introduced for intermetal dielectric (IMD) applications for sub-0.5 micron processes. Most systems will be introduced on multichamber cluster platforms. Metal CVD will continue to exhibit strong growth, driven by blanket tungsten for contact and via plugs, CVD barrier metals such as CVD titarvium nitride, and dichlorosilane (DCS) ttmgsten silicide for shrink 16Mb and 64Mb DRAMs. For these reasons, the forecast for nontube CVD systems outperforms tube furnaces.

Sputter deposition systems are forecast to grow at an annualized rate of

19.5 percent for 1994 through 2000. As in the case of dry etch and CVD equipment, continued expansion of multilevel intercormect process capacity is the primary driver behind sputter system growth. Rapid growth in average system ASP has helped to drive total revenue growth in 1994, primarily from the rapid and expanding dominance of Applied Materials in the market. With Applied Materials now accounting for more than 50 percent of the market, the bulk of the ASP increases are behind us. Revolutionary changes in system architecture, pioneered by the Applied

Materials Endura system, will continue to yield improvements in film properties, equipment productivity, and defect density. This is a market segment that will be somewhat buffered from a slowdown in DRAM investment as the fundamental growth in the number of metal layers in

ASICs and logic devices drives a more stable outlook.

The shift from batch to single-wafer epitaxial systems has been the primary driver of epitaxial deposition systems, as 200mm epitaxial wafer capacity was needed. In 1994 we saw the Applied Materials system become an "Intel-accepted supplier," so silicon companies are investing to add this capacity. However, this capacity is more expensive than wafer suppliers would like, so we expect the concept of "minibatches" to emerge as a viable production strategy, as it has in CVD. Moore Technologies is known to be in the process of releasing such a new product. A strong automotive, power, and discrete device market has increased demand for the specialty batch units. An increased product mix of logic semiconductors, sustained demand for discretes, and 200mm wafer capacity addition will be the primary drivers for epitaxial deposition equipment growth beyond

1995.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Wafer Fab Equipment Forecast 33

Thermal Nondeposition Processes: Diffusion and RTP

Diffusion systems are expected to demonstrate a CAGR of 17.9 percent for

1994 through 2000. The displacement of horizontal tube systems by vertical tube reactors will continue. Newer vertical systems will be configurable as multitube clusters with integrated dry clean capability, to compete with single wafer cluster tools. Tube systems wUl also incorporate small batch capabilities to offer greater flexibility for custom and semicustom circuit meinufacturers.

RTP will grow at an annualized rate of 34.5 percent for 1994 through 2000.

This market grew much faster than we anticipated, nearly doubling in

1994 and sUghtly more than doubling in 1995. The growth in 1994 was primarily fueled by the growing acceptance of self-aligned silicide processes in logic process flows. The growth in 1995 comes from new offerings in the market from Applied Materials, CVC, and Mattson and from the expansion of the application into traditional tube diffusion steps. The real growth for this segment will come from transitioning of the thermal "nondepositing" processes away from diffusion tubes and into single-wafer

RTT systems for 300mm wafers. We have factored a large complement of systems into initial 300mm facilities starting in 1997, largely contributing to the higher-than-market growth. RTP systems are primarily used today for silicide anneals and are primarily driven by logic and ASIC capacity.

Dataquest believes that batch tube systems will continue to resist penetration by RTP in areas such as well drive, BPSG reflow, and thermal oxidation because of the demonstrated cost of ownership benefits in these areas, at least through 200mm wafers. For 300mm wafers, there will also be a strong desire on the part of the semiconductor manufacturer to continue to use batch tube systems because these systems offer much better cost efficiencies.

Ion Implantation

Overall ion implantation system revenue is forecast to grow at a CAGR of

20.2 percent for the years 1994 through 2000. This market segment will continue to be one of the most volatile because of the highly device-specific nature of the implant segments and because of the dependence on new fab capacity for unit growth. The fastest-growing segment is expected to be high-energy implantation, which is evoking intense interest because of its potential for process simplification and manufacturing cost reduction. The first year of true production ramp is expected to occur in 1998 as

0.4-micron technologies becomes mainstream, although early adopters such as Samsimg have placed high-voltage implant into 16Mb DRAMs.

New implant systems will continue to offer improvement in uniformity, particle control, charging, and wafer throughput. The number of implant steps requiring medium-current ion sources is expected to grow faster than high-dose implant steps, again driven by the higher worldwide semiconductor logic component, with the shallow junctions preferentially driving the trend toward medium-current implants.

However, our forecast does not reflect this trend, with medium-current implant sales lagging the market with only 15.1 percent CAGR. Why?

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

34 Semiconductor Equipment, Manufacturing, and IWaterials Worldwide

Recently Eaton, Applied Materials, and Genus introduced expanded capability to existing systems or new systems targeted to compete in more than one segment. For example. Applied Materials' 9500 systems now basically cover both high-current and medium-current capabilities, and Genus' new

1520 model expands the range effectively across high-voltage and medium-current. Traditional batch medium-current systems are effectively being squeezed out of the market. This is occurring because equipment utilization for implanters tends to be among the lowest in the fab.

Semiconductor manufacturers, in an effort to increase utilization and reduce cost, are tending to buy equipment with broader ranges.

There are mediiun-current applications that will still require dedicated medium-current systems, namely high-tilt (>42 degree) implants and Vj adjustment implants. However, ihese will be executed better using implanters witti single-wafer end stations rather than the traditional batch systems.

Dataquest is investigating tiie "redefinition" of ion implant segments in order to capture this market dynamic better.

Segments and Tools of Emerging Importance

"Weld enhancement is the talk of the moment. If a piece of equipment enhances yield directly in any way, particularly if it is an incremental concept and on the lower end of the price scale, it will gain immediate acceptance. Such appears to be the case in three specific areas: photostabilizers, a tool we refer to as a pull dryer, and emerging issues in process control/ overlay.

Photostabilizers are relatively small systems, selling for perhaps $200,000, that employ a UV curing technique to treat exposed photoresist on the wafer before placing the wafer into the etcher. Fusion Systems in the

United States has been one of the leaders in this area. The system improves the quality of the photoresist in preparation for etch and thus dramatically improves the etch system performance. Samsung is known to have installed many imits, and the product segment is starting to attract major interest.

A visit to Steag AG in Germany recently introduced us to the concept of what we currently refer to as a puU dryer. These systems are replacing the spin dryers usually employed within the automated wet stations today.

Wafers are placed in the system after a wet cleaning process to dry, but they do not spin. Instead, they are pulled from a bath of isopropyl alcohol

(IPA) and water in a controlled fashion into an IPA vapor atmosphere. As the wafer is pulled out, the liquid sheets off the surface in such a way as to leave no water spots, which often hold killer residue defects. Micron Technology is known to have purchased many of these systems, and we expect this technology to be introduced by several companies in the wafer cleaning area.

Process control systems are the core of the yield enhancement movement, and it is estimated that close to 80 percent of the failures in ICs (killer defects) may be attributed to particulate contamination. Therefore, preand postprocess particle and defect monitoring and characterization are

SEI\/IIVl-WW-l\/IT-9502 ©1996 Dataquest January 22,1996

Wafer Fab Equipment Forecast 35 key to increasing yield. There are several important developments to watch in the equipment segments of process control, which we believe will experience very strong growth in the years ahead. These segments include wafer inspection and review, CD metrology, and thin-film measurement.

Patterned wafer inspection stations have led the way to in-line use in fabs.

Demand for unpattemed wafer inspection is also gaining momentum in applications such as post-CMP inspection. Defect review technology has evolved from microscope-based systems to automated stations that characterize defects at the coordinates provided by wafer inspection systems.

Laser-based systems, such as Ultrapointe's recentiy introduced product, address the throughput issue along with increased resolution. With higher throughput and automated capabilities, defect review stations should follow the inspection segment to be used in-line.

The CD-SEM market has grown tremendously in the past several years, with the major thrust being the introduction of high throughput automated SEM systems introduced several years ago. Operator "interpretation" of SEM data and measurement has always been a problem in CD metrology. Today's SEM systems have made qualitative improvement to the electron emission source and are equipped with pattern recognition software and hardware, thus automating the interpretation function. In turn, the ASPs have increased two-fold to about $1.2 million. ASPs wiU continue to grow along with demand for these automated high-throughput systems.

Thin-fUm measurement is key to intra- and interlevel metal interconnect and storage capacitance applications. Thin-film measurement systems are used in-line to moiutor the in-etch, lithography (photoresist), deposition, and diffusion steps. Although this market has been driven more by logic applications than by memory production, in the past several years, DRAM manvif acturers have begun integrating thin-film measurement stations into their process lines.

Dataquest Perspective

Wafer fab equipment spending grew 66 percent in 1995 and is expected to grow 36 percent worldwide in 1996, driven by massive spending in the

Asia/Pacific region and a strong investment in the United States. DRAMsensitive Japan investment and an expanding European production base complete the strong picture for 1996. Segment growth in 1994 and 1995 is being led by DRAM or capital spending-sensitive equipment, with steppers, implant, wafer inspection, and factory automation exhibiting significantly stronger-than-market growth. New technology segments such as

CMP, high voltage implant, and RTP were the fastest-growing segments.

We expect no major segment declines in 1996, as capacity additions are broad-based and worldwide. Record backlog levels going into 1996 will be a buffer against a market decline in 1997, but we do expect softness from the DRAM companies to become evident by the end of 1996.

After four strong expansion years in 1993 through 1996, equipment purchases in 1997 should decline markedly, followed by a s l i ^ t decline in

1998. Investment in DRAM capacity will be curtailed as producers elect to

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

36 Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide convert their 4Mb DRAM capacity to 16Mb, which adds bit capacity through the instant increase in bits per square inch. Also, many Japanese

DRAM facilities now running 150mm wafers wUl convert to 200mm wafers, further delaying the need for new equipment. DRAM-sensitive equipment technologies or capital intensive segments such as steppers, implantation, diffusion, and polysilicon etch will be affected more than logic-sensitive technologies such as sputtering, epitaxial reactors, CMP,

RTF, nontube CVD, or metal etch. The next expansion should kick in by

1999, driven by 0.35- to 0.4-micron capacity expansion.

We have factored in an infrastructvire investment in equipment for late

1997 through 1999, which will affect the forecast size of the markets positively. This additional investment will be for initial equipment to fUl some

300mm fabs to run silicon by 1999. However, we believe that a significant

300mm equipment market will not occur until well after 2000.

Yield enhancement is the trend of the time. Any system technology that can be priced relatively low and that has a direct impact on jdeld will gain immediate acceptance in volume. Areas emerging as particularly important today are in cleaning technology, photostabilizers, and process control metrology.

SEiVii\/l-WW-l\/iT-9502 ©1996 Dataquest January 22,1996

Chapter 4

Silicon Wafer Forecast,

Dataquest's current forecast and the underl5dng assumptions behind our expectations for regional silicon wafer demand reflect significant silicon wafer growth in 1995 and 1996, with upward revisions to the 1997 through

2000 forecast, in line with increased semiconductor consumption forecasts worldwide. Our latest forecasts, along with highlights of some of the key factors affecting the regional markets, are presented here.

Silicon Forecast Tables

Tables in this chapter include Dataquest's most recent forecasts of regional unit silicon wafer consumption. Tables 4-1 through 4-5 detail unit consumption by region. Individual forecasts of major product segments such as prime, epitaxial, and test and monitor wafers are included.

Tables 4-6 through 4-10 present regional forecasts for wafer size distribution.

Table 4-1

Forecast of Captive and Merchant Silicon* and Merchant Epitaxial Wafers by Region

(Millions of Square Inches)

Worldwide Total

Silicon + Epitaxial

Merchant and Captive Silicon*

Epitaxial Silicon

N o r t h America Total

Silicon + Epitaxial

Merchant and Captive Silicon*

1993

2,449.7

2,157.1

292.6

720.0

565.5

1994

2,919.0

2,535.3

383.7

832.1

641.9

1995

3,588.6

3,091.6

497.0

1,065.7

iT--.,y _K_

824.0

1996

4,150.1

3,567.4

582.7

1,184.9

/ p " o

900.7

1997

4,734.1

4,101.9

632.2

1,281.9

/ '2<,,

973.1

1998

5,176.9

4,483.7

693.2

1,026.6

1999

5,545.3

4,782.7

762.6

1,366.8

^ • , ' : ' - ;

1,499.3

:^,l-ij/^

1,124.6

2000

6,166.7

5,291.0

875.7

1,668.3

1,239.7

Epitaxial Silicon

Japan Total Silicon +

Epitaxial

Merchant and Captive Silicon*

154.5

1,127.3

1,038.3

Epitaxial Silicon

Europe Total Silicon +

Epitaxial

89.0

291.2

Merchant and Captive Silicon*

Epitaxial Silicon

255.1

A s i a / Pacific-ROW Total

Silicon + Epitaxial

Merchant and Captive Silicon*

36.1

311.2

298.2

Epitaxial Silicon

13.0

•Includes prime, test, and monitor wafers

Source: Dataquest (January 1996)

190.2

119.5

351.2

296.7

54.5

456.0

436.5

19.5

241.7

1,160.2 1,349.8

155.9

408.7

336.4

72.3

608.5

581.4

27.1

284.2

1,547.6

178.7

474.2

387.7

86.5

764.7

731.4

33.3

308.8

1,279.7 1,505.7 1,726.3 1,917.6

1,725.9

191.7

543.2

448.5

94.7

991.4

954.4

37.0

340.2

2,052.0

1,843.4

208.6

598.9

494.9

104.0

1,159.2

1,118.8

40.4

374.7

2,145.8

1,917.3

228.5

650.9

536.9

114.0

1,249.3

1,203.9

45.4

428.6

2,333.7

2,073.6

260.1

721.0

590.0

131.0

1,443.7

1,387.7

56.0

CAGR (%)

1994-2000

13.3

13.0

14.7

12.3

11.6

14.5

10.5

10.2

13.8

12.7

12.1

15.7

21.2

21.3

19.2

1

SEMM-WW-MT-9502 ©1996 Dataquest 37

38

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 4-2

Forecast Growth Rates of Captive and Merchant Silicon* and Merchant Epitaxial Wafers by Region (Percentage of Millions of Square Inches)

Worldwide Total Silicon +

Epitaxial

Merchant and Captive

Silicon*

Epitaxial Silicon

North America Total Silicon

+ Epitaxial

Merchant and Captive

Silicon*

Epitaxial Silicon

Japan Total Silicon +

Epitaxial

Merchant and Captive

Silicon*

Epitaxial Silicon

Europe Total Silicon +

Epitaxial

Merchant and Captive

Silicon*

Epitaxial Silicon

Asia/Padfic-ROW Total

Silicon + Epitaxial

Merchant and Captive

Silicon*

Epitaxial Silicon

•Includes prime, test, and monitor wafers

Source: Dataquest (January 1996)

1993

16.8

16.7

17.7

10.6

8.2

20.6

15.9

17.2

2.3

23.9

20.8

51.0

30.8

30.5

36.8

1994

19.2

17.5

31.1

15.6

13.5

23.1

13.5

11.7

34.3

20.6

16.3

51.0

46.5

46.4

50.0

1995

22.9

21.9

29.5

28.1

28.4

27.1

17.7

16.3

30.5

16.4

13.4

32.7

33.4

33.2

39.0

1996

15.6

15.4

17.2

11.2

9.3

17.6

14.7

14.7

14.6

16.0

15.2

19.6

25.7

25.8

22.9

1997

14.1

15.0

8.5

8.2

8.0

8.7

11.1

11.5

7.3

14.6

15.7

9.5

29.6

30.5

11.1

1998

9.4

9.3

9.6

6.6

5.5

10.2

7.0

6.8

8.8

10.3

10.3

9.8

16.9

17.2

9.2

1999

7.1

6.7

10.0

9.7

9.5

10.1

4.6

4.0

9.5

8.7

8.5

9.6

7.8

7.6

12.4

2000

11.2

10.6

14.8

11.3

10.2

14.4

8.8

8.2

13.8

10.8

9.9

14.9

15.6

15.3

23.3

SEMM-WW-MT-9502

©1996 Dataquest January 22,1996

Silicon Wafer Forecast 39

Table 4-3

Forecast of Captive and Merchant Silicon* Wafers by Region (Millions of

Square Inches)

Worldwide Total

Silicon*

Merchant Silicon

Captive Silicon

North America Total

Silicon*

Merchant Silicon

Captive Silicon

Japan Total Silicon*

Merchant Silicon

Captive Silicon

Ettrope Total Silicon*

Merchant Silicon

Captive Silicon

Asia/Pacific-ROW

Total Silicon*

Merchant Silicon

Captive Silicon

1993

2,157.1

2,032.1

• 125.0

2,398.3

137.0

565.5 641.9

487.5

78.0

1,038.3

1,003.3

35.0

255.1

250.1

5.0

298.2

291.2

7.0

1994

2,535.3

429.5

7.0

551.9

90.0

1,160.2

1,125.2

35.0

296.7

291.7

5.0

436.5

'Includes prime, test, and monitor wafers

Source: Dataquest (January 1996)

1995

3,091.6

3,029.6

62.0

824.0

809.0

15.0

1,349.8

1,314.8

35.0

336.4

331.4

5.0

581.4

574.4

7.0

1996

3,567.4

1997

4,101.9

3,505.4

62.0

900.7

4,039.9

62.0

973.1

885.7

15.0

1,547.6

1,512.6

35.0

387.7

382.7

5.0

731.4

724.4

7.0

958.1

15.0

1,725.9

1,690.9

35.0

448.5

443.5

5.0

954.4

947.4

7.0

1998

4,483.7

1999

4,782.7

2000

5,291,0

4,421.7

62.0

1,026.6

4,720.7

62.0

1,124.6

5,229.0

62.0

1,239.7

1,011.6

15.0

1,843.4

1,808.4

35.0

494.9

489.9

5.0

1,118.8

1,109.6

15.0

1,917.3

1,882.3

35.0

536.9

531.9

5.0

1,203.9

1,224.7

15.0

2,073.6

2,038.6

35.0

590.0

585.0

5.0

1,387.7

1,111.8

7.0

1,196.9

7.0

1,380.7

7.0

CAGR (%)

1994-2000

13.0

13.9

-12.4

11.6

21.5

0

14.2

-25.8

10.2

10.4

0

12.1

12.3

0

21.3

Table 4-4

Forecast Growth Rates of Captive and Merchant Silicon* by Region (Percentage of

Millions of Square Inches)

Worldwide Total Silicon*

Merchant Silicon

Captive Silicon

North America Total Silicon*

Merchant Silicon

Captive Silicon

Japan Total Silicon*

Merchant Silicon

Captive Silicon

Europe Total Silicon*

Merchant Silicon

Captive Silicon

Asia/Padfic-ROW Total Silicon*

Merchant Silicon

Captive Silicon

'Includes prime, test, and monitor wafers

Source: Dataquest (January 1996)

17.2

18.2

-5.4

20.8

21.3

0

30.5

30.9

16.7

1993

16.7

17.6

4.2

8.2

8.1

8.3

16.3

16.9

0

13.4

13.6

0

33.2

33.7

0

1995

21.9

26.3

-54.7

28.4

46.6

-83.3

1994

17.5

18.0

9.6

13.5

13.2

15.4

11.7

12.1

0

16.3

16.6

0

46.4

47.5

0

1996

15.4

15.7

0

9.3

9.5

0

14.7

15.0

0

15.2

15.5

0

25.8

26.1

0

0

6.8

6.9

0

10.3

10.5

0

17.2

17.4

0

1998

9.3

9.5

0

5.5

5.6

11.5

11.8

0

15.7

15.9

0

30.5

30.8

0

1997

15.0

15.2

0

8.0

8.2

0 0

4.0

4.1

0

8.5

8.6

0

7.6

7.7

0

1999

6.7

6.8

0

9.5

9.7

2000

10.6

10.8

0

10.2

10.4

0

8.2

8.3

0

9.9

10.0

0

15.3

15.4

0

SEIVII\^-WW-MT-9502 ©1996 Dataquest. January 22,1996

40

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 4-5

Forecast of Merchant Prime and Test/Monitor Wafers by Region

(Millions of Square Inches)

Worldwide Merchant

Silicon

Growth Rate (%)

Prime

Test and Monitor

North America

Merchant Silicon

Growth Rate (%)

Prime

Test and Monitor

Japan Merchant

Silicon

1993

2,032.1

17.6

1,616.8

415.3

487.5

8.1

385.1

102.4

1,003.3

Growth Rate (%)

Prime

Test and Monitor

Europe Merchant

Silicon

Growth Rate (%)

Prime

18.2

800.1

203.2

250.1

Test and Monitor

Asia/Padfic-ROW

Merchant Silicon

21.3

200.1

50.0

291.2

Growth Rate (%)

Prime

30.9

231.5

Test and Monitor 59.7

Source: Dataquest (January 1996)

1994

2,398.3

1995

3,029.6

18.0

1,892.6

505.7

551.9

26.3

2,280.1

749.5

809.0

15.7

2,568.7

936.7

885.7

15.2

2,943.4

1,096.5

958.1

9.5

3,293.5

1,128.2

1,011.6

6.8

3,589.8

1,130.9

1,109.6

10.8

4,033.9

1,195.1

1,224.7

13.2

430.5

121.4

1,125.2

12.1

894.5

230.7

291.7

16.6

230.4

61.3

429.5

47.5

337.2

92.3

46.6

594.0

215.0

1,314.8

1996

3,505.4

1997

4,039.9

1998

4,421.7

9.5

614.7

271.0

1,512.6

16.9

1,015.8

299.0

331.4

15.0

1,156.2

356.4

382.7

13.6

242.5

88.9

574.4

33.7

427.8

146.6

15.5

266.5

116.2

724.4

26.1

531.3

193.1

8.2

657.9

300.2

1,690.9

11.8

1,290.4

400.5

443.5

15.9

300.0

143.5

947.4

30.8

695.1

252.3

5.6

709.2

302.4

1,808.4

6.9

1,404.4

404.0

489.9

17.4

835.4

276.4

1999

4,720.7

2000

5,229.0

9.7

787.7

321.9

1,882.3

10.4

879.3

345.4

2,038.6

4.1

1,489.4

392.9

531.9

10.5

344.5

145.4

1,111.8

8.6

386.2

145.7

1,196.9

7.7

926.5

270.4

8.3

1,626.8

411.8

585.0

10.0

435.9

149.1

1,380.7

15.4

1,091.9

288.8

CAGR (%)

1994-2000

13.9

13.4

15.4

14.2

12.6

19.0

10.4

10.5

10.1

12.3

11.2

16.0

21.5

21.6

20.9

SEMM-WW-MT-9502

©1996 Dataquest

January 22,1996

Silicon Wafer Forecast

41

Table 4-6

Worldwide Wafer Size Distribution Forecast, 1993-2000

Diameter

Percentage Square

Inches by Diameter

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total

Total (MSI)

Growth (%)

Area

(Sq. In.)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 Inches

3 Inches

3.14

7.07

100mm

125mm

12.17

19.02

150mm

200mm

300mm

27.38

48.67

109.56

Total Wafers

(Millions)

Average Wafer

Diameter (Inches)

Source: Dataquest (January 1996)

1993

0.1

1.6

16.1

28.8

49.6

3.6

0

100.0

2,450

16.8

1.0

5.6

32.5

37.1

44.4

1.8

0

122.5

5.05

1994

0.1

1.4

14.8

24.7

48.0

11.0

0

100.0

2,919

19.2

1.1

5.7

35.5

37.9

51.2

6.6

0

138.0

5.19

1995

0.1

1.1

12.6

21.1

44.9

20.2

0

100.0

3,589

22.9

1.2

5.6

37.0

39.9

58.9

14.9

0

157.5

5.39

1996

0.7

5.5

35.2

39.0

65.8

23.4

0

169.6

5.58

0.1

0.9

10.3

17.9

43.4

27.4

0 ,

100.0

4,150

15.6

1997

0

0.8

8.6

15.1

43.4

32.1

0

100.0

4,734

14.1

5.74

0.6

5.1

33.4

37.6

75.1

31.2

0

183.1

1998

0

0.6

7.5

13.3

42.1

36.4

0

100.0

5,177

9.4

5.86

0.7

4.6

31.9

36.2

79.5

38.8

0.01

191.7

1999

0.4

4.0

30.9

34.7

84.0

44.7

0.06

198.7

5.96

2000

0

0.5

6.8

11.9

41.5

39.2

0.1

100.0

5,545

7.1

0

0.4

6.0

10.4

40.7

42.4

0.1

100.0

6,167

11.2

6.07

0

3.5

30.4

33.6

91.6

53.7

0.08

213.0

SEMM-WW-MT-9502

©1996 Dataquest January 22,1996

42

Semicondtjctor Equipment, Manufacturing, and Materials Worldwide

Table 4-7

North American Wafer Size Distribution Forecast, 1993-2000

Diameter

Percentage Square

Inches by Diameter

2 inches

3 inches

100mm

125mm

150mm

Area

(Sq. In.)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

200mm

300mm

Total

Total (MSI)

Growth (%)

Unit Distribution b y Wafer Starts

(Millions of Wafers)

2 inches

3.14

3 inches

7.07

100mm

125mm

12.17

19.02

150mm

200mm

27.38

48.67

109.56

300mm

Total Wafers

(Millions)

Average Wafer

Diameter (Inches)

Source: Dataquest (January 1996)

1993

0.1

1.2

22.4

26.5

44.0

5.8

0

100.0

720

10.6

0.2

1.2

13.3

10.0

11.6

0.9

0

37.2

4.97

1994

0.1

1.1

20.1

21.7

42.2

14.8

0

100.0

832

15.6

0.3

1.3

13.7

9.5

12.8

2.5

0

40.2

5.14

1995

0.1

1.0

17.1

19.0

39.6

23.2

0

100.0

1,066

28.1

0.3

1.5

15.0

10.6

15.4

5.1

0

48.0

5.32

1996

0

1.5

14.0

10.2

16.0

7.7

0

49.3

5.53

0

0.9

14.4

16.3

36.9

31.5

0

100.0

1,185

11.2

1997

0

0.7

12.7

14.6

37.0

35.0

0

100.0

1,282

8.2

5.66

0

1.3

13.4

9.8

17.3

9.2

0

51.0

1998

0

0.6

11.4

13.0

36.7

38.2

0.1

100.0

1,367

6.6

5.76

0

1.2

12.8

9.3

18.3

10.7

0.01

52.4

1999

0

0.5

10.4

11.3

35.4

42.2

0.2

100.0

1,499

9.7

5.88

0

1.1

12.8

8.9

19.4

13.0

0.03

55.2

2000

0

0.3

9.2

9.6

35.0

45.7

0.2

100.0

1,668

11.3

6.01

0

0.7

12.6

8.4

21.3

15.7

0.03

58.8

SEMM-WW-MT-9502

©1996 Dataquest

January 22,1996

Silicon Wafer Forecast 43

Table 4-8

Japan Wafer Size Distribution Forecast, 1993-2000

Diameter

Percentage Square

Inches by Diameter

2 inches

3 inches

100mm

125mm

150mm

200mm

300mm

Area

(Sq. In.)

Total

Total (MSI)

Growth (%)

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 inches

3.14

3 inches

7.07

100mm

125mm

12.17

19.02

150mm

200mm

300mm

27.38

48.67

109.56

Total Wafers

(Millioris)

Average Wafer

Diameter (Inches)

Source: Dataquest (January 1996)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

1993

0

1.5

11.5

32.2

53.3

1.5

0

100.0

1,127

15.9

0

2.4

10.7

19.1

21.9

0.3

0

54.4

5.14

1994

0

1.4

10.8

29.9

51.6

6.3

0

100.0

1,280

13.5

0

2.5

11.4

20.1

24.1

1.7

0

59.8

5.22

1995

5.37

0

1.1

9.3

26.2

50.9

12.5

0

100.0

1,506

17.7

0

2.3

11.5

20.7

28.0

3.9

0

66.4

1996

0

2.2

11.2

20.8

32.3

6.0

0

72.6

1997

0

0.9

7.9

22.9

51.3

17.0

0

100.0

1,726

14.7

0

0.8

6.9

20.0

52.3

20.0

0

100.0

1,918

11.1

0

0.7

6.2

18.3

52.3

22.5

0

100.0

2,052

7.0

5.50

0

2.2

10.9

20.2

36.6

7.9

0

77.7

5.60

1998

5.68

0

2.0

10.5

19.7

39.2

9.5

0

80.9

1999

0

1.8

9.9

19.1

41.5

10.5

0.02

82.8

5.74

2000

0

0.6

5.6

16.9

52.9

23.9

0.1

100.0

2,146

4.6

0

0.5

5.0

15.2

52.5

26.7

0.1

100.0

2,334

8.8

5.83

0

1.7

9.6

18.6

44.7

12.8

0.02

87.5

SEMM-WW-MT-9502

©1996 Dataquest January 22,1996

44

Semiconductor Equipment, !\/lanufacturing, and IVIateriais Worldwide

Table 4-9

European Wafer Size Distribution Forecast, 1993-2000

Diameter

Percentage Square

Inches by Diameter

2 inches

3 inches

100mm

Area

(Sq. In.)

3.14

7.07

12.17

125mm

150mm

200mm

300mm

ISOnam

200mm

300mm

Total Wafers

(Millions)

Average Wafer

Diameter (Inches)

19.02

27.38

48.67

109.56

Total

Total (MSI)

Growth (%)

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 inches

3 inches

3.14

7.07

100mm

125mm

12.17

19.02

27.38

48.67

109.56

1993

0.2

1.7

22.0

30.1

42.3

3.7

0

100.0

291

23.9

0.2

0.7

5.3

4.6

4.5

0.2

0

15.5

4.89

1994

0.1

0.7

20.5

21.7

43.0

14.0

0

100.0

351

20.6

0.1

0.3

5.9

4.0

5.5

1.0

0

16.9

5.14

1995

0.1

0.5

18.0

19.3

39.1

23.0

0

100.0

409

16.4

5.8

1.9

0

18.4

0.1

0.3

6.0

4.1

5.32

1996

0

0.3

5.5

3.6

6.5

3.2

0

19.2

0

0.4

14.2

14.6

37.8

33.0

0

100.0

474

16.0

5.61

1997

0

0.2

11.0

11.5

35.3

42.0

0

100.0

543

14.6

5.87

0

0.2

4.9

3.3

7.0

4.7

0

20.0

1998

0

0.1

4.5

3.1

7.7

5.6

0

21.0

1999

0

0.1

9.2

9.8

35.2

45.7

0

100.0

599

10.3

0

0.1

7.9

8.3

35.1

48.6

0

100.0

651

8.7

0

0.1

7.1

7.3

35.0

50.5

0

100.0

721

10.8

6.02

8.3

6.5

0

22.0

0

0.1

4.2

2.8

6.14

2000

6.21

9.2

7.5

0

23.8

0

0.1

4.2

2.8

SEMI\/i-WW-MT-9502

©1996 Dataquest January 22,1996

Silicon Wafer Forecast

45

Table 4-10

Asia/Pacific-ROW Wafer Size Distribution Forecast, 1993-2000

Diameter

Percentage Square

Inches by Diameter

2 inches

3 inches

100mm

Area

(Sq. In.)

125mm

150mm

200mm

300mm

Total

Total (MSI)

Growth (%)

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 inches

3 inches

100mm

125mm

150mm

200mm

3.14

7.07

12.17

19.02

27.38

48.67

109.56

300mm

Total Wafers

(Millions)

Average Wafer

Diameter (Inches)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

Source: Dataquest (January 1996)

1993

0.6

2.9

13.0

20.8

56.3

6.4

0

100.0

311

30.8

6.4

0.4

0

15.4

0.6

1.3

3.3

3.4

5.07

1994

0.5

2.4

12.0

17.8

52.3

15.0

0

100.0

456

46.5

8.7

1.4

0

21.2

0.7

1.5

4.5

4.3

5.24

1995

9.6

4.0

0

24.7

0.8

1.5

4.5

4.3

0.4

1.7

9.0

13.5

43.4

32.0

0

100.0

609

33.4

5.60

1996

0.3

1.4

7.1

11.0

39.2

41.0

0

100.0

765

25.7

0.7

1.5

4.5

4.4

10.9

6.4

0

28.5

5.84

1997

0.2

1.1

5.2

8.3

39.1

46.1

0

100.0

991

29.6

14.2

9.4

0

34.3

0.6

1.5

4.2

4.3

6.07

1998

0.7

1.3

4.1

4.0

14.3

12.9

0

37.4

1999

0.2

0.8

4.3

6.6

33.8

54.3

0

100.0

1,159

16.9

0.1

0.6

3.9

5.9

32.4

57.0

0.1

100.0

1,249

7.8

6.28

0.4

1.1

4.0

3.9

14.8

14.6

0.01

38.8

6.41

2000

0.0

0.5

3.4

5.0

31.0

59.9

0.2

100.0

1,444

15.6

0

1.0

4.0

3.8

16.3

17.8

0.03

43.0

6.54

The 200mm Wafer Ramps Up

Dataquest has been studying the subject of 200mm wafers and their ramp rate closely over the last year, particularly in light of the announcements of massive fab capacity planned over the next few years. There are 76 new

200mm fabs expected to come on line from 1995 through 1997, and there are probably several more armouncements to come. The largest proportion comes on line in 1996 alone. Wafer suppliers have answered the need with several new wafer plants and billions in committed investment. We continue to believe that the ramp of 200mm wafers will be supply-constrained through the decade—that is to say, the industry cannot convert from

150mm at will. The biggest factor, however, is the consumption of test wafers—now over the ratio of one test wafer for every product wafer. This compares with the industry average of about one test wafer per seven product wafers for sizes below 200mm.

SEIVlM-WW-IVIT-9502

©1996 Dataquest January 22,1996

46 Semiconductor Equipment, Manufacturing, and Materials Worldwide

In this forecast update, we have revised the forecast—again upward—to reflect the expected fab activity in each region and the supply increases in place. Our current forecast for the year 2000 represents a level about 15 percent above what the wafer manufacturing industry has committed to supply to date. We stop short of forecasting a shortage of 200mm wafers to the point of restricting ramp-up plans, but buyers of wafers will experience firm-to-rising prices and wiU be placed on allocation from time to time over the next several years. One issue that semiconductor manufacturers wiU face is that they may be forced to reduce their test wafer consumption. According to straight demand requirements, the industry will come down only to tihe level of one test wafer for every 1.5 product wafers, but our forecast assumes that this demand for test wafers wiU be smaller, more on the order of one test wafer for every 3.5 product wafers. In a recent Focus Report {Is There a Silicon Shortage Looming? November 13,

1995, SEMM-WW-FR-9502) we have analyzed demand and supply of

200mm wafers in detail. The 200mm wafer may be in the best supply situation of aU areas.

What about 300mm Wafers?

Once the wafer size had been settled and the time horizon for the first

300m plant(s) had been proposed, we initiated a forecast for consumption of 300mm wafers. The level is still commercially zero before the year 2000; however, the recent goal of bringing a fab on line by 1998 or 1999 means that 300mm wafers will be made and significant activity will be occurring in R&D. Although no company has as yet made a firm commitment, we would expect something to be announced soon. This has been quite a story to watch this year. At the beginning of the year, there was great enthusiasm, particularly among Samsung, Motorola, Intel, and a Japanese company. As the year went on, Intel dropped off and both Samsung and

Motorola became more cautious. At this point, it is unclear who will step up to the bar, but we would also not be surprised if this were given a oneor two-year rest. SEMATECH is trying to be the champion, but as we have always said: Without funding, it will not happen.

We are assuming that at least two pilot fabs will seek to come on line by

1999.

Further details and issues regarding the move toward 300mm wafers are included in a recent Market Analysis document ("The Move toward

300mm Wafers," May 22,1995, SEMM-WW-MA-9503). Please refer to that document for a more detailed discussion.

We believe the first commercially productive plant will be started in the

2001-to-2003 period (after the feasibility noted above), with serious volimie ramp-up in the 2003-to-2005 period, which would be consistent with shrink 0.25-micron technology being produced primarily on 300mm wafers. This means that 200mm wafers represent at least a two-technology-generation wafer size, and fabs being built today may have longer lives than history would indicate.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Silicon Wafer Forecast 47

Highlights of the North American Silicon Wafer Market and Forecast

Silicon consumption in North America is forecast to grow 28 percent in

1995 to 1,066 million square inches (MSI), followed by mild, 11 percent growth in 1996 to 1,185 MSI. Microprocessor and other logic chip demand has been and will continue to be the key driver behind increased silicon demand in North America, and epitaxial wafer demand is expected to grow faster than the overall market throughout this decade.

Merchant epitaxial wafer consumption will increase 27 percent to 242 MSI, driven in large part by microprocessor manufacturers such as Intel and

AMD, which build their microprocessors on epitaxial wafers. The automotive and discrete segments of the chip market have also been much stronger than expected. By 1998, epitaxial silicon wiU accoimt for 25 percent of tixe square inches consumed in North America—the highest concentration in any region.

Dataquest's longer-term forecast for North American silicon consumption has increased, in part because of the increase in the overall semiconductor market revision since the last update, but also because we believe that the

United States will attract a larger share of foreign multinational fabs, such as those recentiy announced by Hyvmdai, TSMC, Toshiba (joint venture fab), and Samsung. We are projecting that total silicon MSI will grow at a

12.3 percent CAGR for 1994 through 2000.

Highlights of the Japanese Silicon Wafer Market and Forecast

Our Japanese silicon consumption forecast has sUghtiy increased from our last update, with the silicon market expected to have grown 18 percent to

1,506 MSI in 1995, with continued strong growth in 1996 and with softening in 1997 through 1999. The persistent high price of 4Mb DRAMs through 1995, wifii low yields restricting economical production of the more silicon-efficient 16Mb density, is the reason for tiie continued nearterm optimism. As conversion to tiie 16Mb density occvirs, silicon squareinch demand will ease.

Unlike North America, with its sizable CMOS epitaxial wafer market,

Japan's merchant epitaxial wafer market is more focused on discrete and bipolar applications. Therefore, a recovery in the economy will have more of an effect on the growth of Japan's epitaxial wafer market, and early indications are that it has kicked into gear. Epitaxial demand is expected to have grown 30 percent in 1995 after a 34 percent increase in 1994. Another good growth year is expected in 1996.

Dataquest's longer-term growth scenario for silicon wafer demand in

Japam remains moderately conservative as the country continues to work through difficulties with the economy and semiconductor production infrastructure (too heavily dependent on DRAM). Recent investment patterns, however, indicate that Japanese semiconductor manxifacturers are willing to come to the table and invest, preserving their stake in the memory business against Korean companies. The desired shift of the Japanese product mix to higher value-added semiconductors is apparentiy on the back burner until the current memory cycle subsides, but wiU come to the forefront soon. We are estimating that silicon demand will grow at a

SEI\/lM-WW-l\/IT-9502 ©1996 Dataquest January 22,1996

48 Semiconductor Equipment, N/lanufacturing, and Materials Worldwide

10.5 percent CAGR for the years 1994 through 2000, the slowest growth of all regions.

Highlights of the European Silicon Wafer Market and Forecast

Demand for silicon wafers in Europe, as well as wafer fabrication equipment, remains heavily dependent on the fab activities of foreign semiconductor firms. With increased presence from European companies, the outlook for silicon consumption has brightened.

European silicon demand has basically remained imchanged from our previous forecast at a level of 409 MSI in 1995, up just over 16 percent from

1994 levels. Siemens Dresden and other DRAM production and U.S. multinationals Intel, Motorola, and Texas Instruments will continue to ramp to answer the PC production boom in Europe, helping silicon consumption grow another 16 percent in 1996. These driving forces have caused us to call Europe the second-fastest-growing region for silicon consumption, growing at a rate of 12.7 percent CAGR through the year 2000.

Epitaxial wafer demand in the region wiU increase from 36 MSI in 1993 to

72 MSI in 1995, nearly double consumption, a direct result of Intel in Ireland. Beyond 1995, epitaxial wafer demand will come from European producers, primarily from the power/discrete market, leading to a CAGR more than three percentage points above polished bare wafer consumption through the decade.

Highlights of the Asia/Pacific-ROW Silicon Wafer Market and Forecast

Silicon consumption is growing at a 33 percent pace in 1995, the largest growth rate of any region in the world. Production is expanding at a fierce pace and not expected to ease through 1997, as many large 200mm fab projects are in various stages of construction and start-up. Because the

Asia/Pacific region collects nearly the largest proportion of 1995 capital, the trend for high silicon consumption growth will continue unabated.

The phenomenal growth in Asia/Pacific silicon consumption in 1995 is tied directly to the manufacturing activities of Korean DRAM producers.

Growth will be seen here, but increased production in Taiwan, Singapore, and the most recently announced Thailand foundries will also contribute.

Foundry-related capital spending in these Asia/Pacific countries has exploded, increasing from about $900 million in 1994 to over $1.8 billion in

1995. As these fabs come on line in 1995 and beyond, they will provide some regional consumption stability as memory-related silicon consumption cools in 1997 and 1998, Taiwan, with its many new DRAM producers coming on line in 1996, will cause silicon consumption to grow by nearly

26 percent again in 1996. Asia/Pacific-ROW remains the fastest-growing silicon consumer, with a forecast five-year CAGR of 21.2 percent.

Silicon Wafer Revenue Forecast

Dataquest has been tracking silicon wafer revenue and market share since

1985, but has always provided forecast information in terms of square inches of area and unit wafer size distributions. With the announcement of the initial public offering of MEMC Electroruc Materials in July 1995, we believe it is now important to consider the revenue-generating capability of the industry.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Silicon Wafer Forecast 49

We believe a revenue forecast will benefit our clients, gaining increased visibility for the capital markets. Table 4-11 contains the revenue forecast for silicon wafers worldwide. Our analysis concludes that the revenue forecast would resemble the semiconductor industry more closely than the capital spending markets. The concept of semiconductor revenue per square inch is more closely tied to silicon consumption than raw wafer capacity of the industry. IThe six-year CAGR of 18.9 percent is about one percentage point below the semiconductor forecast of 20.1 percent, consistent with the model that semiconductor manufacturers will attempt to control the costs associated with manufacturing, which includes using silicon more efficiently in the future. Yet, based on the fact that silicon manufacturers are attaining price increases in a sellers' market, this forecast may prove conservative in the near term.

Table 4-11

Worldwide Merchant Silicon Wafer Revenue Forecast^ 1992-2000 (Includes Polished,

Virgin Test, and Epitaxial Silicon; Millions of U.S. Dollars)

Worldwide

Growth (%)

1992

2,991

2.7

Source: Dataquest (January 1996)

1993

3,554

18.8

1994

4,592

29.2

1995

6,012

30.9

1996

7,470

24.3

1997

8,995

20.4

1998

10,199

13.4

1999

11,256

10.4

2000

12,951

15.1

CAGR (%)

1994-2000

18.9

Dataquest Perspective

The silicon market, driven by a stronger long-term picture for semiconductors in general, will grow faster over the next six years than it has in the recent past. As the industry moves to a 200mm baseline, the outlook for silicon wafer manufacturers becomes brighter Silicon manufacturers have answered the call for 200mm capacity, and the semiconductor market has again responded with a cry for more. We believe silicon manufacturers' ramp plans in 200mm have been strategically and smartly measured and that the overcapacity situations of 1985 are being remembered, and we are not expecting that scenario to be repeated.

Although there will be activity with 300mm wafers, this is expected to be focused on R&D and confined to low volumes imtil after the turn of the decade.

SEIVII\/l-WW-l\/IT-9502 ©1996 Dataquest January 22,1996

Chapter 5

Semiconductor Consumption Forecast

This chapter presents data on the worldwide semiconductor market by region. The regional semiconductor market, or regional semiconductor consumption, deals with where chips are consumed; this contrasts with regional semiconductor production, which deals with where chips are manufactured. The data presented here is for the merchant market and does not include the value of chips made by captive semiconductor manufacturers for internal use.

This is an excerpt from the Semiconductor Five-Year Forecast, published by Dataquest in October 1995 (SCND-WW-MT-9502). Further details about this forecast can be found in that publication.

Yearly exchange rate variations can have a significant effect on the 1988through-1995 data in the following tables. For more information about the exchange rates used and their effects, see Appendix B.

Semiconductor Consumption

Table 5-1 shows revenue and growth from semiconductor shipments for

1988 through 1994, by region. Table 5-2 shows revenue and growth from semiconductor shipments for 1994 through 2000, by region.

Table 5-1

Worldwide Semiconductor Consumption by Region—Historical (Includes Merchant

Semiconductor Companies Only; Millions of U.S. Dollars)

North America

Growth (%)

Japan

Growth (%)

Etirope

Growth (%)

Asia/Padfic-ROW

Growth (%)

1988

15,844

23.2

20,772

39.2

8,491

30.7

5,752

45.0

Worldwide

Growth (%)

Source: Dataquest (January 1996)

50,859

33.0

1989

17,070

7.7

21,491

3.5

9,498

11.9

6,280

9.2

54339

6.8

1990

16,540

-3.1

20,257

-5.7

10,415

9.7

7,333

16.8

54,545

0.4

1991

16,990

2.7

22,496

11.1

11,014

5.8

9,194

25.4

59,694

9.4

1992

20,430

20.2

20,579

-8.5

12,218

10.9

12,034

30.9

65,261

9.3

1993

27,926

36.7

24,645

19.8

15,461

26.5

17,486

45.3

85,518

31.0

1994

35,939

28.7

31,010

25.8

20,819

34.7

22,812

30.5

110,580

29.3

CAGR (%)

1989-1994

16.1

7.6

17.0

29.4

15.3

SEMM-WW-MT-9502

©1996 Dataquest 51

52

Semiconductor Equipment. Manufacturing, and Materials Worldwide

Table 5-2

Worldwide Semiconductor

Consumption by Region—Forecast (Includes Merchant

Semiconductor Companies

Only; Millions of U.S. Dollars)

1994

North America

Growth (%)

Japan

35,939

28.7

Growth (%)

Europe

Growth (%)

31,010

Asia/Padfic-ROW

Growth (%)

Worldwide

Growth (%)

Source: Dataquest (January 1996)

25.8

20,819

34.7

22,812

30.5

110,580

29.3

1995

47,805

33.0

41,843

34.9

28,995

39.3

31,145

36.5

149,788

35.5

1996

60,307

26.2

47,638

13.8

35,517

22.5

39,446

26.7

182,908

22.1

1997

71,508

18.6

52,760

10.8

39,674

11.7

46,419

17.7

210,361

15.0

1998

80,947

13.2

58,446

10.8

44,895

13.2

54,311

17.0

238,599

13.4

1999

92,908

14.8

65,241

11.6

51,859

15.5

65,736

21.0

275,744

15.6

2000

112,325

20.9

75,680

16.0

62,231

20.0

81,250

23.6

331,486

20.2

CAGR (%)

1994-2000

20.9

16.0

20.0

23.6

20.1

SEMM-WW-MT-9502

©1996 Dataquest January 22,1996

Chapter 6

Semiconductor Production Forecast

This chapter presents data on worldwide semiconductor production by region. Semiconductor production is defined by the place where the wafers are fabricated, and regional semiconductor production includes all production in the region, including merchant and captive producers and all foreign producers. For instance. North American semiconductor production includes Digital Equipment Corporation and Delco fabs, as well as the North American fabs of Japanese and European companies.

Yearly exchange rate variations can have a significant effect on the 1988through-1995 data in the following tables. For more information about the exchange rates used and their effects, see Appendix B.

The semiconductor industry has a global manufacturing business. Production of semiconductors is constantly shifting among regions as new capital flows to areas of relatively lower capital cost and areas of higher growth in consumption. Dataquest reviews some of the trends and potential impacts for the future.

Historical Semiconductor Production

Table 6-1 shows historical semiconductor production for 1988 through

1994, by region. Dataquest follows a methodology that uses our fab database, estimating the memory, microcomponent, and logic production components separately and estimating net production among regions for foundry activity. This approach provides insight in observing and forecasting production trends.

Because of the reclassification as merchant of the MOS portion of IBM

Microelectronics' business, the captive production figures changed dramatically in 1993. However IBM's bipolar production, which is consumed internally, is still classified as captive by Dataquest.

Captive Semiconductor Production

Semiconductor production from captive manufacturers is estimated to be

$1.98 billion in 1994, down from just over $2 billion in 1993. IBM has restructured and entered the merchant semiconductor market as of 1993.

Dataquest has reclassified IBM's MOS semiconductor production to merchant, but the bipolar products (exclusively used internally) are still reported as captive. This part of IBM's business will be converted to MOS over the next four to five years and had resulted in a lower production figure for captive production in 1994 and the future.

Captive producers may consider moving to merchant to take better advantage of the worldwide growth of semiconductors, leveraging their investments in plant and equipment for higher return in a larger end-user base.

Others may elect to take advantage of the evolving and maturing foundry business, electing to contract out their manufacturing rather than investing in expensive new facilities for their relatively small production base.

We have not, however, included any such movement to merchant or fabless status in our captive production forecast.

SEMM-WW-MT-9502 ©1996 Dataquest 53

54

SemicontJuctor Equipment, Manufacturing, and Materials Worldwide

Table 6-1

Worldwide Semiconductor Production by Region—Historical (Merchant and Captive

Semiconductor Company Sales; Millions of U.S. Dollars)

Total North America

Growth (%)

Percentage of

Worldwide

Merchant

Captive

Total Japan

Growth (%)

Percentage of

Worldwide

Merchant

Captive

Total Europe

Growth (%)

Percentage of

Worldwide

Merchant

Captive

Total Asia/Pacific-ROW

Growth (%)

Percentage of

Worldwide

Merchant

Captive

Worldwide

Growth (%)

Merchant

Growth (%)

Captive

Growth (%)

NA = Not applicable

Source: Dataquest (January 1996)

1988

20,533

20.8

37.3

17,326

3,207

26,732

40.5

48.6

26,388

344

5,854

23.9

10.6

5,277

577

1,868

71.8

3.4

1,868

N A

54,987

31.4

50,859

33.0

4,128

15.2

1989

22,232

8.3

37.6

18,464

3,768

28,527

6.7

48.2

28,119

408

6,451

10.2

10.9

5,782

669

1,974

5.7

3.3

1,974

NA

59,184

7.6

54,339

6.8

4,845

17.4

1990

24,202

8.9

40.8

20,453

3,749

26,384

-7.5

44.5

25,977

407

6,350

-1.6

10.7

5,723

627

2,392

21.2

4.0

2,392

NA

59,328

0.2

54,545

0.4

4,783

-1.3

1991

26,039

7.6

40.4

22,275

3,764

28,338

7.4

44.0

27,925

413

6,979

9.9

10.8

6,396

583

3,097

29.5

4.8

3,097

NA

64,453

8.6

59,693

9.4

4,760

-0.5

1992

29,457

13.1

41.8

25,248

4,209

28,023

-1.1

39.8

27,664

359

8,589

23.1

12.2

7,957

632

4,391

41.8

6.2

4,391

NA

70,460

9.3

65,260

9.3

5,200

9.2

1993

33,446

13.5

38.2

31,745

1,701

34,744

24.0

39.7

34,744

0

11,772

37.1

13.4

11,452

320

7,577

72.6

8.7

1994

40,268

20.4

35.8

38,508

1,760

44,670

28.6

39.7

44,670

0

15,780

34.0

14.0

15,560

220

11,842

56.3

10.5

7,577

NA

87,539

24.2

85,518

31.0

2,021

-61.1

11,842

NA

112,560

28.6

110,580

29.3

1,980

-2.0

CAGR (%)

1988-1994

11.9

14.2

-9.5

8.9

9.2

-100.0

18.0

19.7

-14.8

36.0

36.0

12.7

13.8

-11.5

SEMM-WW-MT-9502

©1996 Dataquest January 22,1996

Semiconductor Production Forecast 55

Tlie IViove toward Asia Continues, Europe Growtii Rests Temporarily

The production trends of the last three years may contain two or three surprises. Of no surprise is the strong growth in Asia/Pacific production, over 10 percent of worldwide production in 1994. The strength of Asian

DRAM producers and the emergence of the foundry market have been and will continue to be the key drivers in that growth. Increased, and somewhat uninhibited, capital spending recently wiU certainly continue this trend.

What may be the first surprise is the strength in European production, expanding from just under 11 percent of the semiconductors produced in

1991 to 14 percent in 1994. This is also remarkable in that the last three years have been good overall growth years, resulting in the region's production more than doubling in three years. Why the move to Europe?

With the region's economies recovering and the PC boom continuing,

Europe has attracted PC production, particularly in the United Kingdom.

Semiconductor production has moved along with the PC, with Intel and

DRAM producers worldwide taking part. Also, the acceleration of telecommunications-related semiconductor production benefits European companies. Dataquest believes that, although multinationals will continue to invest heavily in Europe, the trend is in a holding pattern in part because of the concentration of spending growth in Asia.

The second surprise may be the relative decline in the percentage of the world's production being done in the Uruted States. Over the last two years. North American production decreased from about 42 to 36 percent of the world's production overall. Several factors are at work here. First,

North American multinational companies have been investing heavily overseas. North America has been a net exporter of capital for several years now, as foreign companies have yet to balance the scales with investments inside the United States. This trend should stabilize over the next several years because Japanese and Korean companies have started to accelerate their investment in the United States.

Second, although U.S. companies are recognized as technology leaders, they have recently begim calling on foreign producers to manufacture their products in the foundry market. Fabless companies have been the key drivers of the market to this point, but starting in 1994 we saw a major shift in increasing use of foimdries by IDMs. In 1995 we expect U.S. IDMs to increase their foundry purchases by 100 percent, with most of this production overseas. The imbalance in the concentration of foundry capacity is starting to make U.S. semiconductor companies a little nervous and they have actually impressed upon key fovmdry suppliers the need to begin building production in the United States. In fact, TSMC recentiy announced plans to build a major fab in the United States, and we believe that other foundry companies are likely to follow. Clearly, users of foimdry have stated a preference for close access to the fab. Foundry providers with capacity in the United States are likely to have a more stable customer base.

And third, although Japanese and European companies have invested somewhat outside their own countries, these companies have remained

"patriots of the domestic economy" and have kept the vast majority of

SEI\/ll^-WW-l\/lT-9502 ©1996 Dataquest January 22,1996

56 Semiconductor Equipment, IVIanufacturing, and IVIaterials Worldwide their investment within their region. This, along with the strong DRAM market over the last two years, has stabilized the Japanese production proportion over the last two years, keeping the share of the production market at about 40 percent (this could be surprise No. 3). However the DRAM market is cyclical, and Japanese foundries will feel pressure from Asian producers, so we expect a resumption of the gradual decay in the base of production in Japan through the rest of the decade.

Semiconductor Production Trends: Accelerating Sliift to Asia/Pacific

Table 6-2 shows forecast semiconductor production by region for the period from 1994 through 2000. The major trend is the growth of the Asia/

Pacific region, mostly at the expense of Japan. Companies such as TSMC and UMC in Taiwan, Chartered Semiconductor in Singapore, and SubMicron Technology in Thailand will accelerate capacity rapidly beginning in

1996. New DRAM companies (Vanguard, Powerchip Semiconductor, and

Nan Ya Technology) have sprung up in Taiwan and likely to further erode

Japanese DRAM share. By 2000, Dataquest believes that Asia/Pacific-

ROW will expand to over 14 percent on a revenue basis.

North America will remain steady on a percentage basis as the lower cost of capital and clear leadership in technology and design motivate companies to invest in the United States. Europe production share is expected to expand slightly, with a product mix shifting toward a higher memory component, driven by the need for proximity to PC production as more

DRAM capacity is added by large international companies over the next several years. Japan's share of production is likely to continue to erode as the cost of capital remains high and as Japanese companies increasingly invest in capacity overseas.

/

SEI\/ll\/l-WW-MT-9502 ©1996 Dataquest January 22,1996

Semiconductor Production Forecast 57

Table 6-2

Worldwide Semiconductor Production by Region—Forecast (Merchant and Captive

Semiconductor Company Sales; Millions of U.S. Dollars)

CAGR (%)

1994-2000

19.9

Total North America

Growth (%)

Percentage of

Worldwide

Merchar\t

Captive

Total Japan

Growth (%)

Percentage of

Worldwide

Merchant

Captive

Total Europe

Growth (%)

Percentage of

Worldwide

Merchant

Captive

Total Asia/

Pacific-ROW

Growth (%)

Percentage of

Worldwide

Merchant

Captive

Worldwide

Growth (%)

Merchant

Growth (%)

Captive

Growth (%)

NA = Not applicable

Source: Dataquest (January 1996)

1994

40,268

20.4

35.8

38,508

1,760

44,670

28.6

39.7

44,670

NA

15,780

34.0

14.0

15,560

220

11,842

56.3

10.5

11,842

NA

112,560

28.6

110,580

29.3

1,980

-2.0

1995

55,181

37.0

36.4

53,474

1,707

58,867

31.8

38.8

58,867

NA

20,381

29.2

13.4

20,221

160

17,226

45.5

11.4

1996

66,049

19.7

35.8

64,384

1,665

70,968

20.6

38.4

70,968

NA

25,353

24.4

13.7

25,241

112

22,315

29.5

12.1

1997

75,487

14.3

35.6

73,837

1,650

78,885

11.2

37.2

78,885

NA

29,951

18.1

14.1

29,871

80

27,768

24.4

13.1

1998

85,641

13.5

35.6

83,987

1,654

87,089

10.4

36.2

87,089

NA

34,406

14.9

14.3

34,358

48

33,165

19.4

13.8

17,226

N A

151,655

34.7

149,788

35.5

1,867

-5.7

22,315

NA

184,685

21.8

182,908

22.1

1,777

-4.8

27,768

NA

212,091

14.8

210,361

15.0

1,730

-2.6

33,165

NA

240,301

13.3

238,599

13.4

1,702

-1.6

1999

99,595

16.3

35.9

97,889

1,706

99,268

14.0

35.8

99,268

NA

40,013

16.3

14.4

39,983

30

38,604

16.4

13.9

2000

119,715

20.2

35.9

118,009

1,706

117,678

18.5

35.3

117,678

NA

48,095

20.2

14.4

48,065

30

47,734

23.6

14.3

38,604

N A

277,480

15.5

275,744

15.6

1,736

2.0

47,734

N A

333,222

20.1

331,486

20.2

1,736

0.0

20.5

-0.5

17.5

17.5

20.4

20.7

-28.3

26.2

26.2

NA

19.8

20.1

-2.2

SEIVIIVI-WW-IViT-9502 ©1996 Dataquest January 22,1996

58 Semiconductor Equipment, l\/lanufacturing, and Materials Worldwide

Dataquest Perspective

Where the PC goes, so go semiconductors. This is true from the perspective of the business forecast as well as the production line. Europe and

Asia/Pacific, with very large capital spending upticks over the last several years (continuing in 1995 in Asia) will gain share in world production over the next several years.

The shifts and currents in semiconductor production trends mean that equipment and material suppliers will absolutely need a global presence in every sense of the word to remain competitive in the market. Product supply and support can no longer concentrate on local trends, because all major semiconductor companies have made it clear they are investing on a worldwide basis.

Taiwan is clearly the new major production growth area. We would expect

Malaysia and Thailand to become the next major growth countries in three to five years. Evidence of this includes recent joint-venture fab announcements by Texas Instruments and others. Silicon plants are now being strategically placed, such as SEH's Malaysian plant and its recently announced joint venture in Taiwan, Komatsu's joint venture with Formosa

Plastics in Taiwan, and MEMC's joint ventures in both Korea (Posco-Hiils) and Taiwan (Taisil).

Further, the concept of contract manufacturing in semiconductors is clearly here to stay. Equipment and material suppliers could find themselves selling their technical products to an international team from several companies, including the manufacturer and the designer. However, the emergence of the dedicated foundry company, taking ownership of the process and manufacturing flow, will tend to centralize this activity.

Dataquest has started a research service in Semiconductor Contract Manufacturing, with a major report planned for the end of January 1996. This report will explore the key trends in contract manufacturing and foundries, including technology trends and supply/demand balance through the decade.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Appendix A

Economic Assumptions, Fourtli Quarter 1995

Optimism of International Business Executives Continues to Decline

Dun & Bradstreet's most recent quarterly survey of international business executives revealed slipping optimism about the near-term growth of net sales and net profits worldwide. Survey respondents also disclosed reduced optimism about near-term growth of selling prices, employment, and inventories worldwide. The continued decline in expectatioris furthered the trend toward reduced optimism that emerged at the beginning of 1995. Global business executives now anticipate more moderate expansion than they did a year ago when optimism was at its peak for the current business cycle.

Despite Slipping Expectations, tiie Near-Term Outlook for the World Economy

Remains Strong

Although the number of bears grew among international executives, the results of the Dun & Bradstreet survey continued to indicate that executives, on balance, are bullish about the world economy. Solid growth in sales and profits is expected in nearly every region of the world, and most regions anticipate some increase in selling prices accompanied by steady employment and inventories. Only Japanese, Mexican, and Brazilian executives expressed contrarian pessimistic near-term views. In these coimtries, the outlook continues downcast, especially for selling prices, employment, and inventories.

The Long-Term Outlook for the World Economy Likewise Remains Strong

The latest long-term forecasts from Dun & Bradstreet subsidiary A.C.

Nielsen also indicate that the world economy will prosper through 2000.

These forecasts indicate that the world's developed economies will continue to experience moderate but healthy economic growth through 2000.

Growth among developed economies is forecast to average 2.5 percent in

1995, following 2.9 percent growth in 1994. Growth is anticipated to increase gradually to the end of the decade, reaching 3.1 percent by 2000.

Growth among the world's emerging economies will continue to vary by region. Emerging Asia/Pacific economies should continue to post impressive rates of growth. Growth among these economies is forecast to average

7.9 percent in 1995, following 8.2 percent growth in 1994. Growth is anticipated to slow somewhat by 2000 but even then will still be an enviable 6.7 percent annually. Emerging Latin American and Eastern European economies will continue to lag behind their Asia/Pacific covmterparts. Growth among emerging Latin American economies (Argentina, Brazil, Chile,

Mexico, and Venezuela) is forecast to be 0.5 percent in 1995, largely because of anticipated contractions in Mexico and Argentina. Growth is anticipated to accelerate through the end of the decade, rising to 5.5 percent by 2000. Emerging Eastern European economies are forecast to experience 3.7 percent growth in 1995, up from 0.4 percent in 1994. Growth among these economies is expected to accelerate, reaching 4.4 percent by

2000.

SEMM-WW-MT-9502 ©1996 Dataquest 59

60

Semiconductor Equipment, l^nufacturing, and Materials Worldwide

A summary of gross domestic product (GDP) and constmier price index

(CPI) forecasts for the G7 nations is shown in Tables A-1 and A-2, respectively.

Table A-1

Gross Domestic Product/Gross National Product Growth Rates: Outlook as of

December 15,1995—Constant Prices and Exchange Rates, Local Currencies (Percent)

Year

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Source: A.C. Nielsen

United States

1.2

-0.6

2.6

3.1

4.1

3.3

3.1

3.7

3.2

3.1

3.0

Canada

-0.2

-1.8

0.6

2.2

4.6

2.2

2.5

3.0

2.9

2.8

2.8

Japan

4.8

4.3

1.1

-0.2

0.5

0.5

2.0

2.6

3.2

3.5

3.2

France

2.5

0.8

1.3

-1.5

2.9

2.7

2.3

2.9

3.2

3.1

3.0

Germany

NA

NA

1.8

-1.2

3.0

2.3

2.5

2.8

3.1

2.8

2.7

Italy

2.1

1.3

0.7

-1.2

2.2

3.0

2.7

2.6

2.5

2.5

2.5

United

Kingdom

0.6

-2.3

-0.5

2.1

3.9

2.8

2.8

2.8

2.7

2.7

2.7

Table A-2

Consumer Price Index Growth Rates: Outlook as of December 15,1995 (Percentage Change)

Year

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

NA = Not available

Source: A.C. Nielsen

United States

5.4

4.2

3.0

2.9

2.6

2.9

2.8

2.8

2.6

2.5

2.4

Canada

4.8

5.6

1.5

1.8

0.2

2.2

2.5

2.4

2.2

2.1

2.0

Japan

3.1

3.3

1.6

1.3

0.7

0.0

0.2

0.5

0.7

0.9

1.0

France

3.4

3.1

2.4

2.1

1.7

1.9

2.5

2.4

2.3

2.2

2.1

Germany

NA

NA

4.0

4.1

1.8

2.0

2.3

2.4

2.2

2.0

2.0

Italy

6.1

6.4

5.4

4.2

3.9

5.3

4.5

3.9

3.5

3.2

2.9

United

Kingdom

9.5

5.9

3.7

1.6

2.4

3.5

3.0

2.8

2.6

2.5

2.4

SEi\/lM-WW-MT-9502 ©1996Dataquest January 22,1996

Economic Assumptions, Fourth Quarter 1995 61

Americas: Continued Stability for the United States and Canada; Increased

Uncertainty for Mexico and Brazil

Dun & Bradstreet survey results continue to indicate high optimism in both the United States and Canada. Executives in both countries anticipate stable conditions over the near term with only moderately slower economic expansion and sustained high levels of employment. In contrast, conditions in Mexico and Brazil appear increasingly uncertain. Once again, continuing economic and political instability in Mexico made it impossible to gattier any meaningful survey results for the country. In Brazil, all survey indexes registered sharp declines for the second survey in a row. Brazilian expectations for the near-term have now turned deeply pessimistic for net profits and employment. Longer-term growth forecasts for the Americas are as follows:

• The United States' GDP is now forecast to grow 3.3 percent in 1995, down from 4.1 percent growth in 1994. Growth is anticipated to decelerate to 3.1 percent in 1996. Growth is then expected to quicken slightly, averaging 3.2 percent annually through 2000.

• Canada's GDP is now forecast to grow 2.2 percent in 1995, down from

4.6 percent growth in 1994. Grow&i is anticipated to increase to 2.5 percent in 1996. Growth is then expected to further accelerate through 2000, averaging 2.9 percent annually.

• Mexico's GDP is now forecast to contract 5.5 percent in 1995, a sharp reversal from the 3.7 percent growth experienced in 1994. Growth of

2.0 percent is anticipated in 1996. Growth is then expected to accelerate, averaging nearly 5.0 percent annually through 2000.

• Brazil's GDP is now forecast to grow 5.0 percent in 1995, down from

5.7 percent growth in 1994. Growth is anticipated to slow further to

4.0 percent in 1996. Growth is then expected to accelerate through 2000, averaging just over 5.5 percent armually.

Europe: Optimism Remains Buoyant

Dun & Bradstreet survey results reveal buoyant near-term optimism among European executives. All survey indexes remained positive, with some indexes inching up slightly over the values observed in the last survey. Expectations of significantly faster near-term growth in Germany and

Italy offset lower levels of near-term optimism in the United Kingdom and

Belgium. Longer-term growth forecasts for Europe are as follows:

• France's GDP is now forecast to grow 2.7 percent in 1995, down from

2.9 percent growth in 1994. Growth is anticipated to slow further to

2.3 percent in 1996. Growth is then expected to increase, averaging s u b t l y over 3.0 percent annually through 2000.

• Germany's GDP is now forecast to grow 2.3 percent in 1995, down from

3.0 percent in 1994. Growth is anticipated to increase to 2.5 percent in

1996. Growth is then expected to further accelerate through 2000, averaging nearly 2.9 percent Annually.

SEMI\yi-WW-IVlT-9502 ©1996 Dataquest January 22,1996

62 Semicondtjctor Equipment. Manufacturing, and Materials Worldwide

Italy's GDP is now forecast to grow 3.0 percent in 1995, up from 2.2 percent growth in 1994. Growth is anticipated to slow somewhat to 2.7 percent in 1996. Growth is then expected to level off and average slightly over 2.6 percent annually through 2000.

The United Kingdom's GDP is now forecast to grow 2.8 percent in 1995, down from 3.9 percent growth in 1994. Growth is anticipated to remain the same in 1996 at 2.8 percent. Growth is then expected to remain steady through 2000, averaging just over 2.7 percent annually.

Japan and Asia/Pacific: Outlook in Japan Remains Glum

Dun & Bradstreet survey results indicate that pessimism continues to darken Japanese expectations. The survey indexes for selling prices and inventories remain negative. The positive indexes for net sales and employment recorded in the last survey registered marked declines, witii the latter actually turning negative. Although the yen has weakened somewhat since summer, the relative strength of the yen against the U.S. dollar and other currencies continues to impact both Japan and the rest of

Asia. Imports from the United States and elsewhere induced by the strong yen continue to depress Japanese prices and profit margins. Added to this, the strong yen continues to depress exports and encourage the migration of Japanese manufacturing to other parts of Asia. Longer-term forecasts for Japan and Asia/Pacific are as follows:

• Japan's GDP is now forecast to grow 0.5 percent in 1995, unchanged from 1994. Growth is anticipated to accelerate to 2.0 percent in 1996.

Growth is then expected to further accelerate through 2000, averaging

3.1 percent annually.

• China's GDP is now forecast to grow 9.5 percent in 1995, down from

11.8 percent in 1994. Growth is expected to slow in 1996 to 9.0 percent.

Further deceleration of growth is expected through 2000 as the Chinese economy matures.

• Taiwan's GDP is now forecast to grow 6.9 percent in 1995, dov^m from

6.5 percent growth in 1994. Growth is anticipated to slow back to

6.5 percent in 1996. Growth is then anticipated to remain at or near this level through 2000.

• South Korea's GDP is now forecast to grow 9.0 percent in 1995, up from

8.4 percent in 1994. Growth is anticipated to decelerate to 7.5 percent in

1996. Growth is then expected to settle down to 7.0 percent annually through 2000.

Dun & Bradstreet's quarterly International Survey of Business Expectations is modeled on its quarterly Survey of U.S. Business Expectations, which has been conducted in the United States since 1947. Survey participants are asked if they expect increases, decreases, or no change in their sales, profits, prices, inventories, and employment levels in the upcoming quarter, compared with the same quarter a year ago. Results of the U.S. survey have proven highly accurate, with quarterly forecasts closely paralleling actual performance.

SEMM-WW-MT-9502 ©1996 Dataquest January 22,1996

Economic Assumptions, Fourth Quarter 1995 63

The current international survey, completed in September 1995, uses similar sampling and interviewing procedures in all countries. Hence, the results provide unique "apples-to-apples" comparisons of trends in business expectations worldwide. The Dun & Bradstreet index figures discussed here represent the net percentage of survey respondents expecting higher sales, profits, and so on. The indexes are calculated by subtracting the percentage of respondents expecting decreases from those expecting increases. GDP and CPI growth rates quoted here are A.C. Nielsen's most current forecasts from December 1995.

SEiy/il\/l-WW-l\/IT-9502 ©1996 Dataquest January 22,1996

Appendix B

Exchange Rates

Dataquest does not forecast exchange rates per se; however, we do forecast semiconductor-related markets in several regions of the world, and we use the U.S. dollar as a common currency for intermarket comparisons and aggregation. In general, in the forecast period Dataquest assumes that the actual exchange rate of the full month before the month in which the forecast-input assumptions are set will apply throughout all future months of the forecast interval. For the forecasts presented here:

• Actual monthly exchange rates were used for all months in the historical interval up to September 1995.

• The September 1995 exchange rate was then assumed to hold for

October through December 1995 and throughout 1996 to 2000.

Dataquest uses an average annual exchange rate to convert annual revenue from local currency values to U.S. dollar values. Table B-1 outlines the rates used in the forecasts presented here.

Table B-1

Exchange Rates per U.S. Dollar

Country

Argentina (Peso)

Australia (Dollar)

Austria (Schilling)

Bangladesh (Taka)

Belgium (Franc)

Brazil (Cruzeiro)

Bulgaria (Lev)

Canada (Dollar)

Chile (Peso)

China (Yuan)

Colombia (Peso)

C5^rus (Pound)

Czech Republic (Koruny)

Denmark (Krone)

Dominican Republic (Peso)

Ecuador (Sucre)

ECU

Finland (Markka)

France (Franc)

Germany (Mark)

Greece (Drachma)

Guatemala (Quetzal)

Hong Kong (Dollar)

1 Himgary (Forint)

1994

N A

N A

11.40

NA

33.36

N A

NA

N A

0.84

5.21

5.54

1.62

242.06

N A

NA

N A

8.54

NA

N A •

N A

6.35

NA

7.73

N A

N A

394.27

8.35

903.68

N A

N A

5.61

N A

1995*

1.00

1.35

10.09

N A

29.52

0.91

N A

N A

0.77

4.38

4.99

1.44

231.10

N A

7.74

N A

U.S. $ Appreciation

1994-1995 (%)*

N M

N M

-11.49

N M

-11.53

N M

N M

N M

N M

-2.16

N M

N M

N M

-11.62

N M

N M

-8.28

-15.94

-9.92

-11.43

-4.53

N M

0.11

N M

1996-2000*

1.00

1.33

10.18

40.36

29.76

0.95

68.05

1.34

397.49

8.32

957.86

0.45

26.65

5.62

13.81

2,637.58

0.78

4.33

4.98

1.45

234.43

5.84

7.74

132.63

(Continued)

SEMM-WW-MT-9502

©1996 Dataquest 65

66

Semiconductor Equipment, IVIanufacturing, and l\/laterials Worldwide

Table B-1 (Continued)

Exchange Rates per U.S. Dollar

Country

India (Rupee)

Indonesia (Rupiah)

Ireland (Punt)

Israel (New SheqaUm)

Italy (Lira)

Japan (Yen)

Malaysia (Ringgit)

Mexico (Peso)

Morocco (Dirham)

Netherlands (Guilder)

N e w Zealand (Dollar)

N o r w a y (Krone)

Pakistan (Rupee)

Paraguay (Guarani)

Peru (New Sol)

Philippines (Peso)

Poland (Zloty)

Portugal (Escudo)

Romania (Lei)

Russia (Ruble)

Singapore (DoUar)

Slovakia (New Koruna)

South Africa (Rand)

South Korea (Won)

Spain (Peseta)

Sweden (Krona)

Switzerland (Franc)

Taiwan P o U a r )

Thailand (Baht)

Tunisia (Dinar)

Turkey (Lira)

United Kingdom (Poimd)

Uruguay (Peso)

Venezuela (Bolivar)

\ ^ e t n a m (Dong)

Zimbabwe (Dollar)

•Estimated

NA = Not available; not traci<ed until 1995

NM = Not meaningful

Source: Dataquest (December 1995)

NA

NA

165.63

N A

N A

1.53

N A

NA

802.84

133.48

7.70

1.37

1994

31.15

NA

0.67

NA

1,609.34

101.81

2.62

NA

N A

1.82

NA

7.04

NA

NA

NA

26.45

25.36

N A

N A

0.65

NA

NA

NA

NA

1995*

31.91

2,240.77

0.62

N A

1,634.59

93.04

2.50

6.17

N A

1.61

1.52

6.34

N A

NA

2.24

25.69

N A

150.08

N A

N A

1.43

N A

N A

770.75

124.91

7.25

1.19

26.49

24.89

N A

44,573.38

0.63

N A

N A

N A

N A

U.S. $ Appreciation

1994-1995 (%)*

2.44

N M

-6.70

N M

1.57

-8.61

-4.64

N M

N M

-11.54

N M

-10.06

NM

NM

NM

NM

NM

-9.39

-6.41

-5.80

-13.20

0.15

-1.85

NM

N M

-3.33

NM

N M

N M

NM

N M

N M

-6.46

N M

NM

-4.00

25.99

2.38

151.17

2,090.10

4,450.87

1.42

30.37

3.65

770.69

124.41

7.12

1.18

27.28

25.07

0.94

48,473.10

0.64

6.63

169.79

11,027.79

8.69

1996-2000*

1

32.86

2,266.01

0.62

3.03

1,620.50

98.52

2.50

6.32

8.52

1.62

1.53

6.33

31.40

1,965.99

2.25

©1996 Dataquest

SEMM-WW-IVIT-9502

January 22,1996

For More Information...

Clark Fuhs, Director/Principal Analyst (408) 468-8375

Internet address [email protected]

Via fax (408) 954-1780

The content of this report represents our interpietation and analysis of information generally available to the public

DataQuest or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in part to other parties shall be made upon the written and express consent of Dataquest.

©1996 Dataquest—Reproduction Prohibited

AGartnerGrOUpCompany Dataquest is a registered trademark of A.C. Nielsen Company

DATAQUEST WORLDWIDE OFFICES

NORTH AMERICA

Worldwide Headquarteis

251 River Oaks Parkway

San Jose, California 95134-1913

United States

Phone: 1-408-468-8000

Facsimile: 1-408-954-1780

East Coast Headquarters

Nine Technology Drive

RO. Box 5093

Westborough, Massachusetts 01581-5093

United States

Phone: 1-508-871-5555

Facsimile: 1-508-871-6262

Dataquest Global Events

3990 Westerly Place, Suite 100

Newport Beach, Califoriua 92660

United States

Phone: 1-714-476-9117

Facsimile: 1-714-476-9969

EUROPE

European Headquarters

Holmers Farm Way

High Wycombe, Bucks HP12 4XH

United Kingdom

Phone:+44 1494 422 722

Facsimile: +441494 422 742

Dataquest France

Immeuble Defense Berg^res

345, avenue Georges C16menceau

TSA 40002

92882 - Nanterre CTC Cedex 9

France

Phone: +33 1 41 35 13 00

Facsimile: +331 41 35 13 13

Dataquest Germany

Kronstadter Strasse 9

81677 Munchen

Germany

Phone: +49 89 93 09 09 0

Facsimile: +49 89 93 03 27 7

Sales Offices:

Washington, EXZ (Federal)

New York, NY (Financial)

Dallas, TX

LATIN AMERICA

Research Affiliates and Sales Offices:

Buenos Aires, Argentina

Sao Paulo, Brazil

Santiago, Chile

Mexico City, Mexico

Sales Offices:

Brussels, Belgium

Kfar Saba, Israel

Milan, Italy

Randburg, South Africa

Madrid, Spain

JAPAN

Japan Headquarters

Shinkawa Sanko Building

6th Floor

1-3-17, ShirJcawa

Chuo-ku, Tokyo 104

Japan

Phone: 81-3-5566-0411

Facsimile: 81-3-5566-0425

ASIA/PACIFIC

Asia/Pacific Headquarters

7/F China Underwriters Centre

88 Gloucester Road

Wan Chai

Hong Kong

Phone: 852-2824-6168

Facsimile: 852-2824-6138

Dataquest Korea

Suite 2407, Trade Tower

159 Samsung-dong, Kangnam-gu

Seoul 135-729

Korea

Phone: 822-551-1331

FacsimUe: 822-551-1330

Dataquest Taiwan

llF-2, No. 188, Section 5

Nan King East Road

Taipei

Taiwan, R.O.C.

Phone: 8862-756-0389

Facsimile: 8862-756-2663

Dataquest Singapore

105 Cecil Street

#06-01/02 The Octagon

Singapore 069534

Phone: 65-227-1213

Facsimile: 65-227-4607

Dataquest Thailand

12/F, Vanissa Building

29 Soi Chidlom

Ploenchit Road

Patumwan, Bangkok 10330

Thailand

Phone: 662-655-0577

Facsimile: 662-655-0576

Research Affiliates and Sales Offices:

Melbourne, Australia

Beijing, China

DataQuest

A Gartner Group Company

®19^ Dataquest

SEIyM-IW-MT-9502

~ ;. Maria V a l e n z u e l a

r e q u e s t I n c o r p o r a t e d

Dari - l l C O

-•-INTEilNAL DIST.—

Qty: 1

DataQuest

Asia/Pacific Fab Database

Market Statistics

Program: Semiconductor Equipment, IVIanufacturing, and Materials Worldwide

Product Code: SEI\/ll\/l-WW-MS-9506

Publication Date: December 25,1995

Filing: Market Analysis

Asia/Pacific Fab Database

Market Statistics

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-9506

Publication Date: December 25,1995

Filing: Market Analysis

Asia/Pacific Fab Database

Table of Contents

Page

Asia/Pacific Fab Database 1

Background 1

Research Methodology 1

General Definitions 1

Worldwide Geographic Region Definitions and Regional 1

RoU-Ups

Americas 1

Japan 1

Europe, Africa, and Middle East 1

Asia/Pacific 2

Definition of Table Columns 2

SEMM-WW-MS-9506 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, Manufacturing, and Materials Woridwide

List of Tables

Table Page

1 Asia/Pacific Existing Pilot and Production Fab Lines 7

2 Asia/Pacific Future Pilot and Production Fab Lines 11

Note: All tables show estimated data.

SEMM-WW-MS-9506 ©1995 Dataquest Incorporated December 25,1995

Asia/Pacific Fab Database

Background

This document contains the Asia/Pacific portion of Dataquest's wafer fab database. The Semiconductor Equipment, Materials, and Manufacturing Worldwide (SEMM) program uses both primary and secondary research to update this data. The tables in this report cover both merchant and captive production and pilot-line facilities, although our surveys and database also include R&D fabs.

Research Methodology

Dataquest conducts extensive annual surveys, complemented with quarterly secondary research. This data is then supplemented and cross-checked with various other information sources.

General Definitions

Fab line: A fab line is a semiconductor processing facility equipped for all front-end wafer manufacturing. Occasionally, there are two or more separate product-specific fab lines or wafer sizes in a single cleanroom.

In this situation, Dataquest documents the cleanroom as separate fab lines if the company dedicates equipment to each wafer size or product line. Therefore, a company may operate many fab lines at one location.

Front-end wafer processing: Dataquest defines front-end wafer processing as all steps involved in semiconductor processing, beginning with initial oxide and ending at wafer probe.

Production fab: A wafer fab capable of front-end processing more than

1,250 wafers per week is a production fab (type = F).

Pilot fab: A wafer fab capable of front-end processing 1,250 or fewer wafers per week is a pilot fab (type = P).

Worldwide Geographic Region Definitions and Regional Roll-Ups

Americas

Includes Central America (all nations), Canada, Mexico, United States and Puerto Rico, and South America (all nations).

Japan

Japan is the only single-country region.

Europe, Africa, and Middle East

Includes Africa (all nations), Albania, Andorra, Armenia, Azerbaijan,

Belarus, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia,

Finland, France, Georgia, Germany, Gibraltar, Hungary, Iceland, Israel,

Italy, Kazakhstan, Kyrgyzstan, Latvia, Liechtenstein, Lithuania,

Luxembourg, Malta, Middle East (all nations), Moldova, Monaco,

SEMM-WW-MS-9506 ©1995 Dataquest Incotpotated t

Semiconductor Equipment, IVIanulacturing, and IVlaterials Worldwide

Netherlands, Norway, Poland, Romania, Russia, San Marino,

Scandinavia, Slovakia, Spain, Sweden, Sweden, Switzerland, Tajikistan,

Turkey, Turkmenistan, Ukraine, United Kingdom, Uzbekistan, Vatican

City, and Yugoslavia (all nations within the former Yugoslavia).

Asia/Pacific

Includes Australia, Bangladesh, Cambodia, China, Hong Kong, India,

Indonesia, Laos, Malaysia, Maldives, Myarunar, Nepal, New Zealand,

Pakistan, Philippines, Singapore, South Korea, Sri Lanka, Taiwan,

Thailand, and Vietnam.

Definition of Table Columns

Products Produced contains details for seven product categories. The nomenclature used within the seven product groups of the fab database is as follows, with definitions where warranted:

• Analog

• A / D D/A: Analog-to-digital, digital-to-analog converter o AUTOMOTIVE: Dedicated to automobile applications

o CODEC: Coder/decoder

o INTERFACE: Interface IC

o LIN: Linear/analog device o MDIODE: Microwave diode a MESFET: Metal semiconductor field-effect transistor o MEET: Microwave field-effect transistor o MIXSIG ASIC: Mixed-signal/linear ASIC o MODEM: Modulator/demodulator

a MMIC: Monolithic microwave IC

a OP AMP: Operational amplifier

a PWR IC: Power IC o REG: Voltage regulator

a SMART PWR: Smart power

a SWITCHES: Switching device o TELECOM: Telecommunications chip

• Memory

o DRAM: Dynamic RAM

a EEPROM or E2: Electrically erasable PROM o EPROM: Ultraviolet erasable PROM

SEMM-WW-MS-9506 ©1995 Dataquest Incorporated December 25,1995

Asia/Pacific Fab Database p FERRAM: Ferroelectric RAM a FIFO: First-in/first-out memory a FLASH: Flash memory o MEM: Memory

a NVMEM: Nonvolatile memory (ROM, PROM, EPROM, EEPROM,

FERRAM)

a PROM: Programmable ROM o RAM: Random-access memory n ROM: Read-only memory o SPMEM: Other specialty memory (such as dual-port, shift-register, color lookup)

a SRAM: Static RAM

a VRAM: Video RAM

• Micrologic

• ASSP: Application-specific standard product o BIT: Bit slice (subset of MPU functions)

a DSP: Digital signal processor o LISP: 32-bit list instruction set processor for AI o MCU: Microcontroller unit

a MPR: Microperipheral o MPRCOM: MPR digital communication (ISDN, LAN, UART, modem) o MPU: Microprocessor unit o RISC: Reduced-instruction-set computation 32-bit MPU

• Standard logic

a LOG or LOGIC: Standard logic

• ASIC logic a ARRAYS: Gate array

• ASIC: Application-specific IC o CBIC: Cell-based IC o CUSTOM: Full-custom IC (single user) o FPGA: Field-programmable gate array o PLD: Programmable logic device

SEMI\/l-WW-MS-9506 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, IVIanufacturing, and Materials Woridwide

• Discrete o DIODE o DIS or DISCRETE: Discrete

a PET: Field-effect transistor o GTO: Gate hirn-off thyristor o HEMT: High-electron-mobility transistor

a IGBT: Insulated gate bipolar transistor o MOSFET: MOS-based field-effect transistor o PWR TRAN: Power transistor o RECTIHER o RF: Radio frequency

a SCR: Schottky rectifier o SENSOR

a SST: Small-signal transistor o THYRISTOR o TRAN: Transistor o ZENER DIODE

• Optoelectronic

n CCD: Charge-coupled device (imaging)

a COUPLER: Photocoupler o lED: Infrared-emitting diode o IMAGE SENSOR o LASER: Semiconductor laser or laser IC

a LED: Light-emitting diode

a OPTO: Optoelectronic

• PDIODE: Photo diode

a PTRAN: Photo transistor o SAW: Surface acoustic wave device o SIT IMAGE SENSOR: Static induction transistor image sensor

Process Technology column lists four major types of technologies. This column also lists uncommon technologies with information on well types, logic structure, and number of metal levels. Definitions used in the "Process Technology" column are as follows:

• MOS (silicon-based) o CMOS: Complementary metal-oxide semiconductor

SEI\/IM-WW-MS-9506 ©1995 Dataquest Incorporated December 25,1995

Asia/Pacific Fab Database

a MOS: N-chaimel metal-oxide semiconductor (NMOS) and p-channel metal-oxide semiconductor (PMOS).

a Ml: Single-level metal

a M2: Double-level metal a M3: Triple-level metal

a N-WELL o P-WELL

a POLYl: Single-level polysilicon

D POLY2: Double-level polysilicon o POLYS: Triple-level polysilicon

• BiCMOS (silicon-based) o BiCMOS: Bipolar and CMOS combined on a chip o BiMOS: Bipolar and MOS combined on a chip o ECL I/O: ECL input/output o TTL I/O: TTL input/output

• Bipolar (silicon-based) a BIP or BIPOLAR: Bipolar o ECL: Emitter-coupled logic o TTL: Transistor-transistor logic

a STTL: Schottky TTL

• Gallium arsenide and other compound semiconductor materials a GaAs: Gallium arsenide o AlGaAs: Galliunn aluminum arsenide

• GaAs on Si: Gallium arsenide on silicon

a GaP: Gallium phosphide o HgCdTe: Mercuric cadmium telluride

• InAs: Indiim\ arsenide

a InGaAs: Indium gallium arsenide o InP: Indium phosphide o InSb: Indium antimony

• LiNbOS: Lithium niobate

a SOS: Silicon on sapphire

SEMM-WW-MS-9506 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, Manufacturing, and Materials Woridwide

Minimum Geometry is the smallest feature attainable in production volumes, measured in microns, at the critical mask layers.

Wafer Diameter represents the wafer diameter usually expressed colloquially in inches. However, for wafers greater than 3 inches in diameter, the colloquial expression becomes grossly inaccurate. When calculating square inches, Dataquest uses the following approximations:

• Stated diameter 4 inches (100mm) = Approximate diameter

3.938 inches

• Stated diameter 5 inches (125mm) = Approximate diameter

4.922 inches

• Stated diameter 6 inches (150mm) = Approximate diameter

5.906 inches

• Stated diameter 8 inches (200mm) = Approximate diameter 7.87 inches

Estimated Maximum Wafer Starts per Month is the equipment-limited wafer start capacity per four-week period. Start capacity is limited only by the installed equipment in the fab and the complexity of the process it runs, not by ciirrent staffing or the number of shifts operating.

SEMM-WW-MS-9506 ©1995 Dataquest Incorporated December 25,1995

o

CD o

CO f \ 3

cn

CO

CO

cn

CO

cp

CO

en

O

<y>

@

C D

CO

on

K.

D> s:i

o

• 3 o

Table 1

Asia/Pacific Existing Pilot and Production Fab Lines (Including Fabs Beginning Operation duri

City or District

SCIENCE PARK

Country

TAIWAN

Fab Name Products Produced

FABl NA

Process

Technology

CMOS

Est.

Minimum

Geometry

(Microns)

Wafer

Diameter

(In.)

Year of Initial

Production

1988

Company

ADVANCED

MICROELECTRONICS

PRODUCTS

ADVANCED S/C

MANUPOF

SHANGHAI (ASMC)

AMALGAMATED

W1REI.R5S

ASMC {Fonni>rIy

Shfiilgliai Radio No. 7)

BEIJING N0.3

SEMICONDUCTOR

APPLIANCE

CHARTERED

SEMICONDUCKJR

SHANGHAI

SYDNEY

SHANGHAI

BEIJING

SINGAPORE

CHINA

CHINA

CHINA

TAB I

AUSTRALIA NA

FABl

NA

SINGAPORE FABl

SINGAPORE FABD

LIN DIGTTAL IC FOR CMOS

T,V. EPROM

ASIC

LIN DEGPTAL IC for

TV EPROM

REGULATORS,

TELECOM

CMOS

CMOS MOS

MPU MCU MPR

ASIC ANALOG

SRAM E EPROM:

OTHER MEM

LOGIC SRAM

CMOS

CMOS

CMOS

(1,R

1,5

0.8

5

0.5

0.5

1992

1989

1995

1988

1989

1995 CIIARTERED

SEMICONDUCTOR

DAEWOO

DAEWOO

DONGSUNG

EPJSIL

TECHNOLOGIES, INC,

EPISIL

TECHNOLOGIES, INC.

EKSO

FINE MICROELECT.

GENERAL

INSTRUMENT

HOLTEK

SINGAPORE

GURO-DONG, SEOUL

GURO-DONC, SEOUI.

ECHUN, KYUNGKI-DO

SCIENCE PARK

SCIENCE PARK

HStNCHU

SCIENCE PARK

HSI TIEN CrTY

SCIENCE PARK

KOREA

KOREA

KOREA

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

BIPOLAR

MOS

BIPOLAR

FABl

FAB 2

FABl

NA

NA

FABl

ANALtJG

CUSTOM ASIC

DIODE, RECTIFIER

EPl WAFER r-OR

M O D E T R A N & I C i

BIPOLAR IC, POWER

TRAN

4Mb 16Mb DRAM

OPTOTRAN

PWR DIS

BIPOLAR

CMOSNMOS

BIPOLAR

EPI

BIPOLAR

CMOS

NA

BIP

2

1

5

5

5

0.5

12

1986

1988

1983

1986

1991

1991

1970

1985

1991

HUAKO

EI,i;CTRONlCS

HUAYUE

MICRCJELECTRONICS

CO, LTD,

HUAYLit!

MICROELECTRONICS

CO, LTD,

HUAJING

ELECTRONICS GROUP

T A I P O

SHAOXfNC

SHAOXING

WUXI

HONG

KONG

CHINA

CHINA

CHINA

NA

FABl

FAB 2

FABl

MPR, MCU,

TliLECOM

MPU LIN ASIC LOG

SRAM ROM

CONSUMER JCs

ASIC

CMOS

CMOS MOS

CMOS

CONSUMER ICs

ASiC

CONSUMER ICs,

AUDIO VISUAL

CMOS

NA

2,5

3

1.5

1982

1993

1994

1994

Crt

CO

en o

CD

3

cr

CD

-^

t\3 tn

CO

CO

t n

@

C D

CO

cn

CJ

K. a>

X3

o

Table 1 (Continued)

Asia/Pacific Existing Pilot and Production Fab Lines (Including Fabs Beginning Operation dur

Company

IflJALON

MICROELECTRONICS

HUALON

MICROELECTRONICS

HYUNDAI

HYUNDAI

HYUNDAI

City or District

SCIENCE PARK

SCIENCE PARK

ICHIN, KYUNGKl'DO

ICHIN, KYUNGKl-DO

ICHUN, KYUNGKI-DO

Country

TAIWAN

TAIWAN

KOREA

KOREA

KOREA

Fab Name Products Produced

FABl CONSUMER,

TELECOM ICs

FAB IB ROM SRAM

Process

Technology

CMOS

CMOS

Est.

Minimum

Geometry

(Microns)

0.8

Wafer Year

Diameter of Initial

(In.) Production

1988

0.6

0.5

0.35

I

1994

1993

1994

1985

HYUNDAI

HYUNDAI

HYUNDAI

HYUNDAI

HYUNDAI

JINAN NO.l

JINAN N0.2

KODENSHI

KOREAN

ELECTRONIC CO-

KOREAN

ELECTTRONIC CO.

KUKJE

LG SEMICON

LG SEMICON

LG SEMICON

LG SEMICON

LG SEMICON

LG SEMICON

LG SEMICON

MACRONIX, INC

ICHUN, KYUNGKI-DO

ICHUN, KYUNGKl-DO

ICHUN, KYUNGKI-DO

ICHUN, KYUNGKI-DO

ICHUN, KYUNGKI-DO

JINAN

JINAN

IRI

GUMI-CITY, KYUNGBUK

KOREA

GUMI-CITY, KYUNGBUK KOREA

GUMI-CrrY, KYUNGBUK KOREA

SCIKNCE PARK

KOREA

KOREA

KOREA

KOREA

KOREA

CHINA

CHINA

KOREA

GUMI-CITY, KYUNGBUK

KOREA

SHIHUNG KYUNGKI-DO cHONGfu-crrv,

CHOONGBUK

CHONGJU-CnX

CHOONGBUK

CHONGJU-CITY,

CHCXJNGBUK

KOREA

KOREA

KOREA

KOREA

CHONGJU-Ci'IX

CHOONGBUK

KOREA

GUMI-CITY, KYUNGBUK KOREA

TAIWAN

FAB 4

FABS

MOS FAB

1-A

MOS FAB

1-B

MOS FAB

2-A

MOS FAB

2-B

MOS FAB 3

MOS R&D

NA

NA

NA

4Mb 16Mb DRAM

16Mb 64Mb DRAM

256K DRAM SRAM

64K 256K 1Mb SRAM

1Mb 4Mh DRAM

1Mb DRAM

4Mb DRAM

DRAM

LOG OP AMP

IK SRAM 4K DRAtit

OPTOBIP

BIP LINE 1

ANALOG TRANS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

NA

MOS

COMPOUND

BIP

BIPOLAR

BIP LINE 2

ANALOG !Cs ASIC

BIPOLAR

NA

C2, PHASE

2

CI, PHASE

1

CI, PHASE

2

C2, PHASE

1

G-FAB

GUMI

BIPOLAR

GUMI

MOS

FAB 1-A

OITOELECTRONICS

16Mb DRAM

GaAs

CMOS

1Mb 4Mb DRAM

4Mb DRAM

16Mb DRAM

SRAM ROM

MEMORY

ANALOG LOGIC

SRAM ROM LOGIC

FOUNDRY

CMOS

CMOS

CMOS

CMOS

BIPOLAR

CMOS NMOS

CMOS

0.8

0.7

0.6

0.5

0.25

5

5

1.5

1.2

5

0.5

0.8

0.7

0.5

0.5

3

1.5

0.45

1985

1986

1992

1989

1989

1985

1989

1995

1975

1985

1992

1995

1989

1991

1993

1995

1980

1984

1994

• a

o

I ci>

o .

o

CD

O

CB

3

CT

CD

IS3

cn

CD

CO

t n

@

CO

CO en

O

.o

CO

CO

en

Table 1 (Continued)

Asia/Pacific Existing Pilot and Production Fab Lines (Including Fabs Beginning Operation duri

Company

MACRONIX, INC,

City or DistricI

SCIENCE PARK

SCIENCE PARK

Country

TAIWAN

TAIWAN

Fab Name Products Produced

FABl ROM EFROM 1Mb

4Mb FLASH DRAM

FAB 1-A DRAM VRAM

Process

Technology

CMOS

CMOS MOS

Est.

Minimum

Geometry

(Microns)

0.8

Wafer Year

Diameter of Initial

(In.) Production

1992

05

1995 MOSiiL-ViTELIC

CORPORATION

MOTOROLA

MOTOROLA liLECTRONICS

PHOTRONiCS

RAMAX

RCL

SEMICONDUCTORS

RCL

SRPvUCONDUCTORS

RECTRON LTD.

SAM Mi

SAMSUNC^

SAMSUNG

SAMSUNG

SAMSUNG

SAMSUNG

SAMSUNG

SAMSUNG

SAMSUNG

SAMSUNG

SGNEC SI-IOUGANG

SGS-THOMSON

SGS-THOMSON

SCS-THOMSON

TIANJIN

SEREMBAN

NA

MELBOURNE

TAIPO

TAIPO

TAIPEI

YONGIN KYUNGKI-DO

BUCHON-CTTY,

KYUNGKT-nO

BUCHON-CITY,

KYUNGKl-DO

KIHEUNG-UP,

KYUNGKI-DO

KIHEUNG-UP,

KYUNGKI-DO

KIHEUNG-UP,

KYUNGKI-DO

KIHEUNG-UP,

KYUNGKI-DO

KIHEUNG-UP,

KYUNGKI-DO

KIHEUNG-UP,

KYUNGKI-DO

KIHEUNG-UP,

KYUNGKI-DO

BEIJING

ANG MO KIO

ANG MO KIO

ANG MO KIO

CHINA

MALAYSIA

TAIWAN

AUSTRALIA

HONG

KONG

HONG

KONG

TAIWAN

KOREA

KOREA

KOREA

KOREA

KOREA

KOREA

KOREA

KOREA

KOREA

KOREA

CHINA

SINGAPORE

SINGAPORE

SINGAPORE

MOS-17

ISMF

NA

NA

NA

NA

TELECOM ASIC RF

PWRTRANDISSK'

SMALL SIGNAL

OPTO

FERRAM

MEM MPU LOG LIN

TRAN

NA

CMOS

BIPOLAR

NA

CMOS GaAs

CMOS

CMOS

NO. 1

m-v

BIPOLAR

LINE

DIS

LED LD HEMT

ANALOG

NA

GaAs

BIPOLAR

MOS LINE

MPU MCU MPR

LOGIC

MOSIC

CMOS NMOS

CMOS

F A B l

SRAM MASK ROM

CMOS

FAB 2

FAB 3

1Mb DRAM VRAM

SRAM

CMOS

CMOS

FAB 4 4Mb DRAM SRAM

CMOS

FABS

4Mb 16Mb DRAM

CMOS

FAB 6

16Mb 54Mb DRAM

CMOS

R&D

R&D

NA MCU LOGIC

BIPOLAR

COMPLEX

CONSUMER ICs,

PWR AMI'S

ANALOG

BIPOLAR

STANDARD

MOS

ANALOG OPTO

DISCRETE

MPR ASIC ANALOG

CMOS

BIPOLAR

BIPOLAR

CMOS MOS

0.5

3

2.5

2.5

5

3

2

1.5

1.2

0.8

0.6

0.5

0.35

0.3

1.2

3

8

2

1988

1990

1982

1993

1983

1992

1978

1974

1984

1985

1988

1990

1992

1994

1989

1994

1984

1984

1988

©

en

C3

Si lu

X l t=

CO

tn o o

• 3 o

I

a.

CO cu

CD

3

cr

CD fV)

CJ1

CD

CO

en

CO

I

B

Table 1 (Continued)

Asia/Pacific Existing Pilot and Production Fab Lines (Including Fabs Beginning Operation dur

Company

SHANGHAI B131XING

MICROK[.i;crRONICg

SHANGHAI RADIO

N0.7

SUZHOU PLANT

City or District

SHANGHAI

SHANGHAI

SUZHOU

TECH

SEMICONDUCTOR

TRMIC

Tl/ACEK

Tl/ACRR

TSMC

TSMC

TSMC

TISMC

TSMC

UMC

UMC

UMC

VANGUARD

INTERNATIONAL

WIN BOND

WIN BOND

SINGAPORE

SHANGHAI

SCIENCE PARK

SCIENCE PARK

SCIENCE HARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

Country

CHINA

CHINA

CHINA

SINGAPORE

CHINA

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

Fab Name Products Produced

NA TELECOM

CONSUMER ICs DIS

FABl

NA

NA

Process

Technology

CMOS

BIPOLAR

CMOS

BIP TTL MOS

PHASE 1

NA

FABl

FAB 1-B

FABl

FAB 2-A

FAB 2-B

FAB2<:

FAB 3

FABl

FAB 2

FAB 3-A

FABIA

LOG O n t )

CONSUMER

4Mb 16Mb 64Mb

DRAM

HYBRID

4Mb DRAM

4Mb/16Mb DRAM

MEMORY MICRO

LOGIC ANALOG

LOGIC

SRAM

FOUNDRY

DRAM SRAM ROM

LOG CUSTOM

CONSUMER IC

LOGIC TFT-LCDs

Ml'R MI'U SRAM

ROM I/O

SRAM ROM IXX;iC

MPU MPR

4Mb 16Mb DRAM

CMOS

NA

CMOS MOS

CMOS

CMOS

BiCMOS

CMOS

CMOS

NA

CMOS

BiCMOS

CMOSNMOS

CMOS

CMOS

CMOS

TAIWAN

TAIWAN

FABl

FAB 2

TELECOM ICs,

DIALERS

SRAM RISC LAN

CMOS

CMOS

BiCMOS

Est.

Minimum

Geometry

(Microns)

1.2

Wafer

Diameter

(In.)

4

Year of Initial

Production

1989

3

0.5

0.6

0.5

0.8

0.6

0.5

0.6

0.35

0.8

0.5

0.5

0.4

0.8

0.6

5

3

8

6

6

6

8

6

8

6

4

6

a

8

5

6

1991

1990

1993

1991

1995

1988

1990

1992

1994

1995

1982

1989

1995

1995

1988

1992

NA = Not applicable

Fab Types:

F = Product!on-Based Fab

B = Semiconductor R&D ancfor Trial Production Facility

P = Pilot Line (Initial Production or Intended Low Volume)

T = Test and Assembly (Formerly A)

Q = Qtiick-Turn Fab

N = Nondedicated Foundry Service Available

D = Design Center

Source: Dataquest (December 1995)

1

<p

CO t n

C3 en

C/3

@ t o

CD

CJl

£i

(U

X 3

O)

en p - * -

5"

CD

C3. tn

CD

O

CD

cr

CD

Table 2

Asia/Pacific Future Pilot and Production Fab Lines (Including Fabs Beginning Operation durin

Company

ASMC (Formtrly

Shanghai Radio No.7)

BE!sm.[:YIIALL&CO,

(Finsnciers)

CHARTERED

SENDCONDUCTOR

DONGSUNG

I-IOLTEK

HUAJtNG

ELECTROMiCS GROUP

City or District

SHANGHAI

SARAWAK

StNGAPORE

IC5-IUN, K Y U N G K I - D O

SCIENCE PARK

WUXI

HUAl-ON

MICROELECTRGNICS iiYUNDAi

KODENSHI

I.G SEMICON

I.GSEMICON

LG SEMICON

MACRONIX, IhfC.

MOSEL-VFTELIC

CORPORATION

MOSEL-VTTliLlC

CORPORATION

MOSEL-VITELIC

CORTORATION

NANYA

TECHNOLOGY

TOWERCUIT

SEMICONDUCTOR

(ELiTIi GROUP)

SAMSUNG

SAMSUNG

SOENCEPARK

I C m N , K Y U N G K I ^

IRI

CHONG]u-cmr,

CHOONGBUK

CHONGju<:rrY,

CHOONGBUK

GUMI-Cn*Y, KYUNGBHK

KOREA

KOREA

KOREA

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

TAO YUAN

SCIENCE PARK

KIHEUNG-UP,

KYUNGKI-DO

KIHEUNG-UP,

KYUNGKI-DO

BANGKOK

Country

CHINA

MALAYSL^

KOREA

TAIWAN

CHINA

TAIWAN

KOREA

KOREA

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

KOREA

KOREA

THAILAND

Fab Name

FABl

NA

SINGAPORE FAB II

Products Produced

LIN DIGITAL IC for

TV RPSOM

NA

LOGIC SRAM

Process

Technology

CMOS

NA

CMOS

BIPOLAR

FAB 2

FAB 2

DIODE, R E C n n E R

MPRMCTJ

TELECOM

CONSUMER AUTO

ICs ASIC DIS

SRAM

BIPOLAR

CMOS

CMOS

FAB 2

FAB 6

NA

6^«v!bDRAM

OPTOBIP

CMOS

BICMOS

CMOS

COMPOUND

BIP

CMOS C2, PHASE

2

CI, PHASE

3

G-FAB

16Mb DRAM

16Mb 64Mb DRAM

FAB 2

FAB 1-A

CMOS

SRAM ROM

MEMORY

ROM EFROM LOGIC

CMOS

CMOS

DRAM VRAM CMOSMOS

FAB 1-B

FAB 2

FABl

NA

FABS

FAB 7

FABl

4Mb 16Mb MMb

DRAM, SRAM

16Mb DRAM SRAM

16Mb 64Mb DRAM

16MDRAM

64Mb DRAM

15Mb 64Mb DRAM

1-OUNDRY

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

Est.

Minimum

Geometry

(Microns)

0.8

Wafer

Diameter

(In.)

6

Year of Initial

Production

1995

0,35

0.5

5

0.45

0.8

0.5

0.35

0.5

0.35

0.5

0.35

0.5

0,35

0.35

0.45

0.4

0.3

0.35

0.5

8

8

4

6

6

8

8

8

8

8

8

6

6

8

8

8

8

8

8

1997

1995

1996

1997

1997

1996

1996

1995

1995

1997

1995

1997

1995

1996

1998

1996

1996

1997

1996

1996

SUBMICRON

TECHNOLOGY

TECH

SEMlCONDUCrOK

SINGAPORE

SINGAPORE PHASE 2

64Mb 256Mb DRAM CMOS

0.35 8 1997

CO

CO

tn

o

CD

O

CD

3

t n

CO

CD

en

CO

CO

O l

@

CD

CO

t n

O

on

I

Table 2 (Continued)

Asia/Pacific Future Pilot and Production Fab Lines (Including Fabs Beginning Operation durin

Company

TI/ACER

TSMC

TSMC

UMC

UMC

UMC-JVl

UMC-IV2

UMC-JV3

VANGUARD

INTERNATIONAL

VANGUARD

INTERNATIONAL

WIN BOND

City or District

SCIENCE PAJRK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

SCIENCE PARK

Country

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

TAIWAN

Fab Name Products Produced

FAB 1-B 4Mb/16Mb DRAM

FAB 3 DRAM SRAM ROM

\JOG CUSTOM

FAB 4

LOG CUSTC)M MFU

MRM

FAB 3-A

FAB 3-B

FABIDB

SRAM ROM LOGIC

MI'UMPR

SRAM ROM LOGIC

MPUMPR

SRAMMPR

GRAPHIC a i E P S

FOUNDRY

NA

NA

FABIA

FOUNDRY

FOUNDRY

4Mb 16Mb DRAM

Est.

Minimum Wafer Year

Process Geometry Diameter of Initial

Technology (Microns) (In.) Production

CMOS

CMOS

BiCMOS

0.5

0.35

8

8

1995

1995

0.35

8 1996

CMOS

BiCMOS

CMOS 0.5

8 1995

CMOS

CMOS

CMOS

CMOS

CMOS

0.35

0.35

0.35

0.35

0.4

8

8

8

8

8

1996

1997

1997

1997

1995

FAB IB

FABS

16Mb DRAM CMOS

LOGIC SRAM NGN

VOLATILE MEMORY

CMOS

0.35

0.35

8

8

1997

1997

NA = Not applicable

Fab Types;

F = Production-Based Fab

R = Semiconductor R&D and/or Trial Production Facility

P = Pilot Line {Initial Production or Intended Low Volume)

T = Test and Assembly (Formerly A)

Q = Quick-Tum Fab

N = Nondedicated Foundry Service Available

D = Design Center

Source: Dataquest (December 1995)

For More Information...

James Seay, Research Analyst (408) 468-8259

Internet address [email protected]

Via fax (408)954-1780

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our clients Individual companies reported on and analyzed by Dataquest may be clients of this and/or other Dataquest services. This information is not furnished in connection with a sale or offer to sell securities or in cormection with the solicitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their families may, from

J J 3 l ! 3 Q U C S l time to time, have a long or short position in the securities mentioned and may sell or buy such securities.

©1995 Dataquest Incorporated—Reproduction Prohibited

BB ThcDun& Bradstreet CorDOiahon Dataquest is a registered trademark of A.C. Nielsen Company

DATAQUEST WORLDWIDE OFFICES

NORTH AMERICA

Worldwide Headquarters

Dataquest Incorporated

251 River Oaks Parkway

San Jose, California 95134-1913

United States

Phone: 1-408-468-8000

Facsimile: 1-408-954-1780

Dataquest Incorporated

Nine Technology Drive

RO. Box 5093

Westborough, Massachusetts 01581-5093

United States

Phone: 1-508-871-5555

Facsimile: 1-508-871-6262

Dataquest Global Events

3990 Westerly Place, Suite 100

Newport Beach, California 92660

United States

Phone: 1-714-476-9117

Facsimile: 1-714-476-9969

Sales Offices:

Washington, DC (Federal)

New York, NY (Financial)

Dallas, TX

LATIN AMERICA

Research Affiliates and Sales Offices:

Buenos Aires, Argentina

Sao Paulo, Brazil

Santiago, Chile

Mexico City, Mexico

EUROPE

European Headquarters

Dataquest Europe Limited

Holmers Farm Way

High Wycombe, Bucks HP12 4XH

United Kingdom

Phone: +44 1494 422 722

Facsimile: +44 1494 422 742

Dataquest Europe SA

Immeuble Defense Berg^res

345, avenue Georges CMmenceau

TSA 40002

92882 - Nanterre CTC Cedex 9

France

Phone: +33 1 41 35 13 00

Facsimile: +33 1 41 35 13 13

Dataquest GmbH

Kronstadter Strasse 9

81677 Miinchen

Germany

Phone: +49 89 93 09 09 0

Facsimile: +49 89 93 03 27 7

Sales Offices:

Brussels, Belgium

Kfar Saba, Israel

Milan, Italy

Randburg, South Africa

Madrid, Spain

JAPAN

Japan Headquarters

Dataquest Japan K.K.

Shinkawa Sanko Building

6th Floor

1-3-17, Shinkawa

Chuo-ku, Tokyo 104

Japan

Phone: 81-3-5566-0411

Facsimile: 81-3-5566-0425

ASIA/PACIFIC

Asia/Pacific Headquarters

7/F China Underwriters Centre

88 Gloucester Road

Wan Chai

Hong Kong

Phone: 852-2824-6168

Facsimile: 852-2824-6138

Dataquest Korea

Suite 2407, Trade Tower

159 Samsung-dong, Kangnam-gu

Seoul 135-729

Korea

Phone: 822-551-1331

Facsimile: 822-551-1330

Dataquest Taiwan

llF-2, No. 188, Section 5

Nan King East Road

Taipei

Taiwan, R.O.C.

Phone: 8862-756-0389

Facsimile: 8862-756-2663

Dataquest Singapore

105 Cecil Street

#06-01/02 The Octagon

Singapore 0106

Phone: 65-227-1213

Facsimile: 65-227-4607

Dataquest Thailand

12/F, Varussa Building

29 Soi Chidlom

Ploenchit Road

Patumwan, Bangkok 10330

Thailand

Phone: 662-655-0577

Facsimile: 662-655-0576

Research Affiliates and Sales Offices:

Melbourne, Australia

Beijing, China

Dataquest*

a company of

The Dun & Bradstrect Corporation

®1995 Dataquest Incorporated

DataQuest

Europe, Africa, and the

IVIiddle East Fab Database

Market Statistics

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-9505

Publication Date: December 25,1995

Filing: Market Analysis

Europe, Africa, and the

IVIiddle East Fab Database

Market Statistics

Program: Semiconductor Equipment, IVIanufacturing, and IViaterials Worldwide

Product Code: SEMM-WW-MS-9505

Publication Date: December 25,1995

Filing: Market Analysis

Europe, Africa, and the IVIiddle East Fab Database

Table of Contents

Page

Ettrope, Africa, and the Middle East Fab Database

Background

Research Methodology

General Definitions

Worldwide Geographic Region Definitions and Regional

Roll-Ups

Americas

Japan

Europe, Africa and Middle East

Asia/Pacific 2

Definition of Table Columns 2

SEMM-WW-l\/lS-9505 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

List of Tables

Table Page

1 Europe, Africa, and Middle East Existing Pilot and Production

Fab Lines.

2 Europe, Africa, and Middle East Future Pilot and Production

Fab Lines 13

Note: All tables show estimated data.

SEMM-WW-MS-9505 ©1995 Dataquest Incorporated December 25,1995

Europe, Africa, and the IVIiddle East Fab Database ^^.^^^^

Background

This document contains the Europe, Africa, and Middle East portion of Dataquest's wafer fab database. The Semiconductor Equipment,

Materials, and Manufacturing Worldwide (SEMM) program uses both primary and secondary research to update this data. The tables in this report cover both merchant and captive production and pilot-line facilities, although our surveys and database also include R&D fabs.

Research Methodology

Dataquest conducts extensive annual surveys, complemented with quarterly secondary research. This data is then supplemented and crosschecked with various other information sources.

General Definitions

Fab line: A fab line is a semiconductor processing facility equipped for all front-end wafer manufacturing. Occasionally, there are two or more separate product-specific fab lines or wafer sizes in a single cleanroom.

In this situation, Dataquest documents the cleanroom as separate fab lines if the company dedicates equipment to each wafer size or product line. Therefore, a company may operate many fab lines at one location.

Front-end wafer processing: Dataquest defines front-end wafer processing as all steps involved in semiconductor processing, beginning with initial oxide and ending at wafer probe.

Production fab: A wafer fab capable of front-end processing more than

1,250 wafers per week is a production fab (type = F).

Pilot fab: A wafer fab capable of front-end processing 1,250 or fewer wafers per week is a pilot fab (type = P).

Worldwide Geographic Region Definitions and Regional Roll-Ups

Americas

Includes Central America (all nations), Canada, Mexico, United States and Puerto Rico, and South America (all nations).

Japan

Japan is the only single-coimtry region.

Europe, Africa, and Middle East

Includes Africa (all nations), Albania, Andorra, Armenia, Azerbaijan,

Belarus, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia,

Finland, France, Georgia, Germany, Gibraltar, Hungary, Iceland, Israel,

Italy, Kazakhstan, Kyrgyzstan, Latvia, Liechtenstein, Lithuarua,

Luxembourg, Malta, Middle East (all nations), Moldova, Monaco,

SEMM-WW-MS-9505 ©1995 Dataquest Incorporated

Semiconductor Equipment, IVianufacturing, and Materials Worldwide

Netherlands, Norway, Poland, Romania, Russia, San Marino,

Scandinavia, Slovakia, Spain, Sweden, Sweden, Switzerland, Tajikistan,

Turkey, Turkmenistan, Ukraine, United Kingdom, Uzbekistan, Vatican

City, and Yugoslavia (all nations within the former Yugoslavia).

Asia/Pacific

Includes Australia, Bangladesh, Cambodia, China, Hong Kong, India,

Indonesia, Laos, Malaysia, Maldives, Myanmar, Nepal, New Zealand,

Pakistan, Philippines, Singapore, South Korea, Sri Lanka, Taiwan,

Thailand, and Vietnam.

Definition of Table Columns

Products Produced contains details for seven product categories. The nomenclature used within the seven product groups of the fab database is as follows, with defirutions where warranted:

• Analog

a A / D D/A: Analog-to-digital, digital-to-analog converter

o AUTOMOTIVE: Dedicated to automobile applications o CODEC: Coder/decoder

o INTERFACE: Interface IC

o LIN: Linear/analog device

D MDIODE: Microwave diode

a MESFET: Metal semiconductor field-effect transistor

a MEET: Microwave field-effect transistor

o MIXSIG ASIC: Mixed-signal/linear ASIC o MODEM: Modulator/demodulator

o MMIC: Monolithic microwave IC

• OP AMP: Operational amplifier

o PWR IC: Power IC o REG: Voltage regulator

o SMART PWR: Smart power

a SWITCHES: Switching device

o TELECOM: Telecommunications chip

• Memory

a DRAM: Dynamic RAM

a EEPROM or E2: Electrically erasable PROM o EPROM: Ultraviolet erasable PROM

SEMM-WW-MS-9505 ©1995 Dataquest Incorporated December 25,1995

Europe, Africa, and the Middle East Fab Database o FERRAM: Ferroelectric RAM o FIFO: First-in/first-out memory

a FLASH: Flash memory

a MEM: Memory o NVMEM: Nonvolatile memory (ROM, PROM, EPROM, EEPROM,

FERRAM)

n PROM: Programmable ROM o RAM: Random-access memory

a ROM: Read-only memory o SPMEM: Other specialty memory (such as dual-port, shift-register, color lookup)

a SRAM: Static RAM o VRAM: Video RAM

Micrologic

a ASSP: Application-specific standard product o BIT: Bit slice (subset of MPU functions) o DSP: Digital signal processor

a LISP: 32-bit list instruction set processor for AI

a MCU: Microcontroller unit o MPR: Microperipheral o MPRCOM: MPR digital communication (ISDN, LAN, UART, modem) o MPU: Microprocessor xinit n RISC: Reduced-instruction-set computation 32-bit MPU

Standard logic

a LOG or LOGIC: Standard logic

ASIC logic o ARRAYS: Gate array o ASIC: Application-specific IC

a CBIC: Cell-based IC o CUSTOM: Full-custom IC (single user)

• FPGA: Field-programmable gate array

• PLD: Programmable logic device

SEMi\/l-WW-MS-9505 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, IVIanufacturing, and l\/laterials Worldwide

• Discrete o DIODE o DIS or DISCRETE: Discrete o FET: Field-effect transistor o GTO: Gate turn-off thyristor

a HEMT: High-electron-mobility transistor

• IGBT: Insulated gate bipolar transistor

a MOSFET: MOS-based field-effect transistor o PWR TRAN: Power transistor o RECTIFIER

a RF: Radio frequency o SCR: Schottky rectifier o SENSOR

D SST: Small-signal transistor o THYRISTOR

a TRAN: Transistor o ZENER DIODE

• Optoelectronic o CCD: Charge-coupled device (imaging) o COUPLER: Photocoupler a lED: Infrared-emitting diode

a IMAGE SENSOR

a LASER: Semiconductor laser or laser IC o LED: Light-emitting diode

a OPTO: Optoelectronic o PDIODE: Photo diode o PTRAN: Photo transistor

• SAW: Surface acoustic wave device

• SIT IMAGE SENSOR: Static induction transistor image sensor

Process Technology column lists four major types of technologies. This column also lists unconunon technologies witii information on well types, logic structure, and number of raetal levels. Definitions used in the "Process Technology" column are as follows:

• MOS (silicon-based) o CMOS: Complementary metal-oxide semiconductor

SE!\/IM-WW-MS-9505 ©1995 Dataquest Incorporated December 25,1995

Europe, Africa, and the Middle East Fab Database a MOS: N-charmel metal-oxide semiconductor (NMOS) and p-channel metal-oxide semiconductor (PMOS). o Ml: Single-level metal o M2: Double-level metal

a M3: Triple-level metal o N-WELL o P-WELL o POLYl: Single-level poly silicon o POLY2: Double-level poly silicon o POLYS: Triple-level polysilicon

BiCMOS (silicon-based)

a BiCMOS: Bipolar and CMOS combined on a chip

a BiMOS: Bipolar and MOS combined on a chip o ECL I/O: ECL input/output

a TTL I/O: TTL input/output

Bipolar (silicon-based)

a BIP or BIPOLAR: Bipolar o ECL: Emitter-coupled logic o TTL: Transistor-transistor logic o STTL: Schottky TTL

Gallium arsenide and other compound semiconductor materials o GaAs: Gallium arsenide o AlGaAs: Gallium aluminum arsenide

a GaAs on Si: Gallium arsenide on silicon

a GaP: Gallium phosphide o HgCdTe: Mercuric cadmium telluride o In As: Indium arsenide

a InGaAs: Indium gallium arsenide o InP: Indium phosphide o InSb: Indium antimony

• LiNbOS: Lithium niobate o SOS: Silicon on sapphire

SEIVIM-WW-IVlS-gSOS ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, Manufacturing, and IViaterials Worldwide

Minimum Geometry is the smallest feature attainable in production volumes, measured in microns, at the critical mask layers.

Wafer Diameter represents the wafer diameter usually expressed colloquially in inches. However, for wafers greater than 3 inches in diameter, the colloquial expression becomes grossly inaccurate. When calculating square inches, Dataquest uses the following approximations:

• Stated diameter 4 inchep (100mm) = Approximate diameter

3.938 inches

• Stated diameter 5 inches (125nvm.) - Approximate diameter

4.922 inches

• Stated diameter 6 inches (150mm) = Approximate diameter

5.906 inches

• Stated diameter 8 inches (200mm) = Approximate diameter

7.87 inches

Estimated Maximum Wafer Starts per Month is the equipment-limited wafer start capacity per four-week period. Start capacity is limited only by the installed equipment in the fab and the complexity of the process it runs, not by current staffing or the number of shifts operating.

SEIVllVi-WW-IVIS-gsOS ©1995 Dataquest incorporated December 25,1995

D ct>

3

CO

CO

en

CO

t n

o

t n

CD

CO

1—^

~ i

O

O

• 3

O c?

CI.

@

<D

CO o

K.

03

t—>

Table 1

Europe, Africa, and Middle East Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

ABB SEMICONDUCTOR

ABB-! iAFO AB

City or District

LENSBURG

JARFALLA

AEG AG (DAIMLER

BENZ)

AMS AUSTRtA MIKRO

SYSTEMR

INTERNATIONAL

ANALOG DEVICES

ANALOG DEVICES

ANSALEX* TRASPOtrn

AT&T

MICROE-LECTRONICS

ATMELCORPORATION

(ES2)

ATMOS/ELPOL

BANEASA S.A. (IPRS)

BT&D TECHNOLOGIES

ELMOS GMBH

EM

MICROELECTRONICS-

MARIN S,A.

ERICSSON

ERICSSON

FUJITSU

I'UJiTSU

GECPLES3EY

GEC PLESSEY

ULM

LIMERICK

LIMERICK

GENOA

MADRID

ROUSSET CEDEX

WARSAW

BUCHAREST

IPSWICH

DORTMUND

MARIN

KALMAR

KISTA

NEWTON AYCLIFFE

NEWTON AYCLIFFE

LINCOLN

OLDHAM

Country

SWITZERLAND

SWEDEN

Fab Name

NA

NA

GERMANY

UNTERPREMSTATTEN AUSTRL\

IRELAND

IRELAND

ITALY

SPAIN

FRANCE

POLAND

ROMANIA

ENGLAND

GERMANY

SWITZERLAND Fab 3

SWEDEN

SWEDEN

ENGLAND

ENGLAND

ENGLAND

ENGLAND

ULMRSCH

NA

NA

NA

LINITA

NA

ES2

NA

NA

NA

NA

NA

NA

PHASE 1

PHASE 2

NA

NA

Products

Produced

DIS

ASIC MIXED

SIGNAL

CUSTOM

3DICS mm-WAVE OPTO

ASIC ANALOG

Process Tech-

no\ogy

BIPOL/VR

CMOS

GaAs MOS

CMOS BiCMOS

NMOS PMOS

Est.

Minimum

Geometry

(Microns)

1.25

Wafer

Diameter

(In.)

4

4

Year of Initial

Production

1991

4

1991

1983

ANALOG LIN

AD/DA

TELECOM

DSPs ANALOG

PWRDIS

CBIC CUSTOM

DSP

CBIC ARRAYS

CUSTOM

MIXSIGASiC

ASIC

THYRISTOR

DIODE LIN

OPTO LASER

LED

ASIC HV MIXED

SIGNAL

ASIC ANALCKj

BICMOS

CMOS BiCMOS

BIPIM

CMOS

CMOSM2

POLYl

NA

BIP

NA

CMOS

CMOS BiCMOS

PWR DISCRETE

AS)CS(TELECOM

CHIPS)

4Mb/16Mb

DRAM ASIC

16Mb DRAM

BIP

CMOS BICMOS

CMOS

SRAM MPU ASIC

DISCRETE

MPR ASIC

DISCRCTE

CMOS

CMOS

BIPOLAR

2

0.6

2

0,6

0.6

2

1

1

1

0.5

0.8

0.5

1

1

4

6

4

2

4

6

6

6

4

6

4

6

4

6

4

1986

1995

1970

1988

1987

1987

1985

1991

1972

1994

IWl

1994

1981

1964

CO

CO

CJ1

@

<o

CD

cn

O

{u

.a

CD

O

CD

3

ex

CO

—1 r o cn

CO

CO

cn

CO

1

CO

cn

t n

s

I

Table 1 (Continued)

Europe, Africa, and Middle East Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

GEC PLESSEY

GliC PLESSRY

GEC PLESSEY

GiiC I'LE&SEY

GENiJRAL

[NSTRu^IE^^$

HITACHI

IIMT riUGHES l!)M

IBM

IBM

IBM

IBM

IBM

IBM

IBM

!BM

! B M / P H I U ^

!BM/SIEMQ4^

ICCE

IMKC

INSTITUTE OF

ELECTRON

TECHNOLOGY

INTEGRAL

INTEL

City or District

PLYMOUTH

ROBOROUGH

SWINDON

WEMBLEY

CRICKLADE

Country

ENGLAND

ENGLAND

ENGLAND

ENGLAND

ENGLAND

Fab Name

NA

NA

NA

NA

NA

Products

Produced

Process Technology

MPU DSPs ASIC CMOS

ANALOG

ASIC DSP

TELECOM

ANALOG

NA

DIS

CMOS NMOS

M3

BIPOLAR

GaAs

BIP

I.ANDSHLIT

BRUGG

GLENROTHES

GERMANY

SWITZERLAND

SCOTLAND

E2

NA

NA

16Mb DRAM

MCU

CONSUMER !Cs

EPKOM ASIC

MIXED SIGNAL

CORBEIL-ESSONNES

CORBEIL-ESSONNES

FRANCE

FRANCE NA

MMbDRAM

256K DRAM 64K

SRAM

CORBEIL-ESSONNES FRANCE

CORBEIL-ESSONNES

HANNOVER

SINDELTINGEN

SINDELFINGEN

SINDELFINGEN

ZURICH

DOEBtJNGEN

CORBE rL-F,SSONNES

BANEASA

PLIJM

NA ARRAYS LIN

CUSTOM

FRANCE NA

1Mb DRAM

GERMANY

GERMANY

GERMANY

NA

NA

NA

DIB

ARRAYS

PWR DIS

HYBRID

GERMANY NA

DRAM SRAM

DSP MPU

CUSTOM

NA

SWITZERLAND NA

GERMANY NA

4Mb DRAM

FRANCE ACL

16Mb DRAM

ROMANIA NA

BELGIUM

OITO LIN

R&D

PRcnr>rfPiNG

WARSAW POLAND R&D SIU-

CON

ASIC MCU DSP

OPTO

MINSK CfTY

LEIXUP

RUSSIA

IRELAND

NA

FAB 10

NA

MPU (P5, DX4,

P6)

CMOS

MOS

CMOSMOS

CMOS MOS

B!P

CMOS

BIT

BIP

BIP

MOS

GaAs C M q ^

CMOS

CMOS

BU'

CMOS BiCMOS

NMOS/PMOS

CMOS

BIP CMOS

BiCMOS

Est.

Minimum

Geometry

(Microns)

0.7

Wafer

Diameter

(In.)

Year of Initial

Production

1987

0.7

0.5

0.5

0.35

1

0.8

3

2

3

1.5

0.8

0.4

0.25

3

0,6

1987

1961

1992

1970

1989

1989

1992

1982

1978

1994

O

CD

0

o -

CD

CO

CO

@

CO

CO

en

O o>

s-

cz

. 0

l-t-

5"

0

0

• 3

0

1

CD

Q .

Table 1 (Continued)

Europe, Africa, and Middle East Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

INTEL

City or District

MIGDAEL HEIMIK

Country

ISRAEL

Fab Name

FABS

Process Technology

CMOS

Est.

Minimum

Geometry

(Microns)

0.8

Wafer

Diameter

(In.)

6

Year of Initial

Production

1985

INTERNATIONAL

RECTIFIER

INTERNATIONAL

RECrrfflER

ISKRA

ISOCOM

ITALTEL m

ITT

IXYS SEMICONDUCTOR

LUCAS

MICROELECTRONICA

S.A.

MICROELECTRONICS -

IME, LTD.

MICROELECTRONICS -

IME, LTD.

MICRONAS, INC,

TURIN

TURIN

TRBOVLJE

HARTLEPOQIi

ROME

FREIBUKG

FREIBURG

L A M P B R T F ^ ^

SUTTON COLDFIELD

BANEASA

SOFIA

SOFIA

ESPOO

ITALY

ITALY

SLOVENL\

ENGLAND

ITALY

GERMANY

GERMANY

GERMANY

ENGLAND

ROMANIA

BULGARIA

BULGARIA

FINLAND

BORGARO

VENARIA

NA

NA

NA

ICFAB

NA

IXYS

NA

NA

NA

NA

NA

Products

Produced

MPU (386E,486)

MCUMPR

RECTIHER

TFIYRISTOR

RECTIFIER

THYRISTOR

DIS

OPTO

NA

MPU MCU DSf

NVMEM

DIS CUSTOM

DISCRETE

DIODE

THYRiSTORS

PWR DIS

MPU 16K DRAM

LIN

LIN

NA

NA

BIP

GaAs

GaAs

CMOS NMOS/

PMOS

BIPOLAR

BIPOLAR

BIPOLAR

GaAs

MOS

CMOS BICMOS

CMOS BICMOS

0,?

5

4

2

2

2

4

i

3

5

4

5

1

S

4

1960

1988

1990

1970

1986

MICRONAS, INC,

<ASCOM MICRO)

MIETEC ALCATEL

MIETEC MJCAJV.l.

MIKROELEKTRONIK &

TECHNOLOCIE GE3.

MOTOROLA

BEVAIX

OUDENAARDE

OUDENAARDE

DRESDEN

TOULOUSE

SWITZERLAND AM

BELGIUM

BELGIUM

GERMANY

FRANCE

F A B l

FAB 2

SMD

U N C B I C cusroM

ARRAYS

CUSTOM

CUSTOM MIXED

SIGNAL ASIC

CUSTOM MIXED

SIGNAL ASIC

ASIC SRAM DSP

CMOSM2

BIP

CMOS BiCMOS

CMOS BiCMOS

CMOSM2

BIP

5

1

0.35

2

4

4

6

4

1990

1985

1993

1983

MOTOROLA

MOTOROLA

TOULOUSE

TOULOUSE

FRANCE

FRANCE

BIPOLAR 4

SMOSl

TLS POWER

TELECOM OP

ANALOG

PWRTRAN

SMARTMOS

LOW

FREQUENCY

POWER

CMOS

BIP MOS

3

1.5

6

5

1969

i

CJ1

o

CO

O

CD

3

cr en

CO

CO

CJI en

I

CO

O l

o

CJ1

@

C D

CO t n a

Si

(U

X3

O

—^

T3

O a

Table 1 (Continued)

Europe, Africa, and Middle East Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

MOTOROLA

MOTOROLA

MOTOROi.A

City or District

TOULOUSE

EAST KILBRIDE

EAST Kn.BRIDE

Country

FRANCE

SCOTLAND

SCOTLAND

Fab Name

TLS RECTI-

FIER

M O S l

MOS9

Est.

Minimiun Wafer Year

Products Process Tech- Geometry Diameter of Initial

Produced nology (Microns) (In.) Production

RECTIFIERS

MOS

CMOSMOS

CMOS

0.8

0.35

1969

1990

MOTOROLA (Formerly

ISigital Equip.)

MTG(THIiSVSCMBH)

NATIONAL S/C

NATIONAL S/C

NAllONAL S/C

NATIONAL S/C

NEC S E M I C O N D U C T ^

SOUTH QUEENS-

FERRY

ERFUKT

GREENOCK

GREENOCK

GREENOCK

GREENOCK

LIVINGSTON

SCOTLAND

GERMANY

SCOTLAND

SCOTLAND

SCOTLAND

SCOTLAND

SCOTLAND

ENGLAND

MOS16

NA

4

FAB 2

LINEAR 4

LINEAR 6

1 Phase 1

NA

MCU LOGIC

FSRAM DSP

MCU

MPU A f J H A

ANALOG

ASIC

LOG LIN

LAN

LIN

LIN

DRAM SRAM

Ml'U

LINDIS

CMOS BICMOS

CMOS BICMOS

CMOS MI

BIP

BIP Ml M2

BIP

CMOS

BIP

0.5

1.5

2.5

2

5

0.5

1990

1981

1984

1977

1993

1987

NEWMARKin"

MICROSYS.

NEWPORT WAFER PAB

NUOVA MISTRAL S.P.A.

PHILIPS

PHILIPS

PHILIPS

PHILIPS

PHILIPS

PHILIPS

PHILIPS

NEWMARKET

NEWPORT

SERMONETA (LA-

TCMA)

HAZELGROVE,

STOCKPORT

HAZELGROVE,

STOCKPOHT

REDinLL

CAEN

HAMBURG

HAMBURG

HAMBURG

PHILIPS

PHILIPS

PHILIPS

PHILU^

NIJMEGEN

NIJMEGEN

NUMEGEN

NIJMEGEN

WALES

ITALY

ENGLAND

ENGLAND

ENGLAND

FRANCE

GERMANY

GERMANY

GERMANY

NETHER-

LANDS

NETHER-

LANDS

NETHER-

LANDS

NETHER-

LANDS

F2I

NA

BIPOLAR

POWER-

MOS

NA

NA

CONSUMER

DISCRETE

NA

MOS3

NA

FOUNDRY

ZENER DIODE

DIODES SST

TRAN n r o D i ;

RE(7riFiER

DIODE PWR ICs

NA

CONSUMER ICs

CONSUMER ICs

DIS

8- 16-Bit MCU

EEPROM ASIC

CONSUMER ICs

DIS

LOG

CMOS

NA

BIP

MOSMl

GaAs

BIP

BIP

BIP

CMOS NMOS

M1M2

CMOS

CMOS

NA

NA

SRAM

CONSUMER ICs

DIS

CMOS NMOS

M2

MOS BICMOS

BIP

0.7

3

10

3

1.5

1.2

2

1

0.7

3

).8

1.5

1982

1983

1969

1987

1990

1957

1980

1993

1984

1988

1970

o

3

O "

CD

L

CD

CD

cn

Crt

CO

t n

CD

CJ1

@

CO

CO

CJl o

^ j a

CD

CO

i - l -

3

C3

O

Table 1 (Continued)

Europe, Africa, and Middle East Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

PHILIPS

City or District

STADSKANAAL

SILFORD

Country

NETHER-

LANDS

ENGLAND

Fab Name

NA

NA

Products

Produced

RECTIFIER

NA

Process Technology

BIPM3

GaAs

Est

Minimum Wafer Year

Geometry Diameter of Initial

(Microns) (In.) Production

1970

PHILIPS KESEARCH

LABS

PHILIPS RTC

ROBEKT BOSCH

PARIS

REUTLINGHSI

FRANCE

GERMANY

NA

NA

5

0.6

1970

1995

RODEBT SOSCH

SEAGATE MICROELECT.

SEMEFAB

LIVINGSTON

GLENROTHES

SEMIKRON

SGS-THOMSON

SGS-THOMSON

SG&-THOMSON

SGS-THOMSON

SGS-THOMSON

SGS-TTIOMSON

SGS-THOMSON

SGS-THOMSON

SGS-THOMSON

SGS-THOMSON

SGS-THOMSON

SIEMENS

SIEMENS

SIEMENS

RI:UTI.INGEN

NURNBERC

CORaEVILLE

CROLLES

GRENOBLE

RENNES

ROUSSET

TOURS

TOURS

AGRATE

AGRATE (MILAH)

CASTALETTO

CATANIA

V1LLACH

DBtESDEN

R E G E N S B l ^ ; :

GERMANY

SCOTLAND

SCOTLAND

GERMANY

FRANCE

FRANCE

FRANCE

FRANCE

FRANCE

FRANCE

FRANCE

ITALY

ITALY

ITALY

ITALY

AUSTRIA

GERMANY

GERMANY

RtW/FAW

NA

NA

NA

NA

GNB92

NA

NA

MODULE 5

MESA

PLANAR

Rl

FABS

NA

NA

VILLACH

NA

REGENS-

BURG

TRAN

LE^DIS

CUSTOM

LINDIS

CUSTOM

LIN

LIN DIS OPTO

GaAs

BiCMOS

HIP BiCMO^

DIS

NA

AS[C ASSP MPU

TELECOM DSP

LIN SMART PWR

CUSTOM

LLNEAR

TELECOM

EPROM EEPROM

MFUMCU

TELECOM

D[S

DIS

FLASH EPROM

EPROM ARRAYS

MIXSICASIC

LIN

LOG LIN

CUSTOM SMART

PWR

LOG TELECOM leMb 64Mb

DRAM

1Mb 4Mb DRAM

MCU

BIPM2

BIPCMOS

MOS tSlP

GaAs

CMOS BICMOS

HFMOS

BIP CMOS

Bll'OLAR

BiCMOS

CMOSNMOS

NA

NA

CMOS

CMOS

DIP CMOS

CMOS

MOS BIP

CMOS

CMOS

0.5

1.5

2,5

1.2

10

5

0.6

0.6

1.25

1.5

1

0.35

0.7

4

4

5

3

8

4

5

6

4

4

6

6

5

5

4

a

1986

1993

1980

1985

1987

1991

1987

1981

1995

1986

CD

O

CD

3 ro

CJl

CO

CO

t n

Crt

en

CO

t n

O

en

(G)

(D

CO

cn

K.

CU

. a c

CD ja

o

•a

1

Table 1 (Continued)

Europe, Africa, and Middle East Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Est.

Minimum Wafer

Geometry Diameter

Year

of Initial

Company

SOUTH AFRICAN liAlCROI;l.FCrRONlC^

SYSTEMS

TAG

TEMIC

TEMIC MATRA Mi tS

S.A.

TEMIC TIELEFUNKEN

ELECTRONIC

TEXAS INSTRUMENTS

T^XAS INSTRUMENTS

TFiXET

TCIWER

SKMICONDLFCTOR

VAISALA

WESTCODE S/C

ZETEX

ZMD GmbH

City or District

PRETORIA

Z U R i a i

HER.BRONPN

NANTKS

HEILIJRONN

FKEISING

AVEZZANO

NICE

MIGDALIIAEME^

VANTAA

CHIPPENHAM

CHADDERTON I J \ N -

CASHIRE

GRENZSTRASSE

Country

SOUTHAPRICA

Fab Name

NA

SWTFZERLAND

GERMANY

NA

NA

FRANCE

GERMANY

GERMANY

ITALY

FRANCE

ISRAEL

FIN1.AND

ENGLAND

ENGLAND

GERMANY

FABl

NA

FRIS

AMOS

NA

NA

NA

NA

NA

Z V E l

Products

Produced

CBIC ASIC

Process Technology

CMOS

DIS

ANALOG

DISCRETE

NA

BIPOLAR

256K SCRAM

MCU RKC ASIC

LIN

ANALOG

CMOS BLCMOS

M2

MPR

4Mb DRAM ASSP

CBIC

DIS

BIPOLAR

BLCMOS

CMOS BiCMOS

CMOS

NA

CMOS

MPU MCU MPR

DSP ARRAYS cusrroM

LIN

DIS

DISCRCTE

ANALOG

DRAM SRAM

NVMEM ASIC

ANALOG

CMOS

BIP

BIPOLAR

NMOS/PMOS

CMOS NMOB/

PMOS

1,5

0.6

0.5

0.8

0.6

0.5

0.8

5

2

0.8

6

6

4

6

3

2

4

1992

1994

1986

1970

1977

1991

1986

1977

1970

1983

NA = Not applicable

Fab Types:

F = Production-Based Fab

R = Semiconductor Fl&D and/or Trial PfoducHon Facility

P = Pilot Line (tnitial Production or Intended Low Volume)

T = Test and Assembly (Formerly A)

Q = Quick-Tum Fab

N = Nondedicated Foundry Service Available

D = Design Center

Source: Dataquest (December 1995)

CO

o

CD o

CD

3

C3-

CD

r>o

CO

CO en

Ei

CO

C D cn

g

@

CO

C D

C71

O

J 3

C

CD

CO

5" o o

o

1

CD

O .

Table 2

Europe, Africa, and Middle East Future Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

AMD

ANALOG DEVICES

ATMEU CORPORATION

FRAUNHOFER

INSTrn/TE

Fujnsu

City or District

DRESDEN

LIMERICK

AIX EN PROVENCE

rrzHOE

NEWTON AYCLiri^E

Coiuilry

GERMANY

IRELAND

FRANCE

GERMANY

ENGLAND

Fab Name

NA

NA

AIX EN

PROVENCE

ITZHOE

Products

Produced

K5

DSPs ANALOG

CUSIOM MIXED

SIGNAL ASICs

ASIC

Process Technology

CMOS

CMOS BiCMOS

CMOS BiCMOS

CMOS

Est.

Minimum

Geometry

(Microns)

0.25

0.6

0.4

Wafer

Diameter

(In.)

8

6

6

Year of Initial

Production

1997

1995

1997

0.8 6 1996

FAB 2 0.32 8 1997

GEC PLFASIiV

INTEL

INTEL

MtTSUBiSlil

NEC SiiMICONCUCTOR

NEC SEMICONDUCTOR

ORBIT

SEMICONDUCTOR, INC

ROBOROUGH

LEiXUF

KIRYAT GAT

Alsdorf

LIVINGSTON

LtVINGCTON

EILAT

PHILIPS

ROBERT BOSCH

SGS-THOMSON

SGS-THOMSON

SIEMENS

SIEMENS

TEMIC

NIJMEGEN

REUTLINGEN

SOUSSET

CATANIA

NEWCASTI£;

DRESDEN

NANTES

ENGLAND

IRELAND

ISRAEL

GERMANY

SCOTLAND

SCOTLAND

ISRAEL

NETHER-

LANDS

GERMANY

FRANCE

ITALY

ENGLAND

GERMANY

GERMANY

TEXAS INSTRUMENTS

AVEZZANGt

ITALY

NA = Not applicable

Fab Types:

F = PfOduction-Based Fab

R = Semiconductor R&D and/or Trial Production Facility

P = Pilot Line {Initial Production or Intended Low Volume)

T = Test and Assembiy (Formerly A)

Q = Ouick-Tum Fab

N = Nondedicated Foundry Service Available

D = Design Center

Source: Dataqusst (December 1995)

NA

FAB 14

FAB 18

NA

2 Phase 1

2 Phase 2

NA

MOS4

NA

ROUSSET

2000

NA

NA

NA

NA

PHASE 2

16Mb 64Mb

DRAM

ASIC

LOGMPU

PLASH

16M DRAM

16M/64M DRAINIE

!6M/64M DHAM

ARRAYS

CUSTOM

MIXSIGASIC

CONSUMER Ifeit

CMOS

CMOS

BiCMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

LENDIS

CUSTOM

MCU EEPROM

NVMEM

EPROM FLASH

ASIC

16Mb e4Mb

DKAM

SRAM MCU

ASIC

16Mb DRAM

BiCMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

0.5

0.25

0.35

0.35

0.35

0.35

0.8

0.35

0.6

0.35

0.5

0.35

0.35

0.35

0.5

8

6

8

8

8

8

8

8

6

8

8

8

8

8

8

1996

1998

1997

1997

1996

1998

1996

1996

1995

1998

1996

1997

1995

1996

1996

For More Information...

James Seay, Research Analyst (408) 468-8259

Internet address [email protected]

Via fax (408)954-1780

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness It does not contain material provided to us in confidence by our clients. Individual companies reported on and analyzed by Dataquest may be clients of this and/or other Dataquest services This information is not furnished in cormection with a sale or offer to sell securities or in connection with the solicitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their families may, from

jJ3t3Q!LlCSt '"'^^ *° dme, have a long or short position in the securities mentioned and may sell or buy such securities.

©1995 Dataquest Incorporated—Reproduction Prohibited

lEJS TheDunKBradsticct Corporation Dataquest is a registered trademark of A.C. Nielsen Company

DATAQUEST WORLDWIDE OFFICES

NORTH AMERICA

Worldwide Headquarters

Dataquest Incorporated

251 River Oaks Parkway

San Jose, California 95134-1913

United States

Phone: 1-408-468-8000

Facsimile: 1-408-954-1780

Dataquest Incorporated

Nine Technology Drive

RO. Box 5093

Westborough, Massachusetts 01581-5093

United States

Phone: 1-508-871-5555

Facsimile: 1-508-871-6262

Dataquest Global Events

3990 Westerly Place, Suite 100

Newport Beach, California 92660

United States

Phone: 1-714-476-9117

Facsimile: 1-714-476-9969

Sales Offices:

Washington, DC (Federal)

New York, NY (Financial)

Dallas, TX

LATIN AMERICA

Research Affiliates and Sales Offices:

Buenos Aires, Argentina

Sao Paulo, Brazil

Santiago, Chile

Mexico City, Mexico

EUROPE

European Headquarters

Dataquest Europe Limited

Holmers Farm Way

High Wycombe, Bucks HP12 4XH

United Kingdom

Phone: +44 1494 422 722

Facsimile: +44 1494 422 742

Dataquest Europe SA

Immeuble Defense Berg^res

345, avenue Georges CMmenceau

ISA 40002

92882 - Nanterre CTC Cedex 9

France

Phone: +33 1 41 35 13 00

Facsimile: +33 1 41 35 13 13

Dataquest GmbH

Kronstadter Strasse 9

81677 Mtinchen

Germany

Phone: +49 89 93 09 09 0

Facsimile: +49 89 93 03 27 7

Sales Offices:

Brussels, Belgium

Kfar Saba, Israel

Milan, Italy

Randburg, South Africa

Madrid, Spain

JAPAN

Japan Headquarters

Dataquest Japan K.K.

Shinkawa Sanko Building

6th Floor

1-3-17, Shinkawa

Chuo-ku, Tokyo 104

Japan

Phone: 81-3-5566-0411

Facsimile: 81-3-5566-0425

ASIA/PACIFIC

Asia/Pacific Headquarters

7/F China Underwriters Centre

88 Gloucester Road

Wan Chai

Hong Kong

Phone: 852-2824-6168

Facsimile: 852-2824-6138

Dataquest Korea

Suite 2407, Trade Tower

159 Samsung-dong, Kangnam-gu

Seoul 135-729

Korea

Phone: 822-551-1331

Facsimile: 822-551-1330

Dataquest Taiwan

llF-2, No. 188, Section 5

Nan King East Road

Taipei

Taiwan, R.O.C.

Phone: 8862-756-0389

Facsinrule: 8862-756-2663

Dataquest Singapore

105 Cecil Street

#06-01/02 The Octagon

Singapore 0106

Phone: 65-227-1213

Facsimile: 65-227-4607

Dataquest Thailand

12/F, Vanissa Building

29 Soi Chidlom

Ploenchit Road

Patumwan, Bangkok 10330

Thailand

Phone: 662-655-0577

Facsinrule: 662-655-0576

Research Affiliates and Sales Offices:

Melbourne, Australia

Beijing, China

Datacaiest'

a company of

The Dun & Bradstreet Corporation

®1995 Dataquest Incorporated

DataQuest

Japan Fab Database

Market Statistics

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-9504

Publication Date: December 25,1995

Filing: Market Analysis

Japan Fab Database

Market Statistics

Program: Semiconductor Equipment, IVIanufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-9504

Publication Date: December 25,1995

Filing: Market Analysis

Japan Fab Database

Table of Contents

Page

Japan Fab Database 1

Background 1

Research Methodology 1

General Definitions 1

Worldwide Geographic Region Definitions and Regional 1

RoU-Ups

Americas 1

Japan 1

Europe, Africa, and Middle East 1

Asia/Pacific 2

Definition of Table Columns 2

SEMM-WW-MS-9504 ©1995 Datequest Incorporated December 25,1995

Senniconductor Equipment, Manufacturing, and Maieriais Woridwide

List of Tables

Table Page

1 Japan Existing Pilot and Production Fab Lines 7

2 Japan Future Pilot and Production Fab Lines 19

Note: All tables show estimated data.

SEMM-WW-MS-9504 ©1995 Dataquest Incorporated December 25,1995

Japan Fab Database ^^^^^^^^^^^^^^^^^^^^^^^^

Background

This document contains the Japan portion of Dataquest's wafer fab database. The Semiconductor Equipment, Materials, and Manufacturing

Worldwide (SEMM) program uses both primary and secondary research to update this data. The tables in this report cover both merchant and captive production and pilot-line facilities, although our surveys and database also include R&D fabs.

Research Methodology

Dataquest conducts extensive annual surveys, complemented with quarterly secondary research. This data is then supplemented and crosschecked with various other information sources.

General Definitions

Fab line: A fab line is a semiconductor processing facility equipped for all front-end wafer manufacturing. Occasionally, there are two or more separate product-specific fab lines or wafer sizes in a single cleanroom.

In this situation, Dataquest documents the cleanroom as separate fab lines if the company dedicates equipment to each wafer size or product line. Therefore, a company may operate many fab lines at one location.

Front-end wafer processing: Dataquest defines front-end wafer processing as all steps involved in semiconductor processing, beginning with initial oxide and ending at wafer probe.

Production fab: A wafer fab capable of front-end processing more than

1,250 wafers per week is a production fab (type = F).

Pilot fab: A wafer fab capable of front-end processing 1,250 or fewer wafers per week is a pilot fab (type = P).

Worldwide Geographic Region Definitions and Regional Roll-Ups

Americas

Includes Central America (all nations), Canada, Mexico, United States and Puerto Rico, and South America (all nations).

Japan

Japan is the only single-country region.

Europe, Africa, and Middle East

Includes Africa (all nations), Albania, Andorra, Armenia, Azerbaijan,

Belarus, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia,

Finland, France, Georgia, Germany, Gibraltar, Hungary, Iceland, Israel,

Italy, Kazakhstan, Kyrgyzstan, Latvia, Liechtenstein, Lithuania,

Luxembourg, Malta, Middle East (all nations), Moldova, Monaco,

SEI\/IM-WW-MS-9504 ©1995 Dataquest Incorporated 1

Semiconductor Equipment, Manufacturing, and Materials Woridwide

Netherlands, Norway, Poland, Romania, Russia, San Marino,

Scandinavia, Slovakia, Spain, Sweden, Sweden, Switzerland, Tajikistan,

Turkey, Turkmenistan, Ukraine, United Kingdom, Uzbekistan, Vatican

City, and Yugoslavia (all nations within the former Yugoslavia).

Asia/Pacific

Includes Australia, Bangladesh, Cambodia, China, Hong Kong, India,

Indonesia, Laos, Malaysia, Maldives, Myanmar, Nepal, New Zealand,

Pakistan, Philippines, Singapore, South Korea, Sri Lanka, Taiwan,

Thailand, and Vietnam.

Definition of Table Columns

Products Produced contains details for seven product categories. The nomenclature used within the seven product groups of the fab database is as follows, with definitions where warranted:

• Analog

o A/D D/A: Analog-to-digital, digital-to-analog converter

• AUTOMOTIVE: Dedicated to automobile applications

a CODEC: Coder/decoder

• INTERFACE: Interface IC a LIN: Linear/analog device

a MDIODE: Microwave diode

o MESFET: Metal semiconductor field-effect transistor

o MEET: Microwave field-effect transistor o MIXSIG ASIC: Mixed-signal/linear ASIC

a MODEM: Modulator/demodulator o MMIC: Monolithic microwave IC

a OP AMP: Operational amplifier

a PWR IC: Power IC

a REG: Voltage regulator

o SMART PWR: Smart power o SWITCHES: Switching device

o TELECOM: Telecommunications chip

• Memory

a DRAM: D3Tiamic RAM o EEPROM or E2: Elecfa-ically erasable PROM

a EPROM: Ultraviolet erasable PROM

SEMi\/l-WW-l\/iS-9504 ©1995 Dataquest Incorporated December 25,1995

Japan Fab Database o FERRAM: Ferroelectric RAM

a FIFO: First-in/first-out memory o FLASH: Flash memory o MEM: Memory a NVMEM: Nonvolatile memory (ROM, PROM, EPROM, EEPROM,

FERRAM) o PROM: Programmable ROM

D RAM: Random-access memory o ROM: Read-only memory

a SPMEM: Other specialty memory (such as dual-port, shift-register, color lookup)

D SRAM: Static RAM o VRAM: Video RAM

• Micrologic

a ASSP: Application-specific standard product

a BIT: Bit slice (subset of MPU functions) o DSP: Digital signal processor o LISP: 32-bit list instruction set processor for AI

a MCU: Microcontroller unit

a MPR: Microperipheral o MPRCOM: MPR digital communication (ISDN, LAN, UART, modem)

• MPU: Microprocessor unit o RISC: Reduced-instruction-set computation 32-bit MPU

• Standard logic o LOG or LOGIC: Standard logic

• ASIC logic o ARRAYS: Gate array o ASIC: Application-specific IC o CBIC: Cell-based IC

• CUSTOM: Full-custom IC (single user) o FPGA: Field-programmable gate array

a PLD: Programmable logic device

SEMM-WW-MS-9504 ©1995 Dataquest incorporated December 25,1995

Semiconductor Equipment, Manufactuiing, and Materials Woridwide

• Discrete

o DIODE

a DIS or DISCRETE: Discrete

o FET: Field-effect transistor

o GTO: Gate turn-off thyristor

• HEMT: High-electron-mobility transistor

o IGBT: Insulated gate bipolar transistor

o MOSFET: MOS-based field-effect transistor a PWR TRAN: Power tiransistor

• RECTIHER o RF: Radio frequency a SCR: Schottky rectifier

a SENSOR o SST: Small-signal transistor

o THYRISTOR

o TRAN: Trar\sistor o ZENER DIODE

• Optoelectronic

o CCD: Charge-coupled device (imaging)

• COUPLER: Photocoupler

o lED: Infrared-emitting diode a IMAGE SENSOR o LASER: Semiconductor laser or laser IC

a LED: Light-emitting diode

a OPTO: Optoelectronic

o PDIODE: Photo diode

o PTRAN: Photo transistor

o SAW: Surface acoustic wave device

o SIT IMAGE SENSOR: Static induction transistor image sensor

Process Technology column lists four major types of technologies. This column also lists tmcommon technologies with information on well types, logic structure, and number of metal levels. Definitions used in the "Process Technology" column are as follows:

• MOS (silicon-based) o CMOS: Complementary metal-oxide semiconductor

SEI\yilVi-WW-l\/IS-9504 ©1995 Dataquest incorporated December 25,1995

Japan Fab Database

a MOS: N-channel metal-oxide semiconductor (NMOS) and p-channel metal-oxide semiconductor (PMOS).

a Ml: Single-level metal o M2: Double-level metal o M3: Triple-level metal o N-WELL o P-WELL

• POLYl: Single-level polysilicon o POLY2: Double-level polysilicon o POLYS: Triple-level polysilicon

• BiCMOS (silicon-based)

a BiCMOS: Bipolar and CMOS combined on a chip o BiMOS: Bipolar and MOS combined on a chip o ECL I/O: ECL input/output o TTL I/O: TTL input/output

• Bipolar (silicon-based) o BIP or BIPOLAR: Bipolar

a ECL: Emitter-coupled logic

a TTL: Transistor-transistor logic o STTL: Schottky TTL

• Gallium arsenide and other compoimd semiconductor materials o GaAs: Gallium arsenide o AlGaAs: Gallium aluminum arsenide o GaAs on Si: Gallium arsenide on silicon o GaP: Gallium phosphide o HgCdTe: Mercuric cadmium telluride o InAs: Indium arserude o InGaAs: Indium gallium arsenide

a InP: Indium phosphide o InSb: Indium antimony

a LiNbOS: Lithium niobate

• SOS: Silicon on sapphire

SEMM-WW-MS-9504 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Minimum Geometry is the smallest feature attainable in production volumes, measured in microns, at the critical mask layers.

Wafer Diameter represents the wafer dianneter usually expressed coUoqxiially in inches. However, for wafers greater than 3 inches in diameter, ttie colloquial expression becomes grossly inaccurate. When calculating square inches, Dataquest uses the following approximations:

• Stated diameter 4 inches (100mm) = Approximate diameter

3.938 inches

• Stated diameter 5 inches (125mm) = Approximate diameter

4.922 inches

• Stated diameter 6 inches (150mm) = Approximate diameter

5.906 inches

• Stated diameter 8 inches (200mm) = Approximate diameter 7.87 inches

Estimated Maximum Wafer Starts per Month is the equipment-limited wafer start capacity per four-week period. Start capacity is limited only by the installed equipment in the fab and the complexity of the process it runs, not by current staffing or the number of shifts operating.

SEMM-WW-MS-9504 ©1995 Dataquest Incorporated December 25,1995

CO

CD

O

CD

3.

CD

—1

ro

CJl

C D

CO

CJ1

@

(£)

O

d

CD

CO

l—f

3 o o

—1

O

CD

O .

s: cp

CD cn

g

Table 1

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

AiSHIN SEJKl

AISHIN SEIKI

A S A i n KASEI

MICRO SYSTEMS

ASAHI KASSt

MICRO SYSTEMS

CANON

CANON

CANON DENSM

CASIO

FUJI ELECTRIC

\-V\l ELECTRIC

run

ELECTRIC

FUJI ELECTRIC

FUJI FILM

MICRODEVICE

FUJI XEROX

FUJITSU

FUJITSU

FUJITSU

FUJITSU

I'UJTTSU

FUJITSU

FUJITSU

FUJITSU

FUJITSU

City or Dislricl

HANDA-^SH)

IIEKJNAN-SHT

ATSUGl-SHI

N08E0KA-SHI

HIRATSUKA-SHl

HIRATSUKA-SHI

CHICHIDU-SHI

HACraOJI-SHl

MATSUMOTO-SHI

MATSUMOTO-SHI

MATSUMOTO-Sm

MATSUMOTO-SHI.

KUROKAWA-GUN

SUZUKA-Sm

JAPAN

JAPAN

JAPAN

JAPAN

NA

NA

NA

NA

JAPAN NA

JAPAN

NA

AIZU WAKAMATSU-SHI

AIZU WAKAMATSU-Sm

JAPAN

JAPAN

BLDG. #1

BLDG. #2

AIZU WAKAMATSU-SM

JAPAN BLDG. #2-2

AIZU WAKAMATSU-Sm

AIZU WAKAMATSU-SHI

JAPAN

JAPAN

VLSIl

VLSI 2

AIZU WAKAMATSU-Sm

ISAWA-GUN

ISAWA-tiUN

JAPAN

JAPAN

JAPAN

VLSI 3

NO. 2

NO. 3

ISAV/A-aUN

Country

JAPAN

JAPAN

JAPAN

Fab Name

HANDA

SHDMKAWA

NA

JAPAN NA

JAPAN

NA

JAPAN NA

JAPAN NA

JAPAN NA

JAPAN NO. 4

Products Produced

AUTOMOTIVE

AUTOMOTIVE

TRAN CUSTOM

Process

Technology

NA

NA

MOS

SRAM, FU1.I.,

CUSTOM, OTHER

MOS LOGIC

AMORPHOUS

IMAGE SENSORS

CMOS

GaAs

ASIC

CCD

ASIC

CUSTOM ASSP

CUSTOM ASSP

CMOS

MOS

NA

MOS BiCMOS

BIPOLAR

CMOS BiCMOS

BIPOt-AR

MOS BIPOLAR MOSFET IGBT High

Vollage Diode

DIODE PWR TRAN

MOSFET

CCD Convener

Full-Custom

BIPOLAR

CMOS

PWK ICs IMAGE

SENSOR L<X;

ARRAYS LOGIC

CMOS

ARRAYS CMC

32-Bit MCU

ARRAYS U g i c CBIC

MPU

DIS A / D D / A

CMOS BIPOLAR

CMOS

CMOS

BIPOLAR

256K DRAM SRAM

EPROM MPU

DRAM SRAM ROM

ROM EPROM

1Mb iMb DRAM,

SRAM, ROM

4Mb/16Mb DRAM lMb/4Mb n.ASH

SRAM ASIC

CMOS MOS

CMOS MOS

MOS

CMOS MOS

POLY3

CMOS MOS

POLY3

Est.

Minimum

Geometry

(Microns)

Wafer

Diameter

(In.)

6

3

5

Year of

Initial

Production

1991

1990

1987

SiS;

6

1993

3

2

0.8

3

6

1

3

1.2

0.7

0.35

1.5

1

1

1 5

0.8

0.5

3

6

5

4

4

6

5

4

6

5

6

6

6

5

5

6

5

6

6

im

1990

1995

1985

1992

1986

1985

1990

1995

1983

1984

1988

1984

1987

1990

o

• 3 o

i-

@

CO

CO

CJ1

o

B

J 3

CI

CO

(n

5 " o

O

CO

o

CD

3

—1 r o

O l

C D

CD

cn

Crt

1

C D

c n

2

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

(JUJTTSU

FUJITSU

FUJITSU

FUJITSU

Fujrrsu

PUJITSU-AMD

SKMICONDUCrOR n J j r r S U O u a n h j m

Device nJHTSU-Quantum

Device

HAMAMATSU

PHOTONICS

HITACHI

HITACHI

HnAcm

HITACHI

HITACHI

HITACHI

HTTACHI

HITACHI

HITACHI

HITACHI

HTTACHI

HITACHI

HITACHI

HITACHI

HTTACHI

HITACHI

City or District

KAWASAKI-SHI

RUWANA-GUN

KUWANA-GUN

KUWANA-CUN

KUWANA-GUN

AIZU WAKAMAT^U-SHI

NAKAKOMA-GUN

NAKAKOMA-GUN

Country

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

Fab Name

NA

NO. 1

NO. 2

NO. 3 PHASE

1

NO. 3 PHASE

2

Phase 1

No.l

No.2

Products Produced

3D ICs JOSnpHSON lUNCTJON

ARRAYS

LOG ARRAYS 4Mb

DKAM

4Mb/16Mb DRAM

SRAMMPU

4MIi/16Mb DRAM

4 / 1 6 / 3 2 M b Hash

FPROM

FET LIN O I T O

HEMT

HEMTAStC

Process

Technology

NA

CMOS MOS

CMOSBIP

CMOS

CMOS

CMOS

GaAs

GaAs

JAPAN NA

HAMAMATSU-SHI

CHITOSE-SHI

CHITOSF-SHI

GOSHOGAWARA-SHI

HPTACHI-SHI

HTTACHINAKA-SHI mTACHINAKA-Sm

KODAIRA-SHI

KODAIRA-SHI

KODAIRA-SHI

KOMORO-SHI

MOBARA-Sm

MOBARA-SHI

NAKAKOMA-GUN

NAKAKOMA-GUN

NAKAKOMA-GUN

NAKAKOMA-GUN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

Chitose IF

Chitose 2F

NA

NA

N l - 1

N2-1

NA

R & D l

R&D 2

KOMORO

D l

D3

K2-1F

K2-2F

K4-1F

K4-2F

OPTO NA

4M DRAM MCU

4M DRAM MCU

1M/4M DRAM

PWR GTO

THYRISTERS TTL l M b / 4 M b D R A M

16Mb DRAM

4Mb/16rvtb DRAM

MPUMEMCI3(C

MPU SRAM DRAM

ARRAYS CBIC

LASER TELECOM

CMOS

CMOS

CMOS

BIP

CMOS

CMOS

CMOS

CMOSM2

CMOSM2

ASIC MCU EPROM

4Mb DRAM

4Mb DRAM SRAM

MCU

16Mb DRAM

MROM MPU l O G l C

MROM

CMOS GaAs

BiCMOS

MOS CMOS

CMOSM2

CMOS

CMOS

CMOS

CMOS

1.5

1.3

0.8

2

0.5

1.3

1

0.8

0.5

0.8

1.2

0.5

0.8

0.8

0.8

1.5

Est.

Minimum

Geometry

(Microns)

0.35

Wafer

Diameter

(In.)

5

Year of

Initial

Production

1988

1

0.8

0.5

6

6

1984

1987

6

B

1992

1994

0.25

0.5 8

1.5 4

1994

1984

1.5 4 1992

3

5

6

6

8

5

6

3

6

6

6

5

6

8

6

5

6

1980

1982

1990

1990

1995

1983

1988

1988

1990

1990

1983

1983

1994

1990

1987

1985

@

CO

0 3

CJl o

(U

.o

C O o

CD

o

CD

3

C T

CD

N3

cn

CO

CJl

C/3

I

<£>

cn

g

1

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company_

HITACHI

HITACHI

HrrAcm

HITACHI

HITACHT

HTTACHI

HONDA

IBM

IBM

[WATSU rvc

KAWASAKI STBBt,

City or District

NAKAKOMA-GUN

OME-SHI

TAKASAKJ-SHI

TAKASAKl-SHI

TAKASAKI-SHI

TAKASAKI-SHI

HAGA-GUN

YASU-GUN

YASU-GUN

HACHIOJl-SHI

YOKOSUKA-SHI

UTSUNOMIYA-SHI

UJI-SHT

NiSHJ-SHINJUKU

NISHIWAKI-SHI

Counliy Fab Name

JAPAN K4-3

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

D4/D5

No. 1

No. 2

No. 3

No. 4

Tochigi Lab

NA

NA

NA

NA

NA

PLANTS

LAB

F a b l

Products Produced

4Mb DRAM 1Mb

SRAM EPROM

64M DRAM 64M

Flash ASIC

DISCRETE DIP

ANALOG OFTO

MEM BIP LIN

CMOSMOS

CMOS

BIPMOS

16Mb DRAM

256K/lMbSRAM

ENG. CONTROL

SENSORS MMIC

ARRAY 1Mb DRAM

MPUROM

4M/16M DRAM

NA

IK ARRAYS DSP

CUSTOM

256K SRAM CBIC

AIU^AYS

OPTO DISCRETE

OFTO

16M DRAM ASIC

BIP BICMOS

CMOS

CMOS

CMOS BICMOS

GaAs

CMOS

CMOS

CMOS

CMOS

CMOS

GaAs GaP BIP

NA

CMOS

Est.

Minimum Wafer Year of

Process Geometry Diameter Initial

Technology (Microns) (In.) Production

0.8

0.35

2

1.2

0.5

0.5

0.6

1.5

3

0.8

0.5

6

B

S

6

8

6

3

1989

1994

1986

1988

1995

1992

1990

1990

1986

1991

1984

1992

KODENSHI

KONICA

KTI

SiiMICONUUCTOR

KYOCRRA

KYOTO

SEMICONDUCTOR

LSI LOGIC (NIHON

S/C)

LSI LOGIC (NIMON

S/C}

MATSUSHITA

MATSUSHTTA

MATSUSHTTA

MATSUSHITA

MATSUSHITA

MATSUSHITA

KANSAI-SHI

KYOTO-SHI

TSUKUBA-SHI

TSUKUBA-SHI

ARAI-SM

ARAI-SHI

ARAI-SHI

ARAI-SHI

ARAI-SHI

HIOKI-GUN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

NA

NA

F A B l

FAB 2

FABB

FAB C-1

FAB C-2

FAB D-1

FAB D-2

FAB A

NA

LED TRAN IMAGE

SENSOR

ASIC CBIC MPU

MPR SRAM

ASIC CmC MPU

MPRSRAM

MCU LOGIC ASSP

Analog

CCD

ANALOG

ANALOG

ANA1DG+E102

NA

GaAs GaP

CMOS BICMOS

CMOS BICMOS

BIP BiCMOS

BIP

CMOS

BIP

BIP

BIPOLAR

BiCMOS

0.6

1.5

2

0.8

1.8

1.5

2

4

5

4

5

5

5

1992

1989

1993

1982

1984

1984

1985

1985

1980

@

CO

CO

CJl

K.

(U

.£3

O

CD

O

CO

3

C T

CD

—I to

cn

CD

CO u i

Crt

cp

CJD en g

CD

O .

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHFTA

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHITA

MATSUSHFTA

MATSUSHITA

MATSUSHITA

MEIDENSHA

MITSUBISHI

MITSUBISHI

MITSUBISHI

MITSUBISHI

MITSUBISHI

MITSUBISHI

MTTSUBISHI

MITSUBISHI

City or District

HIOKI-GUN

KADOMA-Sm

KYOrO-SHI

KYOTO-SHI

KYOTO-SHI

NAGAOKAKYOrSHI

NAGAOKAKY&SHl

MAGAOKAKYO-SHt

NAGAOKAKYO-SHI

TON AMI-SHI

UOZU-SHI

UOZU-SHI uozu-sm

UOZU-SHI

UOZU-SHI

UTSUNOMIVA-SHI

MUMAZU-SHI rUKUOKA-SHJ l=UKUOKA-SHI

IIAMI-SHI rTAMI-Sf-n

KAMI-GUN

KAMK;UN

KIKUCHl-GUN

KDCUCHl-GUN

Country Fab Name

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

NA

S / C R 6

KAGOSHIMA

KYOTO R&D

TOYODENPA

A

B

C

V

P A B l

FAB A-1

FABB

FAB B-2

FAB C-1

FAB C-2

NA

NA

#1

#2

NA

ULSI

TAl

TA2

B-IF

B-2F

Products Produced

OFTO I.no T,ASF,R

16Mb DRAM 64-bit

MPU 64Mb DRAM

DISCRETEfSS Tr)

DRAM

D[SCRETE{SS Di)

ANALOG

Discrete (Small

Signa!)

Discrete (Pnwer)

MCU LOGIC

MOSFET

MCU ASIC

MCU

MCU DSP LOGIC

ASSP

MPU MCU

MPU MCU l O G I C

ASSF

MCU

DISCRETE (SS Tr,

Vaiictip)

GTO THYRiSTOR

PWRTRAN DIODE

BIT

BIP LIN A / D D/A

DIS Micro

FET HEMT LAZRE

DI

16Mb/64Mb/256Mb

DRAM ASIC ia.ASH

4Mb DRAM 1Mb

SRAM ASSF

8 b / 1 6 b / 3 2 b M C U

ASIC

EPROM

ARRAYS MCU

Est.

Minimum Wafer Year of

Process Geometry Diameter Initial

Technology (Microns) (In.) Production

G a A s G a P

CMOS

3

0.35

2

6

1974

1991

BIPOLAR

CMOS

BIP

BIP

BIP

BIP

MOSCMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

BIPOLAR

NA

BIPMOS

BIP

GaAs

CMOS

CMOS

CMOS

CMOS

CMOS

5

0.35

5

5

4

3

2

0.5

1.2

0.8

0,8

1

0.6

5

3

1.5

0.35

0.35

0,7

0.5

1.3

1

2

8

4

4

4

4

4

6

5

6

6

6

6

4

5

4

5

3

8

6

6

5

4

1978

1991

1975

1967

1970

1980

1982

1994

1985

1991

1995

1987

1990

1983

1985

1976

1984

1960

1993

1990

1988

1970

1975

O

CD

O

CD

3 t3-

CD ho p i

Crt

Crt

1

CO en

g

i - h

13

O

O

• 3 o

@

CO

CO en

O

03

s-

CD

CD

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

MnSUBESHI

MITSUBISHI

MirSUDISHl

MITSUBISHI

MITSUBISHI

MiTSUBISm

MITSUBiSHI

MITSUMI

MORiRJCA

ELECTRONICS

MOTOROLA

City or District

KIKUCHI-GUN

KIKUCHI-GUN

KiKUCHI-C,UN

SAIJO-SHI

SAIJO-SHI

SAgO-SHI

SAIfO-SHl

ATSUGI-SHI

YOKOHAMA-SW

YAMA-GUN

MIJRATA

MANUFACrURD^

NEC

YASU-GUN

NEC

NEC

HIGASHI

HIROSHIMA-SHI

FCGASHl

HIROSHIMA-SHI

KAWASAKI-SHI

NEC

NEC

NEC

NEC

NEC

NEC

NEC

NEC

SAGAMIHARA-SHl

SAGAMIHARA-SHI

SAGAMIHARA-SHl

TSURUOKA-SHI

TSURUOKA-SHl

ASA-GUN

ASA-GUN

KUMAMOrO-EBHt

Countiy

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

Fab Name

C-IF

C-2F

D-IF

B

C

JAPAN

JAPAN

JAPAN

JAPAN

SA2A

SA2B

NA

NA

Products Produced

1Mb SRAM 1Mb

ROM MCU FL.'^H

EPROM MCU MFU

16Mb DRAM

1Mb DRAM

ANALOG

DRAM ab MCU

ASIC

4Mb/16Mb DRAM

4Mb/16Mb DRAM

LOGDtS

OPTO

Process

Technology

M 0 5 C M 0 S

CMOS

CMOS

CMOS Ml

CMOSM2

CMOSM2

CMOSM2

BIP

Gap

JAPAN

MOS7A

LOGIC ANALOG CMOSMOS

BIPOLAR

BiCMOS

GaAs JAPAN

Yasu

I!ET MMIC

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

Dif-1

Dif-3

NA

BLDG. UL

G-1

G-2

4Mb DRAM SRAM

MFU 4Mb ROM

16Mb/64Mb DRAM

ASIC RISC

ASIC lEPROM MCU

MPUOPTO

EPROM ASIC MPU

MCU

16Mb DRAM ASIC

MPU 4Mb ROM

16Mb/64Mb DRAM

ASIC MPU

HIP LOG LIN DIS

CMOS

CMOS

CMOS N p S ;

CMOS

CMOSBI^Wi©?:

CMOS

BIP

JAPAN

JAPAN

JAPAN

TSURUOKA

Works 1

TSURUOKA

Works 2

Dif-1

JAPAN

JAPAN

LOG LIN

Dif-2 (BIdg,C)

Dif-3

4Mb DRAM 1Mb

SRAM MFU crOlC

4Mb/16Mb DRAM

ASIC

EPROM ROM

BfP

CMOSB^MOS-

CMOS

CMOSMOS

Est.

Minimum

Geometry

(Microns)

0.8

Wafer

Diameter

(In.)

5

Year of

Initial

Production

1989

1.3

0.35

0.9

5

8

5

5

1990

1994

1984

0.8

05

0.5

6

8

4

1988

1991

1994

1984

0.8

0.8

0.6

0.35

1.4

1.2

0.8

0.35

1

0.8

0.8

0.5

2 e

6

S

5

5

6

8

4

6

6

6

5

1994

1993

1990

1985

1985

1988

1990

1976

1993

1988

1993

1974

g o

CD

O

CD

3 cr

CD

</>

@

CO

CO

Si c:

•—1-

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

NEC

NEC

NEC

NEC

NEC

NEC

NEC

NEC

NEC

City or District

KUMAMOTO-SHI

KUMAMOTO-SHI

KUMAMOTO-SHI

KUMAMOTO-SHI

KUMAMOTO-SHI

OTSU-SHI

OTSU-SHI

OTSU-SHI

OTSU-SHI

NEC

NEW JAPAN RADIO

OTSU-Sm

KAMIFUKUOKA-SHI

NEW JAPAN RADIO KAMIFUKUOKA-SHI

*a

NEW JAPAN RADIO o

1 •

CD

NEW JAPAN RADIO

KAMIFUKUOKA-SHI

KAMIHJKUOKA-SHI

NIPPON PRECISION

CIRCUITS

NIPPON PRECISION

CIRCUITS

NIPPON STEEL

NASU-GUN

NASU-GUN

S A G A M I H A R A ^

Country

JAPAN

JAPAN

JAPAN

Fab Name

Dif-4

Dif-5

Dif-6

JAPAN

Dif-7

JAPAN Dif-8 -1

JAPAN

JAPAN

JAPAN

Dii-2

Dif-3

JAPAN

Dif-4

JAPAN

JAPAN

GaAs

FABl

JAPAN

FAB 2

JAPAN FAB 3

JAPAN GaAs

JAPAN BldgH

JAPAN

Dif-1

BldgS

Products Produced

BIP LCM MEM ASIC

LO(.iIC DRAM

1Mb DRAM MPU

ARRAYS

MOJ 4Mb DRAM

ASIC

Process

Technology

BIP

CMOSMOS

CMOSMOS

POLY2

CMOS BICMOS

CMOS i6Mb DRAM 4Mb

SRAM RISC MPU

PWR TRAN

LINEAR

MCU LCD Driver

SRAM 4Mb DRAM

MICRO ASIC

16bit MCU LCD

Driver ASIC

BIP

CMOSMOS

CMOSMOS

CMOS

DIODE OPTO

ANALOG OP AMP

OPTO

ANALOG A / D D /

A

LOGIC ANALOG

MCU

ANAIXX^

DISCRF.TEi OPTO

A / D D / A DSP LOG

ASSP

LOG LIN A / D D / A

MODEM

ASIC

GaAs

BIP

BIP

CMOS BIP

BiCMOS

GaAs GaAlAs

CMOS

CMOS

NA

Est.

Minimum

Geometry

(Microns)

1.4

1.2

1

Wafer

Diameter

(In.)

5

5

6

Year of

Initial

Production

1983

1978

1987

0.8

0.35

2

1

1

0-8

3

4

2.5

0.8

0.5

0.8

2

0.8

6

8

4

6

6

6

4

3

4

S

3

6

4

6

1988

1994

1978

1981

1983

1989

1988

1977

1981

1986

1984

1990

1984

1991

NIPPON DENSO

KAHIYA-Sm

JAPAN

JAPAN

S/C DEVICE

R&D CTR.

BLDG. 1 MOS

1.5 5 1987

NIPPONDENfSO

NffiSAN

NITTETSU

SEMICONDUCTOR

NHTETSU

SEMICONDUCTOR

NUKATA-CUN

YOKOSUKA-SHI

TATEYAMA-SHI

TATEYAMA-SHf

JAPAN

JAPAN

JAPAN

705

R&D Center

M l

JAPAN M2

LOG CUSTOM

MCU OPTO

MCU CUSTOM

MCU CUSTOM

MCU Logic (LCD

Driver) lMb/4MbDRAM

FLASH Logic

MOS

CMOS

CMOS

CMOS Ml

1.5

2

1.2

0.8

6

5

5

6

1993

1987

1985

1988

CO

CO t n

C/5

1

CO

0 1

g

CB

O

CD

3

CD no

_tn

CO

CO en

O)

@

CO

CO en

O j a

CD

a.

• 3 o

a.

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

NITTETSU

SEMICONDUCTOR

NKK

OKI

OKI

OKI

OKI

OKI

OKI

OKI

OLYMPUS

OLYMPUS

OMRON

City or District

TATEYAMA-Sm

AYASE-SHI

OMRON

KOUKA-GUN

ORIGIN ELECTRIC

PIONEER VIDEO

RICOH

OYAMA-SHl

KOFU-SHl

IKEDA-SHI

RICOH

RICOH

ROHM

ROHM

HACMOJI-SHI

HACmOJl-SHI

HACHIOJI-SHI

KUROKAWA-GUN

MIYAZAKl-GUN

MIYAZAKI-GUN

MIYAZAM-CUN

HACHIO|I-SHI

KAMIINA-GUN

KOUKA-GUN

KEDA-SHI

KATO-GUN

CHIKUGO-SHI

KASAOKA-SHI

Country

JAPAN

Fab Name

M3

JAPAN Phase 1

JAPAN

HI

JAPAN U l

JAPAN V3

JAPAN

SI

JAPAN

JAPAN

MI

M2

JAPAN

M3

JAPAN

JAPAN

JAPAN

NA

S/C Technology Cente

NA

JAPAN

NA

JAPAN

JAPAN

JAPAN

NA

NA

F a b l

JAPAN Fab 2

JAPAN Fab 3

JAPAN

JAPAN

APOLLO

ELECTRON-

ICS

Wako

Products Produced lMb/4Mb DRAM

Process

Technology

CMOS

ZS6K/lMb/4Mb

SRAM Flash MKOM

RISC MPU ASEC

Micro Logic Analog

CMOS

64M/256M DRAM

Micro Logic

16M/64M DRAM

Micro Gate Array

4Mb DRAM VRAM

1Mb SRAM 16M

MROM

Micro Logic

1Mb DRAM 2S6k

SRAM 4M MROM

16M DRAM 4M

DRAM

IC

SrriMAGE

SENSOR

OPTO IMAGE

SENSOK

OPTO IMAtiE

SENSOR

TRAN DIODE DIS

MPR ASIC Analog

25fik ROM G/A

CBIC VolRcg.

1Mb ROM 8b MCU

G / A CBtC

4M/8Mb ROM 16b

MCU C / A CBIC

TRAN DIS

C M O S B L C M O S fJIP

CMOS BiC

Compound

CMOS BiC

Compound

CMOS

CMOS BiCM(^

CMOS

CMOS

CMOS

CMOS

BIPGaP

BIPGaP

BIP

CMOS

CMOS Bip

CMOS

CMOS

BIP

Est.

Minimum

Geometry

(Microns)

0.5

Wafer

Diameter

(In.)

6

Year of

Initial

Production

1990

0.5

8 1992

3

0.3

0.5

0.5

2

0.8

0.4

3

3

0.8

1.5

0.8

1.5

3

2

6

6

6

4

5

6

5

4

4

4

5

6

4

6

4

1988

1992

1989

1981

1984

1991

1967

1986

1975

1987

1985

1982

1986

1990

1981

LINEAR L-pi BIP 1.2 4

1968

O)

CO

O l

g

o

CD

O

CD

3

CO

@

CD

CO

CJl

o

.a

»—••

5" o o

•a

o

3

CD

O .

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

ROHM

ROHM

ROHM

ROHM

ROHM

SANKEN

SANKEN

SANKEN

SANKEN

SANKEN

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

SANYO

City or District

KYOTO-SHI

KYOTO-SHI

KYOTO-SHI

KYOTO-SHI

K Y O T O T S H I raCASHINE-SHI

HIGASHINE-SHJ

NOZA-SHl

NnZA-SHI

NIHONMATSU-SHI

ANPACHI-GUN

ANTACHI-GUN

Ojn'A-SHI

Ojn-A-SHI

opYA-sm

OJIYA-SHI

OURA-GUN

OURA-GUN

OURA-GUN

OURA-GUN

OURA-GUN

OURA-GUN

OURA-GUN

OURA-GUN

OURA-GUN

TOTTORI-SHI

Cotinliy Fab Name

JAPAN

LSI R&D 1

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

LSI R&D 2

NA

NA

NA

NA

NA

NA

NA

NA

C 3

Microelectronics

A l

B l

B 2

C

B i p l

Bip2

Bip3

MOS2

R & D l

R&D 2

Tr

T r l

Tr2

Tottori

Products Produced

MPU LASER

MODEM TRAN

LED

MCUGA

MCU ARRAYS

SRAM EEPROM

DISCRETE OPTO

LASER OPTO

PWR TRAN DIODE

LED

PWR TRAN DIODE

PWR TRAN DIODE

LED

PWR TRAN DIODE

LED

LED

SRAM, EEPROM,

Logic (CD-ROM

LSI), Custom, CCD

SRAM, Custom,

CCD lM/4Mb DRAM, 4 /

8b MCU, DSP

EEPROM FLASH

LOGIC ASSP

ANALOG

DRAM

Analog

Analog

Analog

Ijogic ASSP MCU

Memory, Logic

Memory, I.ogic

1/)gic

TRAN (SST, Pw.T)

TRANSISTORS SST

POWER

LED, La^r diode

Process

Technology

CMOS

CMOS BiCMOS

CMOS BICMOS

BIPGaAs

BIPGaAs

Bip

Bip

Bip

Bip

Compound

CMOS

CMOS

CMOS

CMOS

BiCMOS, Bipolar

CMt)S

Bipolar

Bipolar

Bipolar

CMOS, MOS

Dipolar

CMOS

CMOS, MOS

MOS, Bipolai'

MOS, BIPOLAR

Compound

Est.

Minimum

Geometry

(Microns)

0.8

Wafer Year of

Diameter Initial

(In.) Production

1989

0,5

0.8

0.8

0.4

0.8

0.7

1

0.6

2

1.6

1.2

1

1.8

0.6

1.2

1

2

6

6

6

4

4

4

3

4

5

4

5

6

2

4

4

5

4

3

3

5

6

5

6

3

1994

1989

1981

1989

1970

1986

1991

1986

1990

1985

1989

1987

1994

1967

1981

1991

1984

1984

1984

1980

1980

CO

CO en

O

(^

CD

3 ro

O l

C O

C O

cn

CO en

1

C D

cn

g

@

C O

C O

cn

O

K.

cz

CO

C/i

o

• o

o i

CD

Q .

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

SEIKO El'SON

SFDCO EPSON

SEIKO EPSON

SEIKO EPSON

SEIKO

[NSTRUMENTS

SEIKO

INSTRUMENTS

SHARP

SHARP

SHARP

SHARP

SHARP

SHARP

SHARP

SmMADZU

SMNDENGEN

SHINDENGEN

SHINDENGEN

SHINDENGEN

SHINDENGEN

SHINDENGEN

SHINDENGi-N

SONY

SOHY

SONY

City or District

SAKATA-SHI

SUWA-GUN

SUWA<;UN

SUViA-GUN

MATSUDO-Sm

MATSUDO-SHI

FUKUYAMA-SHI

FUKUYAMA-SHl

FUKUYAMA-SHI

TENRI-SHI

YAMATO

KORfi-AMA-SHI

ATSUGI-SHI

HANNO-SHI

HANNO-SHl

HIGASHINE-SHI

HTGASHINE-SHI

HIGASHINE-SKI

HONJO-SIU

HONJO-SHl

ATSUGI-SHI

ATSUGI-SfD

ATSUGI-SHI

Country

JAPAN

JAPAN

Fab Name

Bldg3

BLDG. A

JAPAN BLDG. B

JAPAN

JAPAN

BLDG. D

BLDG. B

JAPAN NA

JAPAN FACTORY 1

JAPAN FACTORY 2

JAPAN

FACTORY 3

KITA KATSURAGI-GUN

TENRI-SHI

JAPAN

JAPAN

NA

FACTORY 2

JAPAN

FACTORY 3

JAPAN NA

JAPAN

JAPAN

Atsugi

R&D Center

JAPAN Trial

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

BLDG. 1

BLDG. 2

BLDG. 3

Ohura -1

Ohura -2

NA

NA

JAPAN NA

Products Produced

FPGA PLD CBIC

ARRAYS J56K

SRAM EFROM

ARRAYS CBIC

SRAM EEFROM

IMli SRAM ASIC

TELECOM

Process

Technology

CMOS

CMOS

CMOSMOS

CMOS BICMOS

CMOS

Est.

Minimum

Geometry

(Microns)

0.65

1.5

Wafer

Diameter

(In.)

6

5

Year of

Initial

Production

1991

1985

2.5

0.8

2

4

6

4

1981

1989

1983

SRAM ARRAYS

CBIC liEPROM

MROM 8bil MCU

ARRAYS CBIC

16Mb MROM,

DRAM, SRAM

PlHsh, 32M MROM,

SRAM

Optocoupler

SRAM, MROM,

MCU, LCD

SRAM, MROM,

MCU, ASIC

LASER, LED, OFTO

CMOS

CMOS

CMOS

CMOS

GaAs

CMOS BIP

CMOSBIP

GaAs

Laser Diode

PWR MOSFET

I-nfBRID

PWR MOSFET

HYBRID

TRANSISTOR

MOSniT

POWERTRAN

DIODE THYRISrOR

DEODE THYRiSTOR

LIN

EEPROM -IMb

VRAM 4Mb SRAM

FET LASER CCD

HEMT

Compound

MOSBIP

MOSBIP

MOSBIP

MOSBIP

MOSBIP

BIP

BIP

BIP

CMOS

GaAs

0.8

0.8

0.6

0.6

0

1.2

0.35

5

5

10

10

10

10

10

2

0.8

6

5

6

8

3

5

5

3

4

4

4

3

3

4

2

2

4

6

3

1987

1985

1989

1991

1981

1977

1980

1989

1991

1985

1987

1993

1972

1983

1987

1988

O

CD

O

CO

3

D -

CO

yi

CO

CO t n

@

CO

CD

0 1

O

a>

55c

CO

CO

•3 o i

C2.

<y>

Crt

2

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Cant pa ny

SONY

SONY

StlNY

SONY

SONY

SONY

SONY

SONY

STANLEY

STANLEY

SUMITOMO METAL

INDUSTRIES

TEXAS

INSTRUMENIS

TEXAS

INSTRUMENTS

TEXAS

INSTRUMENTS

TEXAS

INSTRUMENTS

TEXAS

INSTRUMENTS

TOHOKU

SEMICONDUCTOR

TOHOKU

SEMICONDUCTOR

TOHOKU

SEMICONDUCTOR

TOKCN

TOKO

TOKO

KOKUBU-SHI

HADANO-SHI

AMAGASAKI-SHl

KATOGAYA-SHI

HAYAMI-GUN

HAYAMI-GUN

INASHIKK5UN

[NASHIKI<]UN

SENDAl-SHI

SENDAI-SHI

SENDAI-Sm

SENDAI-SHI

IRUMAGUN

IRUMA-GUN

TOSf-DBA

TOSHIBA

City or District

ISAHAYA-SHI

ISAHAYA-SH3

ISAHAYA-SHI

KOKUBU-SHI

KOKUUU-SHl

KOKUBU-SHI

KOKUBU-SHI

HIMEJl-SITI

HIMEJISHI

Country

JAPAN

JAPAN

Fab Name

IG

2G

JAPAN 3G

JAPAN

JAPAN

JAPAN

JAPAN

#2 PHASE!

#2 PHASE2

#4

#6

JAPAN

JAPAN

JAPAN

JAPAN

NA

NA

NA

NA

JAPAN Hato

JAPAN Hijil

JAPAN HijiS

JAPAN MihoS

JAPAN Miho 6

JAPAN STEPl

JAPAN STEP 2

JAPAN STEP 3

JAPAN

JAPAN

JAPAN

NA

NA

NA

JAPAN No.l

JAPAN No.2

Products Produced

1Mb SRAM [.OG

CCD 256Kb SRAM

1Mb SKAM

1Mb SRAM 4Mb

VRAM CCD

DIS

LIN A / D D / A

SRAM MCU CCD

E.OG MEM MCU

LIN

CCD

LASER LED

LED

4Mb DRAM

ARRAYS

Analog LCD Driver

ASSP

I-ogic Linear G/A

Process

Technology

CMOS

CMOS

CMOS

BIP

BIP

MOSCMOS

CMOS BiCMOS

CMOS

Compound

Compound

NA

CMOS NMOS

BIP

Eat.

Minimum

Geometry

(Mkrons)

1

0.8

Wafer

Diameter

(In.)

6

6

Year of

Initial

Production

1989

1989

0.5

3

2

1 3

0 8

0.6

0.8

1

1

6

4

5

5

6

5

4

6

5

5

1991

1975

1976

1978

1992

1973

1986

1993

1991

1982

1974

4M/16M DRAM

ASSP ASIC MPU

DSP CBIC

IM/4M DRAM

ASSP MTU

1Mb DRAM MCU

MPU

4Mb DRAM MPU

MCU

I6Mb/64Mb DRAM

MOS

C - M O S f f i p ( ^

C^OS

C M O S W q R p i

CMOS

POWER srr

NA

A / D D / A

TELECOM DIODE

PWR tliT TRAN

DIODE

TRAN DIODE

CMOSBtCMC?

BIP

MOS

BIT

MOS BIP

BIP NMOS

0 8

1

0 8

1

0.8

0.35

3

3.5

2

2

6

5

6

6

6

8

3

5

5

4

5

1982

1988

1991

1995

1990

1990

1990

1982

O

CD

O

CD

3

cr

CO

O i

C O

C O

en

CO crt

t o

0 1

g

5" o o

1

CD

Q .

@

C O

C O

en

O

s-

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

TOSHIBA

TPSHIHA

TOSHIBA

TOSHIBA

TOSHTOA

TOSHIBA

TOSHTOA

TOSHIBA

TOSHIBA

TOSHIBA

TOSHIBA

TOSHIBA

TOSHIBA

TOSHIBA

TOSHIBA

TOSHIBA

TOSHIBA

City or District

KAWASAKl-SM

KAWASAKI-Sm

KTTA KYUSHU-SHI

KITA KYUSHU-SHI

KITA KYUSHU-SHI

KITAKAMI-SHI

KUAKAMl-Sm

K[TAKAMI-SH]

KITAKAMl-SHi

KITAKAMl-SHI

NOMI-GUN

OriA-SHI

orrA-sHi

0!TA-SHI

OITA-SHl

YOKKAICHI-Sm

YOKKAICHI-Sin

TOSHIBA

COMPOMENTS'

TOYOUA

AUTOMATIC LOOM

WORKS

TOYODA GOSEI

KIMITSU-SHl

ODU-SHI

INAZAWA-SHI

Country

JAPAN

Fab Name

Bldg.108 D-1

JAPAN Bldg.108 D-2

JAPAN

JAPAN

JAPAN

JAPAN

Kubik 1

Kubik 2

Kubik 3

Bldg.lOl, D-1

JAPAN Bldg.lOl, D-2

JAPAN

JAPAN

Bldg.102, D-3

Bldg.102, D-4

JAPAN Bldg.106, D-5

JAPAN

JAPAN

JAPAN

JAPAN

NA

C-CUBED 1

C-CUBED 2

C-CUBED 3

JAPAN

JAPAN

JAPAN

JAPAN

C-CUBED 4

Y-CUBED,

#l-Mod 1

Y-CUBED,

#l-Mod 2

PHASE 1 & 2

Products Produced

PVVR TRAN LIN

16Mb/64Mb DKAM

FLASH

BET Analog Opto

BIT Analog

Analog

CCD ASIC MPU

MCU MROM

CCD AStC MPU

MCU MROM

Asrc

EPROM MROM

MPU ASIC

EPROM MROM

MPU ASIC

PWRTRAN

MCU SRAM LOGIC

MCU SRAM LOGIC

MCU ASIC DRAM

SRAM

4Mb/16MDRAM

4Mb/16Mb DRAM

Process

TechnolDgy

m' BiCMOS

GaA!;

CMOS

Bn> GaAs

BtCMOS BiF

BIP BICMOS

CMOS

CMOS

CMOS

CMOS

BICMC$;ElB(C3g

BIP

CMOS

CMOS

CMOS

CMOS

CMOS

4Mb/16Mb DRAM CMOS

JAPAN KYOWA

DIODE RECTIFIER

THYRISTOR

POWER TRAN

BIP

Bip

Est.

Minimum Wafer Year of

Geometry Diameter Initial

(Microns) (In.) Production

1970

0.35

1.5

1.2

1

1

1

0.8

0.8

2

1.2

1.2

1

0.5

0.5

0.35

4

4

5

6

5

4

4

5

5

6

8

8

4

4

1990

1986

1988

1993

1984

1986

1989

1991

1993

1992

1986

1987

1989

1991

1993

1994

1970

1990

LED COMPOUND

25

Z

1993

TOYODA MACHINE

WORKS

TOYOTA MOTOR

KARIYA-Sm rOYOTA-SHl

UNIZON ITAMI-SHI

JAPAN

JAPAN

JAPAN

Technology

Center

HIGASHI

KARIYA

ME

JAPAN

NA

ASIC CMOS

MCU PWR ICs

CUSTOM

ZENER DIODE REG

ARRAYS

CMOS BIP

BIP

5

5

5

1990

1990

CD o

CD

3

cr

CD r\3

en

( O

CO f

@

i

1

CO

Table 1 (Continued)

Japan Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

C/5

1

g

Company

YAMAHA

City or District

AIRA-GUN

Country

JAPAN

Fab Name

F a b l

YAMAHA

YAMAHA

YAMAHA

YOKOGAWA IMT

AIRA-GUN

IWACA-GUN

TOYOOKA-MURA

KAMI INA-GUN

JAPAN

JAPAN

JAPAN

JAPAN

Fab 2

Building U

NA

Products Produced

LIN ROM CBIC

ASSP MPR

ROM CBIC ASSP

CBIC LOC;

ASIC MFR

TRAN DIODE

OPTO Analog

Process

Technology

CMOS MOS

CMOS

CMOS

CMOS M 2 / M 3

BIPCMOS

NA = Not applicable

Fab Types:

F = Product) on-Based Fab

R = Semiconductor H&D and/or Trial Production Facility

P = Pilot Line (Initial Production or Intended Low Volume)

T = Test and Assembly (Formerly A)

Q = Quick-Turn Fat)

N = Nondsdicated Foundry Service Available

D = Design Center

Source: Dataquest (December 1995)

Est.

Minimum

Geometry

(Microns)

0.65

Wafer

Diameter

(In.)

5

Year of

Initial

Production

1988

0.65

0.8

0.5

3

6

6

6

4

1995

1986

1990

1988

CO

1

C/J

1

C D

CJ1

g

@

CO

CO cn

- O

CD

C/3

» — * •

5" o o

CD

K ) cn

U3

CD

CJ1

Table 2

Japan Future Pilot and Production Fab Lines (Including Fabs Beginning Operation during 199

Compiny

HJJl ELECTRIC

rujrrsu

FUJITSU

HJJITSU

FUJITSU

FUJITSU-AMD

SEMICONDUCTOR

HITACHI

HiTACHI

HITACHI

HITACHI

KAWASAKI ST P.F.I,

City or District

MATSUMOTO-SM

AKDCAWA

ISAWA-GUN

AIZU WAKAMATSU-SHI

JAPAN

JAPAN

JAPAN

NA

NO. 4-2

Phase 2

CHTTOSE-SHt

HTTACHINAKA-SHl

NAfCAKOMA-GUN

TAKASAKI-SHI

UTSUNOMiYA-SHl

Country

JAPAN

Fab Name

NA

AIZU WAKAMATSU-SHI JAPAN

AIZU WAKAMATSU-SHI JAPAN

BLDG. #2-2

BLDG. #2-3

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

Chitose2

N2-2

K2-2F

No.3

PHASE 2

KTI

SEMICONDUCTOR

MATSUSHTTA

MATSUSHTTA

MITSUBISHI

MITSUBISHI

NEC

NEC

OKI

TOHOKU

SEMICONOpra;]^

TOSHIBA

YAMAHA

YAMAHA

N(SH1WAKI-SHI

TONAMJ-SHJ

UOZU-SHI

KIKUCHI-GUN

SAUO-SHI

HIGASHI

HlROSHIMA-SFn

KUMAMOTO-Sm

KUROKAWA-CUN

SENDAI-SHI

YOiCKAlCHr-SHI

AIRA-GUN

TOYGOKA-MURA

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

JAPAN

Fab 2

NA = NoS applicable

Fab Types:

F = Production-Based Fab

R = Semiconductor R&D and/or Trial Pro^BffBon Facility

P = Pilot Line (Initial Production or I n t e n d ^ Low Volume)

T = Test and Assembly (Formerly A)

Q = Quick-Turn Fab

N = Nondedicated Foundry Service Availatilb

D = Design Center

Source: Dataquest (December 1995)

FAB 2

FAB B-2

D-lF-2

SA2F

Dif-2

Dif-8-2

S2

STEP 3

Y-CUBED, #2

Fab 2

Building 11-2

Products Produced

MOSFET IGBT Higii

Voltage Diode

Process

Technology

MOS BIPOLAR

ARRAYS Logic CBIC

MPU

ARRAYS Logic CBIC

MPU

CMOS

CMtlS

16Mb DRAM

16/32Mb Flash

CMOS

CMOS

CMOS

64Mb DRAM l6Mb/MMb DRAM

16Mb DRAM

16Mb DRAM

DRAM SRAM

ARRAYS

16M/64M DRAM

ASIC

16Mb DRAM ASIC

MPU MCU

16Mb DRAM

16Mb &4Mb DRAM

16Mb DRAM ASIC

RISC

64M/2S6M DRAM

16M/64M DRAM

I6Mb/64Mb DRAM

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOSNB^

CMOS

CMOS

CMOS

CMOS

16Mb/64Mb DRAM

ROM CBIC ASSF

ASIC MI^R

CMOS

CMOS

CMOS M2/M3

Est.

Minimum

Geometry

(MicronsI

3

Wafer

Diameter

(In.)

5

Year of

Initial

Production

1995

0.35

0.35

0.3

0.35

0.35

0.35

0.35

0.5

0.5

0.5

0.35

0.35

0.8

0.35

0.35

0.35

0.25

0.4

0.35

0.35

0.65

0.5

6

8

8

8

8

8

8

8

8

8

8

8

6

6

8

6

8

8

8

8

8

8

1995

1996

1996

1996

1997

1998

1996

1995

1995

1996

1997

1996

1995

1996

1998

1996

1998

1996

1995

1996

1995

1996

For More Information...

James Seay, Research Analyst (408) 468-8259

Internet address [email protected]

Via fax (408)954-1780

The content of this report represents our interpretation and analysis of information generally available to the public

_ or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our clients. Individual companies reported on and analyzed by Dataquest n\ay be clients of this and/or other Dataquest services. This infomnation is not furnished in connection with a sale or offer to sell securities or in connection with the solicitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their families may, from

J J 3 t 3 Q U C S t '™^^ to &me, have a long or short position in the securities mentioned and may sell or buy such securities.

©1995 Dataquest Incorporated—Reproduction Prohibited

B3 ThcDun&BradstrcctCorDOiation Dataquest is a registered trademark of A.C. Nielsen Company

DATAQUEST WORLDWIDE OFFICES

NORTH AMERICA

Worldwide Headquarters

Dataquest Incorporated

251 River Oaks Parkway

San Jose, California 95134-1913

United States

Phone: 1-408-468-8000

Facsimile: 1-408-954-1780

Dataquest Incorporated

Nine Technology Drive

RO. Box 5093

Westborough, Massachusetts 01581-5093

United States

Phone: 1-508-871-5555

Facsimile: 1-508-871-6262

Dataquest Global Events

3990 Westerly Place, Suite 100

Newport Beach, California 92660

United States

Phone: 1-714-476-9117

Facsimile: 1-714-476-9969

Sales Offices:

Washington, DC (Federal)

New York, NY (Financial)

Dallas, TX

LATIN AMERICA

Research Affiliates and Sales Offices:

Buenos Aires, Argentina

Sao Paulo, Brazil

Santiago, Chile

Mexico City, Mexico

EUROPE

European Headquarters

Dataquest Europe Limited

Holmers Farm Way

High Wycombe, Bucks HP12 4XH

United Kingdom

Phone: +44 1494 422 722

Facsimile: +44 1494 422 742

Dataquest Europe SA

Immeuble Defense Berg&res

345, avenue Georges Cl§menceau

TSA 40002

92882 - Nanterre CTC Cedex 9

France

Phone: +33 1 41 35 13 00

Facsimile: +33 1 41 35 13 13

Dataquest GmbH

Kronstadter Strasse 9

81677 Miinchen

Germany

Phone: +49 89 93 09 09 0

Facsimile: +49 89 93 03 27 7

Sales Offices:

Brussels, Belgium

Kfar Saba, Israel

Milan, Italy

Randburg, South Africa

Madrid, Spain

JAPAN

Japan Headquarters

Dataquest Japan K.K.

Shinkawa Sanko Building

6th Floor

1-3-17, Shinkawa

Chuo-ku, Tokyo 104

Japan

Phone: 81-3-5566-0411

Facsimile: 81-3-5566-0425

ASIA/PACIFIC

Asia/Pacific Headquarters

7/F China Underwriters Centre

88 Gloucester Road

Wan Chai

Hong Kong

Phone: 852-2824-6168

Facsimile: 852-2824-6138

Dataquest Korea

Suite 2407, Trade Tower

159 Samsung-dong, Kangnam-gu

Seoul 135-729

Korea

Phone: 822-551-1331

Facsimile: 822-551-1330

Dataquest Taiwan

llF-2, No. 188, Section 5

Nan King East Road

Taipei

Taiwan, R.O.C.

Phone: 8862-756-0389

Facsimile: 8862-756-2663

Dataquest Singapore

105 Cecil Street

#06-01/02 The Octagon

Singapore 0106

Phone: 65-227-1213

Facsimile: 65-227-4607

Dataquest Thailand

12/F, Vanissa Building

29 Soi Chidlom

Ploenchit Road

Patumwan, Bangkok 10330

Thailand

Phone: 662-655-0577

Facsimile: 662-655-0576

Research Affiliates and Sales Offices:

Melbourne, Australia

Beijing, China

Dataquest*

a company of

The Dun & Bradstreet Corporation

®1995 Dataquest Incorporated

DataQuest

Americas Fab Database

Market Statistics

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-9503

Publication Date: December 25,1995

Filing: Market Analysis

Americas Fab Database

Market Statistics

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-9503

Publication Date: December 25,1995

Filing: Market Analysis

Americas Fab Database

Table of Contents

Page

Americas Fab Database 1

Background.,... 1

Research Methodology 1

General Definitions 1

Worldwide Geographic Region Definitions and Regional

RoU-Ups 1

Americas i

Japan 1

Europe, Africa, and Middle East 1

Asia/Pacific i2

Definition of Table Columns 2

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, IVianufacturing, and IVIaterials Worldwide

List of Tables

Table Page

1 Americas Existing Pilot and Production Fab Lines 7

2 Americas Future Pilot and Production Fab Lines 18

Note: All tables show estimated data.

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

Americas Fab Database

Baclcground

This document contains the Americas portion of Dataquest's wafer fab database. The Semiconductor Equipment, Materials, and Manufacturing

Worldwide (SEMM) program uses both primary and secondary research to update this data. The tables in this report cover both merchant and captive production and pilot-line facilities, although our surveys and database also include R&D fabs.

Researcli IVIetliodology

Dataquest conducts extensive annual surveys, complemented with quarterly secondary research. This data is then supplemented and crosschecked with various other information sources.

General Definitions

Fab line: A fab line is a semiconductor processing facility equipped for all front-end wafer manufacturing. Occasionally, there are two or more separate product-specific fab lines or wafer sizes in a single cleanroom.

In this situation, Dataquest documents the cleanroom as separate fab lines if the company dedicates equipment to each wafer size or product line. Therefore, a company may operate many fab lines at one location.

Front-end wafer processing: Dataquest defines front-end wafer processing as all steps involved in semiconductor processing, beginning with initial oxide and ending at wafer probe.

Production fab: A wafer fab capable of front-end processing more than

1,250 wafers per week is a production fab (type = F).

Pilot fab: A wafer fab capable of front-end processing 1,250 or fewer wafers per week is a pilot fab (type = P).

Worldwide Geographic Region Definitions and Regional Roli-Ups

Americas

Includes Central America (all nations), Canada, Mexico, United States and Puerto Rico, and South America (all nations).

Japan

Japan is the only single-country region.

Europe, Africa, and Middle East

Includes Africa (all nations), Albania, Andorra, Armenia, Azerbaijan,

Belarus, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia,

Finland, France, Georgia, Germany, Gibraltar, Hungary, Iceland, Israel,

Italy, Kazakhstan, Kyxgyzstan, Latvia, Liechtenstein, Lithuania,

Luxembourg, Malta, Middle East (all nations), Moldova, Monaco,

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated 1

Semiconductor Equipment, Manufacturing, and Materials Woridwide

Netherlands, Norway, Poland, Romania, Russia, San Marino,

Scandinavia, Slovakia, Spain, Sweden, Sweden, Switzerland, Tajikistan,

Turkey, Turkmenistan, Ukraine, United Kingdom, Uzbekistan, Vatican

City, and Yugoslavia (all nations within the former Yugoslavia).

Asia/Pacific

Includes Australia, Bangladesh, Cambodia, China, Hong Kong, India,

Indonesia, Laos, Malaysia, Maldives, Myanmar, Nepal, New Zealand,

Pakistan, Philippines, Singapore, South Korea, Sri Lanka, Taiwan,

Thailand, and Vietnam.

Definition of Tabie Coiumns

Products Produced contains details for seven product categories. The nomenclature used within the seven product groups of the fab database is as follows, with definitions where warranted:

• Analog o A / D D/A: Analog-to-digital, digital-to-analog converter o AUTOMOTIVE: Dedicated to automobile applications a CODEC: Coder/decoder

a INTERFACE: hiterface IC a LIN: Linear/analog device

o MDIODE: Microwave diode

o MESFET: Metal semiconductor field-effect transistor a MEET: Microwave field-effect transistor

a MIXSIG ASIC: Mixed-signal/linear ASIC

• MODEM: Modulator/demodulator o MMIC: Monolithic microwave IC o OP AMP: Operational amplifier

a PWR IC: Power IC

a REG: Voltage regulator o SMART PWR: Smart power

a SWITCHES: Switching device o TELECOM: Telecommunications chip

• Memory

a DRAM: Dynamic RAM

o EEPROM or E2: Electi-ically erasable PROM

a EPROM: Ulbraviolet erasable PROM

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

Americas Fab Database o FERRAM: Ferroelectric RAM o FIFO: First-in/first-out memory o FLASH: Flash memory o MEM: Memory o NVMEM: Nonvolatile memory (ROM, PROM, EPROM, EEPROM,

FERRAM) a PROM: Programmable ROM

a RAM: Random-access memory o ROM: Read-only memory

• SPMEM: Other specialty memory (such as dual-port, shift-register, color lookup) o SRAM: Static RAM o VRAM: \rideo RAM

• Micrologic

a ASSP: Application-specific standard product

• BIT: Bit slice (subset of MPU functions) o DSP: Digital signal processor

a LISP: 32-bit list instruction set processor for AI o MCU: Microcontroller imit o MPR: Microperipheral o MPRCOM: MPR digital communication (ISDN, LAN, UART, modem) a MPU: Microprocessor unit a RISC: Reduced-instruction-set computation 32-bit MPU

• Standard logic o LOG or LOGIC: Standard logic

• ASIC logic

a ARRAYS: Gate array o ASIC: Application-specific IC

a CBIC: Cell-based IC

. o CUSTOM: Full-custom IC (single user) o FPGA: Field-programmable gate array o PLD: Programmable logic device

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, Manufacturing, and Materials Wortdwide

• Discrete

a DIODE a DIS or DISCRETE: Discrete

o FET: Field-effect transistor

a GTO: Gate turn-off thyristor

• HEMT: High-electron-mobility transistor

o IGBT: Insulated gate bipolar transistor

a MOSFET: MOS-based field-effect transistor

o PWR TRAN: Power transistor

o RECTIFIER

o RF: Radio frequency

o SCR: Schottky rectifier

a SENSOR

o SST: Small-signal transistor o THYRISTOR

o TRAN: Transistor

o ZENER DIODE

• Optoelectronic

o CCD: Charge-coupled device (imaging)

o COUPLER: Photocoupler

o lED: Infrared-emitting diode

o IMAGE SENSOR

o LASER: Semiconductor laser or laser IC

o LED: Light-emitting diode

a OPTO: Optoelectronic

a PDIODE: Photo diode o PTRAN: Photo transistor

o SAW: Surface acoustic wave device o SIT IMAGE SENSOR: Static induction transistor image sensor

Process Technology column lists four major types of technologies. This column also lists uncommon technologies with information on well tjrpes, logic structure, and number of metal levels. Definitions used in the "Process Technology" column are as follows:

• MOS (silicon-based)

o CMOS: Complementary metal-oxide semiconductor

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

Americas Fab Database

• MOS: N-channel metal-oxide semiconductor (NMOS) and p-channel metal-oxide semiconductor (PMOS). o Ml: Single-level metal

• M2: Double-level metal o M3: Triple-level metal o N-WELL o P-WELL o POLYl: Single-level polysilicon o POLY2: Double-level polysilicon o POLY3: Triple-level polysilicon

• BiCMOS (silicon-based)

a BiCMOS: Bipolar and CMOS combined on a chip

a BiMOS: Bipolar and MOS combined on a chip o ECL I/O: ECL input/output o TTL I/O: TTL input/output

• Bipolar (silicon-based) o BIP or BIPOLAR: Bipolar o ECL: Emitter-coupled logic o TTL: Transistor-transistor logic o STTL: Schottky TTL

• Gallium arserude and other compound semiconductor materials o GaAs: Gallium^ arsenide o AlGaAs: Gallium aluminum arsenide o GaAs on Si: Gallium arsenide on silicon

• GaP: Gallium phosphide o HgCdTe: Mercuric cadmium telluride o InAs: Indium arsenide o InGaAs: Indium gallium arsenide

a InP: Indium phosphide

a InSb: Indium antimony o LiNb03: Lithium niobate o SOS: Silicon on sapphire

Minimum Geometry is the smallest feature attainable in production volumes, measured in microns, at the critical mask layers.

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

Semiconductor Equipment, Manufacturing, and Materials Woridwide

Wafer Diameter represents the wafer diameter usually expressed colloquially in inches. However, for wafers greater than 3 inches in diameter, the colloquial expression becomes grossly inaccurate. When calculating square inches, Dataquest uses the following approximations:

• Stated diameter 4 inches (100mm) = Approximate diameter

3.938 inches

• Stated diameter 5 inches (125mm) = Approximate diameter

4.922 inches

• Stated diameter 6 inches (150mm) = Approximate diameter

5.906 inches

• Stated diameter 8 inches (200mm) = Approximate diameter 7.87 inches

Estimated Maximum Wafer Starts per Month is the equipment-limited wafer start capacity per four-week period. Start capacity is limited only by the installed equipment in the fab and the complexity of the process it runs, not by current staffing or the number of shifts operating.

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

Americas Fab Database

a

e g

i

1-1 Pvl

g

"X

111

= g i

^r

111

CM IT) t-s ^ 3 t o " ^

00 00 OO 00 OO OO ^

^ O^ O^ CT^ C^ O^ O^ n i n sD vo CO \D m

£H^

S'

III

2 m CD CQ

I

a:

<

NJ g

1

g

o o s. g lO"

g

2

f

o g

\ D

1

sT

•D

0)

3

O

O

VO LD U^

^o n m ^o

2

§

y = ^ 1 1 t! ^ 5 ^ > n

iPi^iP

ifeg < | | i <

<Uaj O S F ^ U < <

O

U

D

„ S =^ I

0) 9^

•r^ r-<

' - ' 60

CTJ

2 ci u .S

O V

(t: 0^

T f

C BO o

C!

C

PLH bO

01

«.P0

01

* ^ N

**

V)

^

rt

X b

W

bO

CA rt u T3

f-l

QJ

g

^

u

< 6i

g

<

2

<

2

•V

CD

IT)

CD

2

in

IN

CCi

g

a

Vl

< <

2 2

< < < < < < < ; - i < c / i i / i c / i c / i u ^ c / i u 5 t / J c/5 i b 3 i 3 D 3 3 b

< <

t/S tri

3 3

i

S d a c <

2 U

S f-

§n

I

s

U

Q

8

O n

tn t n

§

^

£ y

i>

zi

• y

<

2 u

< <

I

8

2

i

o c

i y

a

i I

<

<

I

sg

< <

i n tft

3 b

2 g-

<

z

< < < < < < < <

en c f t ; / i tn c / i t n u i t n

3 b b 3 3 3 3 3

O

5

g 8

O

( J n S^ y < tn < en

2 S S S o S o

5 D a 0 2 O 2 g S iS OS O £

O OS « u ffi U S

a a s a

y

S

E -

s

SEMM-WW-MS-9503 ©1995 Dataquest Incorporated December 25,1995

• a o

CO

o.

O

CO

o

CD

3 c r

CD

ro en

CD

CO t n

@

CO

CO

en

O

. o en

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

en

I

CO

en o

CO

Company

AVAtvrrtK

BALL AEROSPACE

BELL NORTItERN RESEARCH

UIPOLARICS INC.

City or

District

State Countiy Fab Name

OTTAWA, ONTARIO C N U.S.A. NA

BURRDROWN CORPORATION TUCSON

CAUrORNIA MICRODEVICES MILPITAS

CAIJFORNIA MICRODEVICES TEMPE

AZ U.&A. MICROTECH

CA U.&A, THIN FILMS

AZ U,S.A. MICRO DIV

CALOGIC

CELERTTEK INC.

CHERRY SEMICONDUCTOR

COMMOCORE

SEMICONDUCTORS

COMPENSATP.D DEVICES

CRAY RESEARCH

CRAY RESEARCH

CRYSTALONICS

CYPRESS SEMICONDUCTOR

CYPRESS SEMICONDUCTOR

CYPRESS SEMICONDUCTOR

CYPRESS SEMICONDUCllQR

DALLAS SEMICONDUCTOR

IJALLAS SEMICONDUCTOR

DAVID SARNOf=r LABS

FREMONT

SAN JOSE

EAST GREENWICH

CA

CA

RI

PA NORRISTOWN

MELROSE

MA

U.S. A,

U.S. A.

U S , A.

U.S.A.

U.S.A.

CHIPPBVA FALLS

CHIPPEWA FALUS

CAMBRIDGE

BLOOMINGTDN

BLOOM INGTON

ROUNDROCK

SAN JOSE

DALLAS

DALLAS

PRINCETON

CA

TX

TX

NJ

WI

Wl

MA

MN

MN

TX

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U S A .

U.S.A.

U.S.A.

NA

NA

BIPOLAR

F A B l

N A

NA

NA

NA

FAB 3

FAB 4

FAB 2

F A B l

FAB 10

FAB 11

SILICON IC CENTER

DELCO ELECTRONICS

CORPORATION

DELCO ELECTRONICS

CORPORATION ni:i,CO ELECTRONICS

CORPORATION

DIGITAL ECJUIPMENT

CORPORATION

DIGITAL EQUIPMENT

CORfORATION

DICfTAL EQUIPMENT

CORIWRATION

KOKOMO liOKOMO

KOKOMO

HUDSON

HUDSON

HUDSON

IN

IN

IN

MA

MA

MA

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

F A B l

FAB 2

FAB 3

FAB 4

FAB 6

PILOT

Products Produced

MMJC PET DIS

MILITARY AEROSPACE

NA

D I S C R E T E ; MICROWAVE

TRANSISTORS

HYBRID A N A 1 0 G

DISCRETE

MIXSIG ASIC CUSTOM

MPU MCU MPR NVMEM

A / D D / A

1-"ET AMI'

DISCRETE

ASIC

DISCRETE D I Q B S - S ^ r o t

DIODE

ARRAYS

ARRAYS

CAP REG DIODE HYBWp

SRAM

SRAM EPROM FPGA

SRAM EEPROM EPROM

Est.

Minimum Wafer

Process Geometry Diameter

Technology (Microns) (In.)

GaAs

NA

GaAs

BIPOLAR

BIPOLAR

BIPOLAR

BiCMOS M 2

POLY2 CMOS

M 2 POLY2

BIPMOS

GaAs

BIPOLAR

BiCMOS

CMOS BiCMOS

Mini Displays

BIPOLAR

BiCMOS CMOS

METAL

BIP CMOS

CMOS

CMOS

CMOS BiCMOS

SRAM LOGIC MPU MPR

SRAM RANDOM LOGIC

BiCMOS CMOS

CMOS

SRAMs RANDOM LOGIC CMOS

M M I M BIP

MICKOCOMPONENTS M 0 5

MPR CUSIDM POWER ICs

CMOS BiCMOS

MOSBIP

SENSORS DISCRETE BIPOLAR

1.2

0.5

3

3

1

3

1.5

0.8

3

0.8

1.5

3

0.65

0.5

0.65

0.5

0.8

0.4

0.5

6

3

6

4

4

4

6

4

3

6

5

3

8

4

4

6

8

6

6

6

6

4

4

2.5 5 ASIC ANALCK POWER ICs BiCMOS

NMOS/PMOS

BIPOLAR

ASIC MPU LINEAR POWER CMOS

BIPOLAR

MCU ALPHA NA

MPU AU'HA

CMOS

1

0 5

0.12

5

6

MPU MCU M F S . ^ B K

CUSIDM

CMOS 1.5

8

6

o

CD

3 or cn

CO

CO cn

@

cn

Si

03

J 3

Crt

CO

1

CJl o

0 0

• 3 o

§.

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

DIONICS INC.

1-C[ SEMICONDUCTOR

EGiiG R E n C O N

EG&C VACTEC

ELANTEC tXEL

City or

District

WESTBURY

SANTA CLARA

SUNNYVALE

ST. LOUIS

MILPTTAS

SANIOSE

FLUKE c o w .

FOXBORO ICT

FREQUENCY SOURCES

Fujrreu

GC CORI'RATl R&D

GE ELECT. LAB

GENNUMCORR

EVERETT

SAN fOSE

LOWELL

GRESHAM

SCHENECTADY

SYRACUSE

BURLINGTON,

ONTARIO

GERMANIUM POWER DEVICES AN DOVER

HANSCORN AFB LEXINGTON

HARRIS SEMICONDUCTOR

IIASBIS SEMICOI^UCTOR

HARRIS SEMiCONDUCTOR

HAKRiS SEMtCONDUCTDR

HARRIS SEMICONDUCTOR

HARRIS SEMICONDUCTOR

HARRIS SEMICONDUCTOR

HARRIS SEMICONDUCTOR

FlBWLFTT-rACK A RD

HEWLfHT-PACKARD

1 lEwi irrr- PACKA RD

11 (: W LIHT-PACKARD i IE WtETT-PAC KA RD

HEWLETT-PACKARD

HEWLETT-PACKARD

HEWLETT-PACKARD

FINDLAY

FINDLAY

FINDLAY

MELBOURNE

MELBOURNE

MELBOURNE

MELBOURNE

MOUNTAINTOP

CORVALLIS

CORVALLIS

TORT COLLINS

PALO ALTO

PALO ALTO

SAN JOSE

SAN JOSE

SAN JOSE

CA

CA

CA

CA

CA

!=L

FL

FL

PA

OR

OR

CO

O H

O H

U S A .

U S A .

O H

FL

U S A ,

U S . A.

U S . A.

U S A

U.S. A.

U S A .

U S A .

U S A ,

U S A .

U.S.A.

U S A .

U S A .

U S A .

U S A .

State

Country Fab Name

NY U S A . NA

CA U S A . NA

CA U . S A 4 INCH FAB

M O

CA

CA

WA

CA

OR

NY

NY

CN

U S A

U S A .

U S . A

U S A .

U S A .

U S A ,

U.S.A.

U S A .

U S A

U S A .

MA

MA

U S A .

U S A .

NA

DIFAB

NA

N A

NA

NA

No. 1

PSF

MMIC FAB

LANDMARK

NA

NA

FAB 1 & 2

FAB 3 & 4

FABS

54E (FAB A)

FAB 54W (FAB C)

FAB 59(VHSIC)

FABD

FAB 6

4-INCH

6-INCH

FTC

HSDL

NA

BIPOLAR

DIODE

OED

Piuduct$ P f o d u i ^

P W R T R A N O n O

Process

Technology

BIPOLAR

ANALOG ASIC DBCRirre

OPTP POWER

ANALOG DRIVERS DPTO

CCD

PDIODE PTRAN

BIPOLAR

CMOS

CMOS N M O S /

PMOS

BIP

ANALOG MONOLITHIC ICs BIPOLAR

64K EEPBOM FLD SRAM

MCU

CMOS BIP

ASIC

DIS PRESSURE SENSORS

DISCRETE ASIC

A S I C I M W ^ M b D R A M

SMART r W R

MMIC

TKAN

DIODE ono

CMOS BiCMOS

BIP

CMOS

CMOS

BiCMOS BIP

GaAs

BIPOLAR

ANALOG AMPS FILTSBS

DRIVERS orroDK

CUSTOM MIL STD

MPU MCU MPR ASIC

MPR ANALOG MIXED

SIGNAL POWER DRWERS

CMOS BiCMOS

BiCMOS MOS

BIPOLAR

MCR MCU DSP POWER ICs CMOS BiCMOS

B I P M I C M O S M P U L I N

CUSTOM ASIC

MPU MPR ANALOG ASrC

MTU ASIC DSPs BRAM

LINEAR

DISCRETE POWER

CBIC

ASIC MPR DSP RISC

MPU ASiC

ASIC OPTO

NA

NA

BIP CMOS

MOS

BIP CMOS M2

BiCMOS M2

CMOS

BIPOLAR

CMOS BiCMOS

BIP MOS

BiCMOS

CMOS

CMOS

CMOS

CMOS

BIPOLAR

GaAs

C M O S M 3

BiCMOS

GaAs

BIPOLAR

GaAs

Est.

Minimum

Geometry

(Microns)

Wafer

Diameter

(In.) Pr

6

2

1.3

1.8

3

1

0.8

2

O.S

1.5

1.2

1.2

1.2

3

08

3

1.5

1

0.8

0.5

0.5

3

5

2

2

3

6

4

6

4

6

6

2

6

CD

O

CD

cr

CD

N) tn

CD

CO t n

@

« i .

C D

C D

C71

O

X 3 c:

CD

O

O

• 3 o

R

CO

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

CO

CO en

s

Products Produced

Est.

Mtnimum Wafer

Process Geometry Diameter

Technology (Microns) (In.) P

DISCRETE POWER GaAs 0.15

LOGIC ASIC BIPOLAR 1.5

ANALOG DISCRETE GaAs BIPOLAR 0.25

ANALOG GaAs 0.25

1Mb 4Mb DRAM 256K SRAM CMOS 0.6

MPU

O P AMP EEPROM LOG

AEJC

C M O S M O S 2.5

OPTOELECTEONTCS

GaAs 10

ANALOG OFTO

BIPOLAR 3 4

0.65 4

SRAM ASIC ANA LJOG:

DISCRETE

MIL STD OPTO ASIC LIN

ASIC LIN

MMIC

ARRAYS CUSTOM MIX

SIGNAL ASIC

^Mb DRAM, MPU

*Mb DRAM

]6Ub DRAM e4Mb 256Mb

ux:

CMOS

BIPOLAR

BiCMOS CMOS

MOS

CMOS

GaAs

CMOS

CMOS

CMOS

CMOS

CMOS

MCU MPU D ^ " MPR

FAST lAK MK 256K 5RAM

SRAM MPU RLSC LOGIC

ASJC ANALOG MPR

3.3 VOLT SRAM

MPU (P6, P7)

MCU

486 MPU

MCU

MPU (P5, DX4)

MPU (P5, P6)

FLASH MEMORY

CMOS,

BiCMOS

BIPOLAR

CMOS

CMOS

BiCMOS CMOS

CMOS BiCMOS

NMOS/PMOS

CMOS BiCMOS

BiCMOS

C M O S M O S

CMOS BiCMOS

CMOS BiCMOS

BiCMOS

BiCMOS

CMOS

1.5

0.5

1

0 8

0 8

0 8

0.4

1.5

0.6

1

0.6

0.4

0.6

3

0.25

0.6

0.5

0.8

0.5

0.35

0.8

4

4

3

4

8

8

6

6

8

4

6

6

5

8

8

8

4

6

6

5

ct>

o

CD

3 en

CO

CO t n

Crt

(75

1

en

O

OO r j

D

O

i

@

CO

CO

tn a

K.

.a

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

INTEL

City or

District

RIO RANCHO iNTin,

It^TTJlNATlONAE.

MICROCIRCUTTS [NC,

INTERNATIONAL RECTlFmR

INTERNATIONAL RECTIFIER

INTERNATIONAL RECTIFIER

STT

KODAK

SANTA CLARA

MILPITAS

EL SEGUNDO

TEMECULA

TEMECULA

ROANOKE

ROCHESTER

KULITE

LAWRENCE LIVERMORE LABS

LINEAR TECHNOLOGY

LEONIA

LIVERMORE

MILPTTAS

LINEAR TliCHNOLOGY

LINFINTTV

MICROEi.ECTROMCS I N C

LITTON MICROWAVE

LITTON SOLID STATE

LOCKHEIl

LOCKEED-MARTIN

LOCKHEED

LOCKHEED

LORAL ELECTRONICS

LSI LOGIC

LSI LOGIC

1^1 LOGIC

M/A-COM

M/A-COM

M i l j r r A S

CA

GARDEN GROiW;

CA

S A N i O S E

SANTA CLARA

SYRACUSE

SANDERS NASHUA

FORT WORTH

SUNNYVALE

TX

CA

LOWELL

FREMONT

MILPITAS

SANTA CLARA

BURLINGTON

BURIJNGTON

M/A-COM

MAGNOVOX

MASS. MICROELECTRONICS

CEtSTTER (MJC)

MATSUSHITA

MAXIM INTECiRATED

PRODUCTS

LOWELL

I^ORT WAYNE

WESTBOROUGH

FUYALLUP

SUNNYVALE

CA

CA

MA

C A

CA

CA

MA

MA

MA

IN

MA

WA

CA

U.S.A.

U.S.A.

U.S.A,

U S A .

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

CA

CA

CA

VA

NY

NJ

CA

CA

State Country Fab Name

NM U.S. A. FAB 9

CA

CA

U S A .

U.SA.

U.S.A.

U.SA.

U.SA.

U.SA.

U.&A

U.S.A.

U.&A.

U.&A.

D2

NA

PPD4

HEXFET

HEXFET-2

DARPA

NA

NA

NA

FABl

FAB 2

NA

NA

NA

E L A B

SANDERS NASHUA

NA

113

S / C

FREMONT FAB

MTLPITAS FAB

R&D PILOT

NA

NA

ADV S/C

NA

ICFF

PUYALLUP

NA

Products Produced

MPU (P5 i960 iS6) MCU

MPR

M r U (PS. n)

ASIC

FWR TRANS M05FET SCR

FWR TRANS MOSFET

CMOSMOS

CMOSMOS

PWR TRANS MOSFET

I J N P W R I C D I S M I L S T D

IMAGING ARRAYS CBIC

CCD

DISCRETE

WAFER-SCALE COMPUTER

UNEAR INTERFACE A / D

D/A

LINEAR

CMOSMOS

GaAs

BIPCMOS

MOS

BIP

C M O S M O S

BIPCMOS

BiCMOS

BIPCMOS

BiCMOS

UNEAR ANALCG

FBTAMP

BIPOLAR

CMOS

GaAs

MMIC CCD

ANALOG DISCRETE

PHEMTs

ANALOG MMICs

NA

ASIC MIL STD RAD-HARD

ASIC

ARRAYS CUSTOM

ASIC MPU

GaAs

MOS

CMOS

RF i MICROWAVE DIOE«S BIP

CMOS

CMOS

CMOS

MMIC DiODE I R A N

DISCRETE

MMIC

ARRAYS CMG HYBRID

ASIC

GaAs

GaAs

GaAs MOS

COMPOUND

GaAs

GaAs

CMOS

MOS BiCMOS

M2

1Mb/4Mb DRAM 4blt Stint

MCU

OP AMPS A / D D/A

CMOS

Process

Technology

CMOS BiCMOS

Est.

Minimum

Geometry

(Microns)

Wafer

Diameter

(In.) P

BiCMOS

CMOS

0,6

1

5

5

2

0.25

1.5

3

0.25

3

8

5

4

5

6

3

4

4

6

4

2

2,5

0.5

0.1

0.15

0.8

1.5

1.5

0.5

0.5

0.3

0.35

0.25

5

2

0.8

3

4

4

3

4

5

3

6

6

6

3

4

6

4

3

3

5

3

3

3

O

CD

O

CO

3

cr

CD

K> t n g

@ en

O

j u

CO

cb en

C3

CO

I

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

MCDONrjELL DOUGLAS

MCDONNELt. DOUGLAS

MRDTCWIC/MICRO-REL

M C R O POWER SYSTEMS

M C R O QUALrrV S / C

M C R O SEMt

M i c R o - c E R C u r r E N G

MICROCf nP TECHNOLOGY

MICROCHIP T F X : H N O L O G Y

MICRON TECHNOLOGV

MICRON TECHNOLOG V

MICRON TECHNOLOGY

MICROPAC INDUSTRIES

MICROSEMI CORP

MICROWAVE TECH.

MFTEL SEMICONDUCTOR

MITSUBISHI

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

City or

Dislrict

HUNTINGTON

BEACH

HUNTINGTON

BEACH

TEMPE

Stale

CA

CA

AZ

CA SANTA CLARA

GARLAND

TORRANCE

TX

CA

WEST PALM BEACH PL

CHANDLER

TEMPE

BOISE

AZ

AZ

ID

Country

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A

U S A

U.S.A

U S A

U.S.A.

U S A

Fab Name

3PIL0T

DVLPMNT

NA

NA

NA

N A

NA

CHANDLER 1

T l

FABl

BOISE

BOISE

FAB 2

FAB 3

GARLAND

BROOMFIELD

TX

C O

FREMONT

BROMQNT, QUEBEC

CA

CN

MORTH DURHAM

GUADALAJARA

AlZU

AUSTIN

AUSTIN

AUSTIN

AUSTIN

AUSTIN

CHANDLER

NC

TX

TX

TX

TX

TX

AZ

U.S.A.

U.S.A.

U.S.A.

CANADA

NA

NA

NA

FABl

U S A .

MEXICO

U.S.A.

U.S.A.

U.S.A.

NA

GUAD POWER

M O S 7

APRDL

MOS 13/14

U.S.A.

U.S.A.

U.S.A.

U.S.A.

M 0 S 2

MOS 3

M O S S

MOS 12

IRVINE

MESA

MESA

ME.SA

CA

AZ

AZ

AZ

U.S.A.

U.S.A.

U.S.A.

U.S.A.

MOS 10

BIPOLAR 1

BIPOLAR 2

BIPOLAR 3

Products Produced

4Kl(iK SRAM 6K ARRAY

MPU

MPU LOG ASIC DIS

Process

Technology

BIP CUSTOM MOS CUSTOM

ASIC MLXSiG ASiC

BIP M l CMOS

M2 POLYl

IJN CUSTOM BiCMOS CMOS

DIP

RECTIFIER MULTIPLffiE

MILSTDDIS

CUSTOM

MCU EF.PROM

BIP

BIP

MOS

MCU EEPROM

256K DRAM 1Mb EaiAM

2S6KSRAM

CMOS

CMOS

CMOS MOS

CMOS 1Mb DRAM i56K SRAM

VRAM

1Mb DRAM 4Mb DRAM

16Mb DRAM

CMOS

MILSTDOFTO HYBRID

S C H O m C Y DIODE

R E C i l l L R

MMIC AMP mx

TELECOM A / D D / A

NA

MOS

GaAs

CMOS NMOS

PMOS l M b / 4 M b DRAM

THYKISTOR

DISCRETE LOGIC ANMiBG CMOS MOS

MPU ASIC MI;M

CMOS MOS

MPU MCU RISC

CMOS BiCMOS

HCMOS

LOG A / D M C U

MCU

CMOS MOS

CMOS MOS

MCU FSRAM DSP L!i* s i s c : CMOS

MCU MPU

CMOS BiCMOS

HCMOS

DSP MUAD CMOS HCMOS

BiCMOS

TELECOM O P AMP

AUTOMOTIVE ANALOG tXJGIC

ANALOG GATE ARRAYS

BIP

BI1> BiCMOS

Est.

Minimum

Geometry

(Microns)

Wafer

Diameter

(In.) P

12

4

0.9

0.8

1.2

«.5

0.5

0.5

0.8

1.8

1

0.25

1

1.2

0.65

0.65

0S5

.1

2.5

1.25

4

4

2

4

6

3

4

5

8

4

4

5

8

6

4

4

4

CO

I

CD

CJl

a

CO

@

.a

CO

CO tn

O

•3 o

i

o

CO o

CD

3

cr

CD

— 1

K)

CD

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

MCrrOROLA

MOTOROLA

N-CHIF

NATIONAL S / C

NATIONAL S/C

NATIONAL S/C

NATIONAL S / C

NATIONAL S / C

NAHONAL S/C

NATIONAL 5 / C

NAHONAL S / C

Cityot

District

MESA

MESA

OAKHILL

PHOENIX

PHOENIX

PHOENIX

PHOENIX

PHOENIX

AZ

AZ

U.S.A.

U.S.A

PHOENIX

PHOENIX

RESEARCH TRIAN-

GLE PARK

TEMPE

SAN JOSE

ARLINGTON

AZ

AZ

N C

AZ

CA

TX

U S . A.

U S A .

US-A.

U S . A.

U S A .

U S A .

ARLINGTON

SANTA CLARA

SANTA CLARA

TX

CA

CA

U.S. A,

U.S. A.

U S A .

SANTA CLARA

SOUTH PORTLAND

CA

ME

U S A .

U S A .

SOUTH PORTLAND

WEST JORDAN

ME

UT

U S A .

U S , A,

State Country Fab Name

AZ

AZ

TX

AZ

U.S.A.

U.S.A.

U S A .

U.S.A.

U.S.A.

U.S.A.

M O S S

M O S 6

M O S l l

C O M l

OPTO

BIPOLAR 5

DFS

M 0 S 4

P H O S J I X POWER

ZENER/RECnFIER

MOS15

CS-1 PHASE 1

NA

C M O S l

FAB 2

ANALOG 5

A T G 6

ATG8

BIPOLAR

CMOS

M 0 S 3

NATL. SECURITY ADMIN,

NATL. SECURITY ADMIN.

NAVAL OCEAN SYS. CTR,

NEC ELECTRONICS

NEC r . L r c T R O N i c s

NORTHS (IN lELECOM

NOVASFJ^SOR

OITEK TECHNOLOGY. INC.

OITEK TECHNOLOGY. INC

OPTO DIODE

FORT MEADE MD U.S. A. NA

FORT MEADE

SAN DIEGO

ROSEVILLE

MD

CA

CA

U S A .

U S . A.

U S A .

NA

NA

K-LINE

ROSEVILLE

NEPEAN, ONTARIO

FREMONT

CA

C N

CA

U.S.A.

CANADA

M-LINE

M 0 D 4

U.S.A.

NA

CARROLLTON

CARROLLTON

NEWBURY PARK

TX U S A . TAB 1

TX U S A . FAB 2

CA U S A . NA

Est.

Mimmum Wafer

Process Geometry Diameter

Technology (Microns) (In.) P

Products Produced

MCU

ASIC ANALOG rXIG

CMOSMOS

CMOS BiCMOS

SRAM DSP MCU MTU RISC CMOS BiCMOS

RP POWFJi

OPTO

LDMOS

GaAs

BIP

0.8

0.35

0.35

1.25

5

6

8

6

2

5

RPl'WR SMALL SIGNAL orro

RECTIFIERS

3

MOSFET SMART PWR

DISCRETE

BIP 3

6

PWRTRAN

ZENER DIODE R E C n i m

8 Bit MCU LOGIC

BIP

BIP

CMOS

10

10

0.8

5

4

6

ANALOG RF GaAs

MCM

ARRAYS MCU EBPROM

MPRCOM

ARRAYS MCU

ANALOG

NONVOLATILE MEM MPR

MCU MFU DSP ASIC

MPR MCU MI'K DSP ASIC

LOG

CMOSMOS

CMOSM2

POLYI

CMOS

BIPOLAR

CMOS BiCMOS

BIPOLAR

CMOS BiCMOS

BIP CMOS

BiCMOS

LOGASatAY

EPROM EEPROM

HvIBEDDED CONTROLLERS

CMOSMOS

ANALOG DISCRETE

CUSTOM MIL STD

CMOS BiCMOS

BIP CMOS

MOS

M n . S T D

NA

CMOS

NA

NMOS CMOS 156K SRAM 256K DRAM

AS!C MCU

4Mb/ 16Mb DRAM

CBIC CUSTOM

Si-BASED PRESSURE

SENSORS

MIL ETD PWR ICs

MIL STD PWR ICs

OPTO DIODE

CMOS

MOS

CMOS

MOS

MOS

GaAs

0.35

3

1.2

0.65

2.5

0.65

0.35

2.5

1

0.6

1

0.3

1

0.5

2

4

S

6

4

5

6

5

6

6

6

4

4

5

6

8

4

5

6

O

CD

O

CD

3 cr

CD tn

CO

CO

CJI

@

CO

CO

CJI

CU

j Q

CO

O )

I

C D

0 1

C D

CO

O

O

I

CD

O .

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

City or

District

ORBIT SEMICONDUCTOR INC SUNNYVALE

ORBIT SEMICONDUCTOR INC SUIiftvlYVALE

PARADIGM TECHNOLOGy INC, SAN JOSE

State Country Fab Name

C A

U S A .

FABl

CA US.A. HAH 1

CA U.S.A, SRAM FAD

PERFORMANCE S/C

PHILIPS SENCCON DUCIORS

PI IIUPS SENOaJN DUCTORS

SUNNYVALE

ALBUQUERQUE

ALBUQUERQUE

CA

NM

U S A

U S A

NM U S A

PAB3

FAB 21

FAB 23

PHILIPS/SIGNEI3CS

POWEREX

POWEREX

SUNNYVALE

YOUNGWOOD

YOUNGWOOD

PRECISION MONO.

PRECISION MONO.

PROTECT RfiVICES

RAMTRON INTERNAtJgSAL

CORP.

SANTA CLARA

S/INTA CLARA

TEMPE

COLORADO

SPRINGS

CA

PA

CA

CA

AZ

CO

U S A .

U S A ,

U S A .

U S A .

U S A ,

US.A.

U S A .

RAYTHEON

RAYTHEON

RAYTHEON

RAYTHEON

ANDOVBR

ANDOVKR

MOUNTAIN VIEW

MOUNTAIN VIEW

RAYTHEON

ROCKWELL

ROCKWELL

ROCKWELL

WALTHAM

NEWBURY PARK

NiWl'OKT BEACH

NEWPORT BEACH

ROCKWELL

K0CKWELL{FORMERLY UTMC)

NEWPORT BEACH

COLORADO

SPRINGS

ROHM

SAMSUNG MICROWAVE

SANDERS ASSOCIATES

SANDIA NAnONAL LABS

SANTA BARBARA RSCH.

SUNNYVALE

MBJPrTAS

NASHUA

ALBUQUERQUE

GOLETA

MA

MA

CA

CA

MA

CA

CA

CA

CA

CO

CA

CA

N H

NM

CA

U S A

U S A .

U S A

U S A ,

U S A .

U S A ,

U S A .

U S A ,

U S A ,

U S A .

U S A

U S A .

U S A .

U S A .

U.S.A.

SEMATECH AUSTIN

FABl

HPWR SCR/DIODE

PLANAR TRANSJSn

FABl

FAB 2

NA

NOKTHG*afS>

GaAs

NA

LINEAR

LSIARAY

NA

GaAs WAFER FAB

FABl

FAB 4

Fab V

UTMC

KIFER PLANT

GaAs

GaAs

RHIC-I

SBRC

Products Produced

ARRAYS CUSTOM MKSIG

ASIC ASIC

FOUNDRY

256K, 1Mb, 4Mb SRAM

SRAM /IRRAYS MIPS RISC

MPU

Process

Technology

CMOS

CMOS

CMOS

CMOS

SRAM MPU ASIC

EPROM MCU MPR DSP

ANALOG

BiCMOS CMOS

CMOS BiCMOS

MOS BIPOLAR

CMOS BiCMOS MCU MPR ASIC ANALOG

DISCRETE OPTO POWER

ICs

LIN SMART PWR OP AMPS

DISCRETE POWER DIODES

SCRCTll

DISCRETE DIODE PWR

TRAN THYRISTOR

BIP

BIPOLAR

BIPOLAR

ASIC

CUSTOM

CMOS

BIP

DIODE

4K SERIAL/PARAMJi..*0:

PARALLEL NVMEM

(FERRAM)

BIP

CMOS

Est.

Minimum

Geometry

(Microns)

L2

Wafer

Diameter

(In.) P

4

0.8

0.55

1

0.7

1,2

0.8

2.5

50

5

3

2.5

25

1

6

6

6

6

4

6

4

3

4

4

4

3

6

NA

ARRAYS CUSTOM

LIN ASIC DIS TRAN SST

JDK ARRAYS

MMIC

MOSITTT HBT MEM ASIC

MODEM TELECOM

TELECOM PMTERFACE

GaAs

CMOS

BIP

BIP CMOS

GaAs

GaAs

CMOS MOS

CMOS MOS

TELECOM CHIPS ANALOG

TELECOM CHIPS ANALOG

CMOS GaAs

BIP CMOS

0.5

0.9

5

1

0.7

2

2

0.5

0.8

3

4

4

4

4

4

3

5

8

5

CUSTOM MisEd-Signal ASIC

GsAi FET MMIC

LIN MMIC

MPU LOG SRAM ASIC

MIL STU E\IFRARED

DETECTOR

BIPOLAR

GaAs

GaAs

CMOS

HgCdTe InSb

3

0.25

0.5

1.5

5

4

3

4

4

8

o

•^

CD

O .

@

CO

CO

CJl

a

55"

. o

c/>

CO

CO

1

CO

CJl

C 3

CO

O

CD

O

CD

3

C T

CO

—1

^^

CJl

CD

CD cn

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

SEMrCOA

SEMTECH CORPORATION

SEMTECf i CORPORATION

City or

District

COSTA MESA

CORPUS CHRIST!

NEWBURY PARK

QUAKERTOWN

MlLPfTAS

Stale

CA

Counlor

U S A

Fab Name

NA

TX

CA

PA

U.S.A.

U.S.A.

CORPUS CHRISTI

NEWBURY PARK

CA

U.S.A.

U.S.A.

NA

PAB I

SENSOR SOLLD STATE

SENSVM [NC

SGS-THOMSON

SGS-THOMSON

BGS-THOMSON

SGS-THOMSON

SGS-THOMSON s i n MICROELECT.

CARROLLTON

CARROLLTON

PmLADEIPIflA

PHOENIX

SAN DIEGO

CONTAGEM

TX

TX

PA

AZ

CA

U.S.A. FAB 4

U.S.A.

U.S.A.

U.S.A.

U.S.A.

BRAZIL

PAB 6

NA

PHOENIX n u i

NA

NA

Sro MICROELECT.

SmWENS

SILICON SVSTEMS (TDK)

CONTAGEM

CUPERTINO

SANTA CKUZ

CA

CA

BRAZIL

U.S.A.

U.S.A.

Sn.lCON TRANSISTOR

SILICONIX rK'COKPORATED

SlUCONIX IMCORPORATED

SIPEX CORPORATION

CHELMSFORD

SANTA CLARA

SANTA CLARA

MtlPITAS

MA

CA

CA

CA

U.S.A.

U.S.A.

U.S.A.

U.S.A.

SOUD POWER CO.

SOLID STATE DEVICES

SOLTTRON DEVICES

SONV

SONY

SPECTRA DIODE LABS

SPECTRO LARS (fTLPGHES)

SPRAGUE (ALLEGRO)

SPRAGUE (ALLEGRO)

STAMIARD MICROSYSTEMS

SUrERTIIX

SYMniOS LOGIC (Toimeriy

Hyundsi (NCR])

SVMBIOS LOGIC (FDiJnetly

Hyundai (NCR))

FARMINGDALE

LA MIRADA

WEST PALM BEACH

SAN ANTONIO

SAN ANTDNIO

SAN JOSE

SYLMAR

WILLOW GROVE

WORCESTER

NY

CA

PL

TX

TX

CA

CA

PA

MA

HAUPPAUGE

SUNNYVALE

COLORADO

SPRINGS

COLORADO

SPRINGS

NY

CA

CO

CO

U S A ,

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U.S.A.

NA

NA

F A B m

BIPOLAR

FAB 2

FAB 3

FABl

NA

NA

NA

Fab 11

Fab 12

SDL

NA

NA

NA

NA

NA

COLORADO SPRINGS 4

COLORADO SPRINGS 8

Products Produced

CUSTOM HI-REL PWR

PHOTO

ANALOG DISCRETE

DISCRETE RECTIFIER

ZENER DIODE

CUSrOM DIS SENSORS

Process

Technology

BIP

BIPOLAR

BIPOLAR

SOLID STATE SENSORS

ANALOG

SRAM ASIC e-ROM

EEPROM

1Mb SRAM ARRAYS

DIS, RF

MPUSRAM

ASIC LOG

L I N P W R T R A N S S T P W R

ECs

PWRICs

LED COUPLERS OPTO

MIXED SIGNAL ASIC

MOS N M O S /

PMOS

BIPOLAR

CMOS

CMOS

BIP

CMOS

CMOS BiCMOS

BIP

DISCRETE POWER

SMART PWR A / D D / A r W R SMART PWK

ASIC MIXED SIGNAL

PWRTRAN

HI-REL CUSTOM

DISCRETE

SRAM

ASIC PLD

LASER DIODE

SOlAECELLARtiAYSt

SRAM ROM P R O a :

CUSTOM DIS

CUSTOM CBIC

ANALOG POWER

MPR MIXED SIGNAL A ^

CMOS

GaAs MOS

CMOS BiCMOS

BIPOLAR

BIPOLAR

CMOS

CMOS

BIPOLAR

BiCMOS

BIP

BIP

BIPOLAR

CMOS

BIPOLAR

GaAs GaAIAs

NA

CMOS

BiCMOS DIP

CMOS MOS

CMOS MOS

CMOS

Est.

Minimum

Geometry

(Microns)

7

Wafer

Diameter

(In.)

P

3

5

3

3

12

0.7

5

0.5

2

30

a

0.8

4

3

1.5

3

20

2

0.5

1.25

2.2

4

1.25

4

0.7

4

2

4

4

4

6

4

8

4

3

4

6

4

4

6

4

6

3

4

4

4

2

4

3

6

4

4

4

MPR ASIC ANALOG CMOS

03

8

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

O)

I

(£>

cn

O

CO

@ to

CO cn

O

K.

J3

T3

O

I

CD

Q.

O

CD

O

Company

SYMBIOS LOtMC (Formerly ttyundsi |NCR))

SYMBtOS LOGIC (Formerly

Hyundai (NCR))

SYNERGY SHIvlICONDUCrOR

City or

District

FORT COLLINS

FORT COLLINS

SANTA CLARA

SYI'EX

TECCOR ELECTRONICS

TtSCCOR ELECTRONICS

TEKTRONIX

MIC ROELECTRONICS

TELCOM DEVICES

Ttil.EDYNE MICROELECT.

TELntJYNE MtCROWAVE

TEl.EDYNI; S/C

TEXAS INSTRUMFMTC

TEXAS tNSTRtJMENTS

TEXAS INSITIUMENTS

TEXAS INSTRUMENTS

TEXAS tNSTRtJMENTS

TEXAS INSTRLMEt^rrS

TTiXAS INSTRUMENTS

TEXAS INSTRUMENTS

TRIQUINT

MILPITAS

IRVING

IRVING

BEAVERTON

CAMARILLO

LOS ANGELES

CA

CA

MOUNTAIN VIEW

MOUNTAIN VIEW

CA

CA

DALLAS

DALLAS

DALLAS

DALLAS

DALLAS

HOUSTON

LUBBOCK

SHERMAN

BEAVERTON

TX

TX

TX

TX

TX

TX

TX

TX

OR

CA

TX

TX

OR

State Country Fab Name

CO U S . A. FT. COLLINS (4)

CO U,S.A.

FT. COLLINS (6)

CA

U,S.A. CORVEN FAB

U.S.A.

US.A.

U.S.A.

US.A.

FABl

SIDEATOR

TRIACS STAND

BIPOLAR

U S . A,

U.S.A.

US.A,

U S . A.

U S . A.

U.S. A.

US.A.

U S . A.

US,A.

U S . A.

US.A.

US.A.

U.S.A.

NA

NA

GaAs

NA

DLIN

DLOG

DMOS

DMOS5

SPDL

H-FAB

LMOS

S-FAB

NA

TRW

TRW

TRW

MANHATTAN

BEACH

REDONDO BEACH

REDONDO BEACH

CA

CA

CA

US.A.

U.S. A.

U.S.A.

NA

Dl

Dl

TRW

TRW SYSTEMS uNrraoot;

UNITROOE; (NTEGRATED

C IRCUirs CORTORATION

VITESSE SEMICONDUCTOR

VLSI TECHNOLOGY

REDONtX) BEACH

LA JOLLA

WATERTOWN

MERRIMACK

UNIVERSAL SEMICONDUCTOR SAN JOSE

CAMARILLO

SAN ANTONIO

CA

CA

MA

N H

CA

CA

TX

U.S.A.

U.S.A.

U.S.A.

U.S.A.

U S A .

U S A .

U S A

NA

NA

NA

NA

NA

CAMARILLO FAB

MODULE A to cn

CD

CO cn

Products Produced

ARRAYS CBIC

ARRAYS CBtC

Process

Technology

FAST-SRAM MPR ASIC

OF AMP CBIC CUSTOM

ASiC ANALOG

DISCRETE THYRISTOR

TELECOM A / D ARRAYS

CUSTOM

I^DIODE LED

BIPOLAR

BiCMOS

CMOS

BIPOLAR

BIPOLAR

BIP

GaAs

H!-REL HYBRID i A / D 0/

A nrr

PWRICsPWRMOSFET

HYBRIDS

BIP

GaAs

BIP BiCMOS

CMOS

U N

HM ASSP

LOG MPU

16Mb DRAM

BIP BiCMOS

MOSCMOS

CMOS

CMOS

CMOS

16Mb, 64Mb DRAM ASSP

AD v e i l ' A S S P ASIC

EPROM MPU MCU DSPs

BIP CMOS

CMOS

BIPTTI.

LOG MPR

M M i c L i N o r r o c m c

ARRAYS

U N TRAN PWRTRAN

IIYBRD

GaAs

MOS

CMOSMOS

BIP CMOS

VHSIC MrLSTD FERRAM

U N TRAN PWRTRAN

HYORD

RFPWR

A / D D/A MULnPLIERS

HYBRID DIS

U N SMART PWR CUSTOM

GaAs

CMOS BIP

BIP

DIP

LINEAR ARRAYS ASIC

POWER ICs

GaAs ARRAYS ASIC MPR

ARRAYS CtitC MPU MPR

CMOS BiCMOS

BIPOLAR

GaAs

C M O S M 3

POLYl

Est.

Minimum

Geomeify

(Microns)

Wafer

Diameter

(In.) P

4

1

6

4

I

4

0.24

0.35

1

4

3

3

4

0 6

1

0 6

1

0.7

1

0 8

0 6

0.5

0 5

1.5

0.5

1

4

6

4

4

2

4

3

4

4

6

6

6

8

6

5

6

5

4

3

4

4

4

4

CD

o

CD

3

CD

ro

CJl

CO

CO

en

CO

CO on

CO

@

CO

C D

C71

O

Si

J 3

T3

O

CD

O .

Table 1 (Continued)

Americas Existing Pilot and Production Fab Lines

(Including Fabs Beginning Operation during 1995)

Company

VLSI TECHNOLOGY

VLSI TECHNOLOGY

City or

District

SAN ANTONIO

SAN ANTONIO

State Counliy Fab Name

TX U S A . MODULE B

TX U S A . MODULE C

VLSI TECHNOLOGY

WESTINGHOUSE

WESTINGHOUSE

SAN JOSE

BALTIMORE

BALTIMORE

CA

MD

MD

U S A .

U S A .

U S . A.

CA U S A .

SAN JOSE

GaAs

SILICON

NA

XEROX PALO ALTO RSCH. PALO ALTO

XICOR

ZENITH MICROCIRC

ZILOG

ZILOG

MILPITAS

ELK GROVE

NAMPA

NAMPA

CA

IL

ID

ID

U.S.A.

US.A.

U.S.A.

US.A.

PHASE 2

HVSR

MODULE 1

MODULE 2

ZILOG NAMPA I D

U S A .

NA = Not applicable

Fab Types:

F = Production-Based Fab

R = Semiconductor R&D and/or Trial ProducSon Facility

P = Pilot Line (Initial Production or Intended Low Volume)

T = Test and Assembly (Formerly A)

Q = Quick-Turn Fab

N = Nondedicated Foundry Service Available

D = Design Center

Source: Dalaquest (December 1995)

MODULE 3

Products Produced

ARRAYS CBIC MPU MPR

ARRAYS CBIC SRAM MPU

E2

Est.

Minimum Wafer

Process Geometry Diameter

Technology (Microns) (In.) P

C M O S M 3

C M O S M 3

PLD ARRAYS CBIC MPR

POWER MMICs

CMOSMOS

GaAs

ASIC DISCRETE OFTD R A »

HARD MEM

CMOS BiCMOS

BIPOLAR

CUSTOM IMAGE

PROCESSING

EEPROM

SOI

CMOS N M O S /

PMOS

HIGH-VOLT DIODE TRIODE BIP

Ml'U MCU CUSTOM

ZB0,OOO MPU MCU

CUSTOM

NMOS

CMOS NMOS

BiCMOS

MPU MCU CUSTOM

0.8

0.8

0.8

0.2

0.6

0.5

20

1.2

0.65

0.6

6

6

5

4

6

4

6

2

5

5

8

w

CO

CJI

o

00

o

CD o

CO

3 rs3

CJl

s

@

i

I

o .

=] o

• 3 o

I

Table 2

Americas Future Pilot and Production Fab Lines (Including Fabs Beginning Operation durin

Company

AMD

AMERICAN MICROSYSTEMS

INC.

CVTRESS SEMICONDUCTOR

DIGITAL EQLr[I*MENT

CORFORATtON

FUjrreu

HARRIS SEMICONDUCTOR

City or

District

AUSTIN

POCATELLO

BLCOMINGTON

HUDSON

GRESHAM

MOUNTAINTOP

State Country

Fab Name

TX U.S.A. FAB 25

ID U.S.A. FAB 2

MN

MA

OR

PA

U.S.A.

U.S.A.

U.S.A.

U.S.A.

FAB 4

FAB 6

No 2-2

POWER MOS i

Products Produced

Process

Technology

ARRAYS PLDCBIC CUSTOM CMOS

MDCSIG ASIC

SRAM EPROM FPGA

MPU ALPHA

CMOS

CMOS

Est

Minimum Wafer

Geometry Diameter

(Microns) <In.)

05

0.5

0.12

t)S2

HITACIII

HYUNDAI

IBM/TOSHIBA

IDT

INTEL

INTTTl.

INTT-L

INTEL

INTERNATTONAL RECriFIER

LINEAR TECHNOLOGY

LSI LOGIC

MEOTONIC/ MICRO-REL

MICRON TECHNOLOGY

MOTOROLA

MOTOROLA

IRVING

EUGENE

MANASSAS

MH.LSBORO

ALOHA

CHANDLER

Hn.I5BORO

ElO RANCHO

TltMECUI.A

CAMAS

GRESHAM

TEMPE

LF.H!

AL^TTN

MESA

TX

OR

VA

OR

OR

AZ

OR

NM

CA

WA

OR

AZ

UT

TX

U.S.A.

U.S.A.

U S A .

U.SA.

U.SA.

U.S.A.

U.S.A.

U.S.A.

U S A .

U.SA.

U.S.A.

U S A .

U.S.A.

U.S.A.

U2

OREGON FAB

NA

N A

D1A2

FAB 12

NA

FAB 11.2

HEXFET-2

FAB 3

FABl

NA

NA

MOS 13/14

A Z U.S.A.

MOS 21

U.S.A.

U S A .

U.S.A.

C O M l

POWER RECT

MOS 19

ISMb 64Mb DRAM

DISCRETE POWER

MCU

16Mb ft4Mh DRAM l6Mb/MMb DRAM D ^

CMOS

CMOS MOS

BIPOLAR

CMOS

CMOS

CMOS

LOGIC SRAM

MTU (F6, FT)

MPU (P6)

LOGIC MPU

MPU (P5, P6)

PWR TRANS MO^BT"

LINEAR

CMOS

BICMOS

BiCMOS

CMOS

BiCMOS

CMOS MOS

BIPCMOS

BICMOS

ASIC CBIC MPU MFR SRAM

NA

16Mb MMb 256Mb DRAM

MPU MCU RISC

CMOS BICMOS

BIP M l CMOS

M2 POLYl

CMOS

CMOS BICMOS

HCMOS

LOGIC ASIC ANALOG

DISCRLTI^

CMOS BICMOS

RF POWER

RECnriEBS

MPU MCU LOGIC

LDMOS

CMOS BiCMOS

035

0.35

035

0.5

0.25

0.4

0.25

0.4

2

0.35

0.8

0.25

0.25

0 8

0 3 5

8

12

MOTOROLA

MOTOROLA

MOTOROLA

MOTOROLA

NATIONAL SEMICONDUCTOR

NATIONAL SEMICONDUCTOR

NATIONAL SEMICONDUCTOR

NATIONAL SEMICONDUCTOR

ORBTT SEMICONDUCTOR INC

ORBIT SEMICONDUCTOR ING

PHOENIX

PHOENIX

RESEARCH

TRIANGLE PARK

WEST CREEK

ARLINGTON

SANTA CLARA

SOUTH PORHAND

SUNNYVALE

SUNNYVALE

NC

VA

TX

CA

ME

CA

CA

ROCKWELL

ROCKWELL

NEWPORT BEACH

NEWPORT BEACH

CA

CA

U.S.A.

U.SA,

U.SA.

U S A ,

U S A .

U.SA,

U S A .

U S A .

FAB 3

ATG8

NA

FABl

FAB 2

FabV

Fab VI

PowerPC MPU

ANALOG LOGIC AaCs

ARRAYS MCU

MPR MCU MPR DSP ASIC t£>G ARRAY

FOUNDRY

ARRAYS CUSTOM MIXSIG

ASIC

TELECOM CHIPS ANALOG

Tin.ECOM CHIPS ANALOG

CMOS

BiCMOS CMOS

CMOS

CMOS BiCMOS

BiCMOS

CMOS

CMOS

CMOS GaAs

CMOS GaAs

0.35

0.35

0.25

0.35

0.25

0.8

0.8

0.5

0.35

o

CD

O

CD

1X3

CJ1

CD

CO t n

CO

CO

I

CO

o oo

@

CO

CO

CJ1

o

5f

X }

Table 2 (Continued)

Americas Future Pilot and Production Fab Lines (Including Fabs Beginning Operation Durin

Company

SAMSUNG

SGS-THC3MSON

SONY

TEXAS INSTTRUMENTS

Tl-XAS INSTRUMENTS

TSMC-)V1

City or

District

AUSTIN

PHOENIX

SAN ANTONIO

DALLAS

DALLAS

NA

TWINSTAK SBMiCONDUCTOR

VLSI TECHNOLOGY

RICHARDSON

SAN ANTONIO

State

TX

AZ

TX

Countiy

U.&A.

U.S.A.

U.S.A.

TX

TX

NA

TX

TX

U.S. A.

U.S.A.

U.S.A.

U.S.A.

U.S.A

Fab NAme

N A

PHOENIX FAB

NA

DMOS5

DMOS 5 PHASE 2

NA

TWINSTAR

MODULE D

Pjoducts Produced

64Mb DRAM

MPUSRAM

HS SRAM ASIC LOGIC

(ASSPs)

16Mb DRAM

64Mb, 256Mb DRAM

LOG CUSTOM MPU MEM

FOUNDRY

16M/64M DRAM

ARRAYS CBIC SRAM MPU

E2

Process

Technology

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOS

CMOSM3

Est.

Minimum

Geometry

(Microns)

Wafer

Diameter

(In.)

0.35

0.5

0.35

8

8

6

0.5

0.35

0.35

0.3

0.6

8

8

8

8

6

NA = Not applicable

Fab Types:

F = Prcjduction-Based Fab

R = Semiconductor R&D and/or Trial Production Facility

P = Pilot Line (Initial Pfoduction or Intended Low Volume)

T = Test and Assembly (Formerly A)

Q = Quick-Turn Fab

N = Nondedicated Foundry Service Available

D = Design Center

Source: Dataquest (December 1995)

1

For More Information...

James Seay, Research Analyst (408) 468-8259

Internet address [email protected]

Via fax (408)954-1780

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our clients. Individual companies reported on and analyzed by Dataquest may be clients of this and/or other Dataquest services, This information is not furnished in coimection with a sale or offer to sell securities or in connection with the solicitation of an offer to

^ ^ J, buy securities. This firm and its parent and/or their officers, stockholders, or members of their families may, from

L/alaQUvSl time to time, have a long or short position in the securities mentioned and may sell or buy such securities.

____ , ©1995 Dataquest Incorporated—Reproduction Prohibited n n acompanyoi -i r r

1 1 0 TheDun&Bradstieet Corporation Dataquest is a registered tradennark of A.C. Nielsen Company

DATAQUEST WORLDWIDE OFFICES

NORTH AMERICA

Worldwide Headquarters

Dataquest Incorporated

251 River Oaks Parkway

San Jose, California 95134-1913

United States

Phone: 1-408-468-8000

Facsimile: 1-408-954-1780

Dataquest Incorporated

Nine Technology Drive

RO. Box 5093

Westborough, Massachusetts 01581-5093

Uruted States

Phone: 1-508-871-5555

Facsimile: 1-508-871-6262

Dataquest Global Events

3990 Westerly Place, Suite 100

Newport Beach, Califoniia 92660

United States

Phone: 1-714-476-9117

Facsimile: 1-714-476-9969

Sales Offices:

Washington, DC (Federal)

New York, NY (Financial)

Dallas, TX

LATIN AMERICA

Research Affiliates and Sales Offices:

Buenos Aires, Argentina

Sao Paulo, Brazil

Santiago, Chile

Mexico City, Mexico

EUROPE

European Headquarters

Dataquest Europe Limited

Holmers Farm Way

High Wycombe, Bucks HP12 4XH

United Kingdom

Phone: +44 1494 422 722

Facsimile: +44 1494 422 742

Dataquest Europe SA

Immeuble Defense Bergferes

345, avenue Georges C16menceau

TSA 40002

92882 - Nanterre CTC Cedex 9

France

Phone: +33 1 41 35 13 00

Facsimile: +33 1 41 35 13 13

Dataquest GmbH

Kronstadter Strasse 9

81677 Miinchen

Germany

Phone: +49 89 93 09 09 0

Facsimile: +49 89 93 03 27 7

Sales Offices:

Brussels, Belgium

Kfar Saba, Israel

Milan, Italy

Randburg, South Africa

Madrid, Spain

JAPAN

Japan Headquarters

Dataquest Japan K.K.

Shinkawa Sanko Building

6th Floor

1-3-17, Shinkawa

Chuo-ku, Tokyo 104

Japan

Phone: 81-3-5566-0411

Facsinule: 81-3-5566-0425

ASIA/PACIFIC

Asia/Pacific Headquarters

7/F China Underwriters Centre

88 Gloucester Road

Wan Chai

Hong Kong

Phone: 852-2824-6168

Facsimile: 852-2824-6138

Dataquest Korea

Suite 2407, Trade Tower

159 Samsung-dong, Kangnam-gu

Seoul 135-729

Korea

Phone: 822-551-1331

Facsimile: 822-551-1330

Dataquest Taiwan

llF-2, No. 188, Section 5

Nan King East Road

Taipei

Taiwan, R.O.C.

Phone: 8862-756-0389

Facsimile: 8862-756-2663

Dataquest Singapore

105 Cecil Street

#06-01/02 The Octagon

Singapore 0106

Phone: 65-227-1213

Facsimile: 65-227-4607

Dataquest Thailand

12/F, Vanissa Building

29 Soi Chidlom

Ploenchit Road

Patumwan, Bangkok 10330

Thailand

Phone: 662-655-0577

Facsimile: 662-655-0576

Research Affiliates and Sales Offices:

Melbourne, Australia

Beijing, China

Dataquest' a company of

The Dun & Bradstreet Corporation

®1995 Dataquest Incorporated

Market

Analysis

S e m i c o n d u c t o r Equipment^ Manufacturings and Materials W o r l d w i d e

Market Analysis

Lithography Stepper Capacity: Bottleneck of

Semiconductor Production

AbStrSCt: semiconductor manufacturings unprecedented and sustained growth of the past

several years has put a significant strain on the infrastructure of the markets that support it. In manufacturing semiconductors, lithography defines capacity and technical capability.

The capacity constraints of the suppliers in this critical lithography stepper market are discussed in this article by looking at each stepper supplier.

By Nitder Pakdaman

Stepper Suppliers Step Up Capacity

Nikon recently announced expansion of its stepper production facility in

Kumagaya, Japan. The company will increase output at its n\ain plant to nearly 1,000 steppers per year from a current capacity of about 650 units.

The primary emphasis will be on the high-end i-line systems (NSR-

2205illD and beyond) and the newly announced scanning stepper

(NSR-S201A deep-UV system). Full capacity wiU be reached gradually through 1996.

At about the same time, Ultratech Stepper and ASM Lithography (ASML) announced that they are holding preliminary discussions to form a strategic alliance. Ultratech and ASML will offer their clients the option of mixing and matching their respective noncritical and critical resolution steppers.

The primary goal of the discussions is to focus on developing faster responses to capacity constraints.

Dataquest Analysis

The news releases on both of these announcements were short but significant. Without steppers, one cannot build a fab and expect to pattern the

Dataquest

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MA-9504

Publication Date: October 23,1995

Filing: Mari<et Analysis

Semiconductor Equipment, IVIanufacturing, and Materials Worldwide circuits on the silicon. Concerns about infrastructure constraints in the semiconductor industry are mounting. Could the current undercapacity in manufacturing curtail the semiconductor industry's goal of surpassing

$300 billion by the end of the century? These concerns are balanced by warnings of historical overinvestment patterns that have taken the industry through some very hard times. There is merit to both of these arguments.

As usual, the reality will probably lie somewhere in between and be more complex than either.

In a recent update of the semiconductor revenue forecast, Dataquest increased its forecast of average growth rates for the industry to over

20 percent until the end of the decade. Growth rates had been just over

15 percent for the first half of the decade. We believe this increase is a fundamental indicator of the growing prevalence of semiconductors in our lives. From set-top boxes to PCs and communications applications, semiconductors are changing all facets of our lives at home and work.

The other fundamental aspect of the industry is the cyclical nature of the markets and investment patterns that have produced cycles of relatively stronger and slower years. As the industry migrates to a new device generation, the cycle begins with investment to address undercapacity. Then, as device shrinks take hold and yield increases, demand is met and a period of overcapacity may occur. This cyclical pattern is seen very clearly in the

DRAM market.

In short, w e are experiencing a fvmdamentaliy stronger demand for semiconductors that the industry is addressing by unprecedented increases in capacity. However, the central cyclical drivers still exist in the market. In

Dataquest's view, the issue of over- and undercapacity will not prevail uniformly in all markets and for all ICs. The Dataquest forecasts for the semiconductor market and the directly related manufacturing equipment and materials market reflect both the growth in demand and the cyclical nature of the market.

Since 1993, megafab announcements have become daily news. Capital spending figures for fabs ranging from $1 billion to $2 biUion are now taken as ordinary events. The technology and capacity of these fabs are defined by the number and capabilities of their lithography steppers.

The lithography tools supply-demand equation has been producing backlogs extending into the first quarter of 1997. This may sound fantastic, but if these backlogs are not addressed in a timely fashion by stepper suppliers, it could spell disaster. As both technology and users' needs evolve, these extended delivery times could translate into changing specifications.

Changes in the market over such an extended period could even spur cancellations. A more balanced supply-demand picture with realistic backlogs would make planning and investment for stepper manufacturers and semiconductor manufacturers a more reasonable and less chaotic exercise.

All of the major stepper manufacturers (ASML, Canon, Nikon, SVG Lithography, and Ultratech Stepper) are facing historic capacity demands. Inability to address the problem could spell loss of market share in a segment that accounts for over 15 percent of the total front-end equipment market, on average. Dataquest estimated the stepper market at $1.8 billion in 1994.

SEMM-WW-MA-9504 ©1995 Dataquest Incorporated October 23,1995

; ^

Semiconductor Equipment, IVIanutacturing, and Materials Worldwide

Current estimates call for this market to reach $5 billion by the end of the decade.

Nikon has led the stepper market with shares about and above 50 percent in the last several years. Nikon's current capacity is approximately 650 stepper vmits a year. For 1995, this capacity should preserve the company's position in a market that we estimate will surpass 1,000 steppers for the first time, with over 1,200 steppers shipped to manufacturers. Nikon plans to sustain and even gain market share by quickly increasing capacity in a market that is highly cycHcal. The peak and trough years of the stepper market are very steep. But w e estimate that the growing demand in production will not be met by capacity until 1997, at least, and that the slowdowns that follow will not be as severe as those of previous cycles. Nikon's lens capacity, or the better availability of glass, is even more critical. It takes over a year to anneal the quartz material used in the optical train of the steppers. We must assume that Nikon has been planning this increase for many months and that its glass capacity will match its 1,000-stepper goal.

Nikon's Japanese counterpart. Canon, has been increasing capacity since early 1994. By our estimate. Canon is now running at a capacity of over

400 steppers per year. We expect Canon to increase its capacity to preserve and perhaps increase its share of about 25 percent of the stepper market.

Canon is on a very aggressive product introduction path as it tries to reestablish itself as a leader for 0.25-micron deep-UV lithography scanning tools.

The Ultratech and ASML announcement is also directly related to the capacity issue. ASML has been working hard with its optical shop, the lens supplier Zeiss of Germany, to increase capacity for its systems. Sophisticated tools and major investments are required to process the glass and manufacture and test the optical train for all stepper manufacturers. Based on announcements made by ASML and Zeiss, Dataquest estimates their lens capacity at fewer than 180 units for 1995. This figure could easily grow in the next several years.

ASML's primary strategy is to compete with Canon and Nikon on the highresolution stepper front as the industry moves to 0.35-micron and smaller geometries. However, resolution is not the only measure of steppers and equipment. Cost-effectiveness and productivity must match technical capabilities. Because of this, stepper manufacturers have introduced mix-and-match lithography as a means of reducing cost and increasing throughput.

Ultratech has led the way in matching its high-throughput systems in a mix with high-resolution steppers on the fab floor. Canon, ASML, and Nikon have all followed suit with their high-throughput systems. By our estimates, Ultratech shipped nearly 70 steppers for semiconductor manufacturing in 1994. The company has been growing strongly in all regions of the world with its i-line steppers for mix and match. Unlike other players in the market, which have product lead times of 18 months or more, Ultratech enjoys turnaround times of six to nine months. This is primarily because of the relatively simple optical design of the Ultratech steppers.

SEI\/lM-WW-MA-9504 ©1995 Dataquest Incorporated October 23,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

The agreement between Ultratech and ASML translates into a marketing strategy that would allow semiconductor manufacturers to choose between two mix-and-match scenarios. ASML and Ultratech will each propose two mix-and-match packages to the customer. ASML will offer one package pairing its high-resolution stepper with its own high-throughput system and one with the Ultratech companion. Ultratech will offer its system partnered with ASML's equivalent to the high-resolution stepper (which the prospective client may have already chosen) or with the non-ASML stepper.

Discussions between the companies are still in the preHiiunary stages. Both will benefit from this agreement even if cooperation goes no further than the outlines of their announcement. Most Ultratech high-throughput steppers are matched with Canon and Nikon steppers because of the massive installed base of these two vendors. Ultratech will be well positioned to enjoy the benefits of ASML's client base and to further solidify its position in the critical mix-and-match arena. In turn, ASML will get much-needed capacity relief for its high-resolution steppers.

It will be interesting to see what other synergy may exist between these two players. In 1994, ASML and Ultratech together held over 18 percent of a market that totaled more than $1.8 billion. Each has been growing over

60 to 70 percent per year in the past several years. Ultratech's presence in

Japem may compensate for ASML's lack of visibility in this critical market.

From a manufacturing point of view, both companies rely heavily on their external suppUers in strategic OEM agreements. Could this similarity and flexibility in nnanufacturing strategy and this partnership in mix and match lead to cooperation and exchange of technology?

We cannot close this discussion without mentioning the other critical player in the market, SVG Lithography (SVGL). For SVGL, more than any other company in the market, capacity does not merely translate into market share but may determine its future. SVGL's deep-UV Micrascan has led lithography's foray into sub-0.5 micron deep-UV scanning lithography.

Dataquest estimates that SVGL holds over 25 percent of the installed base of deep-UV steppers, trailing only Nikon's near-40 percent share of installed deep-UV systems.

Nikon's shipment earlier this year of the first NSR-S201A 0.25-micron deep-

UV scanning system to IBM's East Fishkill, New York, plant has the market awaiting SVGL's next stepper—the Micrascan III. However, semiconductor manufacturers are also interested in seeing how SVGL will build the infrastructure needed to support the deep-UV market into the next century.

SVGL has benefited from the approval and financial backing of several

U.S. semiconductor manufacturers and SEMATECH. Nobody doubts that

SVGL's position in technology is formidable, but annual unit shipments now in the low teens would have to increase to levels of deep-UV stepper capacity comparable to SVGL's competitors —ASML, Canon, and Nikon.

Semiconductor revenue and the ensuing growth in capital spending have translated into long backlogs for equipment suppliers, particularly for stepper manufacturers. If stepper undersupply is gating the supply of chips, the market's ability to spur demand by producing in volume and at lower cost will be greatly inhibited. We believe that by mid-1996, even at the current

SEMM-WW-MA-9504 ©1995 Dataquest Incorporated October 23,1996

Semiconductor Equipment, Manufacturing, and Materials Worldwide

high rate of demand, increases in stepper capacity will relieve the pressure on the market. Through adoption of mix-and-match lithography strategies and/or faster shrinks of IC designs, semiconductor manufacturers will address the growing demand for their products.

SEMM-WW-MA-9504 ©1995 Dataquest incorporated October 23,1995

Semiconductor Equipment, IVIanufacturing, and IVIaterials Worldwide

For More information...

Nader Pakdaman, Senior Industry Analyst (408) 468-8417

Internet address [email protected]

Via fax (408) 954-1780

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness.

It does not contain material provided to us in confidence by our clients. Reproduction or disclosure in whole or in

L / 3 X 3 Q U C S L part to other parties shall be made upon the written and express consent of Dataquest

^ _ _ , ©1995DataquestIncorporated —Reproduction Prohibited n n aoomponyof ^ "^ "^

MMMt TTKDun&Bradstreet Corporation Dataquest is a registered trademark of A.C. Nielsen Company

DataQuest

1994 Silicon Wafer iVIarlcet Share Estimates

Market Statistics

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-g501

Publication Date: June 19,1995

Filing: Market Analysis

1994 Silicon Wafer

Market Share Estimates

Market Statistics

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MS-9501

Publication Date: June 19,1995

Filing: Market Analysis

1994 Silicon Wafer l\/iarl<et Share Estimates

Table of Contents

Page

Section 1. Introduction 1

Definitions and Conventions 1

SUicon Products 1

Silicon Producers 1

Merchant Silicon and Epitaxial Wafer Suppliers 2

Section 2. Historical MSI by Region by Product, 1990-1994 4

Section 3. Historical Wafer Size Distribution by Region, 1990-1994 7

Section 4. Historical Market Share by Company, by Region, and by

Product, 1990-1994 13

Section 5. Wafer Pricing 29

Section 6. Silicon Plants 33

SEMi\/l-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

List of Tables

Table Page

1-1 Japanese and German Exchange Rates 1

1-2 Worldwide Merchant Silicon and Epitaxial Companies, 1994 3

2-1 Shipments of Merchant and Captive Silicon Wafers and

Merchant Epitaxial Wafers to Each Region, 1990-1994 4

2-2 Shipments of Merchant Epitaxial Wafers to Each Region,

1990-1994 4

2-3 Shipments of Merchant and Captive Silicon Wafers to Each

Region, 1990-1994 5

2-4 Shipments of Captive Silicon Wafers to Each Region,

1990-1994 5

2-5 Shipments of Merchant Silicon Wafers to Each Region,

1990-1994 6

2-6 Shipments of Merchant Test and Monitor Wafers to Each

Region, 1990-1994 6

3-1 Worldwide Wafer Size Distribution, 1987-1994 8

3-2 North American Wafer Size Distribution, 1987-1994 9

3-3 Japan Wafer Size Distribution, 1987-1994 10

3-4 European Wafer Size Distribution, 1987-1994 11

3-5 Asia/Pacific-ROW Wafer Size Distribution, 1987-1994 12

4-1 Top 15 Merchant Silicon and Epitaxial Wafer Manufacturing

Comparues Comparison of 1994 and 1993 Ranking by

Worldwide Revenues 13

4-2 Each Company's Revenue from Shipments of Merchant Silicon

Wafers and Epitaxial Wafers to the World, 1990-1994 14

4-3 Each Company's Revenue from Shipments of Merchant Silicon

Wafers to the World, 1990-1994 15

4-4 Each Company's Revenue from Shipments of Merchant

Epitaxial Wafers to the World, 1990-1994 16

4-5 Each Company's Revenue from Shipments of Merchant Silicon

Wafers and Epitaxial Wafers to North America, 1990-1994 17

4-6 Each Company's Revenue from Shipments of Merchant Silicon

Wafers to North America, 1990-1994 18

4-7 Each Company's Revenue from Shipments of Merchant

Epitaxial Wafers to North America, 1990-1994 19

4-8 Each Company's Revenue from Shipments of Merchant Silicon

Wafers and Epitaxial Wafers to Japan, 1990-1994 20

4-9 Each Company's Revenue from Shipments of Merchant Silicon

Wafers to Japan, 1990-1994 21

4-10 Each Company's Revenue from Shipments of Merchant

Epitaxial Wafers to Japan, 1990-1994 22

4-11 Each Company's Revenue from Shipments of Merchant Silicon

Wafers and Epitaxial Wafers to Europe, 1990-1994 23

4-12 Each Company's Revenue from Shipments of Merchant Silicon

Wafers to Europe, 1990-1994 24

Note: All tables show estimated data.

SEMM-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

1994 Silicon Water Market Share Estimates W

List of Tables (Continued)

Table Page

4-13 Each Company's Revenue from Shipments of Merchant

Epitaxial Wafers to Eiirope, 1990-1994 25

4-14 Each Company's Revenue from Shipments of Merchant

Silicon Wafers and Epitaxial Wafers to Asia/Pacific-ROW,

1990-1994 26

4-15 Each Company's Revenue from Shipments of Merchant Silicon

Wafers to Asia/Padfic-ROW, 1990-1994 27

4-16 Each Company's Revenue from Shipments of Merchant

Epitaxial Wafers to Asia/Pacific-ROW, 1990-1994 28

5-1 Regional Average Selling Price of Polished and Epitaxial

Wafers, at Start of Year 1990-1994 Price per Wafer 30

5-2 Regional Average Selling Price of Polished and Epitaxial

Wafers, at Start of Year 1990-1994 Price per Square Inch 31

6-1 Silicon Wafer Plant Exparisions/New Lines Since 1990 34

Note: All tables show estimated data.

SEMl\/l-WW-l\/iS-9501 ©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer IVIarlcet Share Estimates

Section 1 . Introduction

Dataquest's Seiiuconductor Equipment, Manufacturing, and Materials service tracks the silicon wafer industry by examining the merchant silicon and epitaxial wafer market, captive silicon production, wafer price trends, and silicon square-inch consumption.

The information in this document is focused on the silicon and epitaxial wafers used in the manufacturing of integrated circuits.

Definitions and Conventions

The calendar year sales of merchant silicon and epitaxial wafer suppliers are estimated in U.S. dollars and converted to millions of square inches using an average selling price for each region. Currency fluctuations over the last several years affect the dollar value of wafer sales of

Japanese and European companies. Dataquest uses average exchange rates supplied by the International Monetary Fund (IMF) to convert from local currency to U.S. dollars. The average exchange rates for the

Japanese yen and German deutsche mark for 1990 through 1994 are shown in Table 1-1.

Table 1-1

Japanese and German Exchange Rates

1990 1991

1992

Japanese Yen/U.S.$l 144 135 126

German Deutsche Mark/U.S.$l 1.62 1.66 1.55

Source: Dataquest (June 1995)

1993

111

1.65

1994

102

1.62

Please note the convention that the regional designation "United States" includes Canadian semiconductor manufacturing activities.

Silicon Products

The merchant silicon wafer market is categorized into two product segments—silicon wafers and silicon epitaxial wafers. Silicon wafers include prime, test, and monitor wafers grown by both Czochralski and float zone methods. In the silicon database, Dataquest does not include sales of polysilicon, single-crystal silicon ingots (unless noted), silicon materials used in solar applications, or compound semiconductor material substrates such as gallium arsenide.

Silicon Producers

Companies that produce silicon and epitaxial wafers are defined as either merchant silicon companies or captive silicon producers. Merchant silicon comparues are suppliers such as Shin-Etsu Handotai (SEH) of

Japan and Wacker of Germany.

SEMM-WW-MS-9501 ©1995 Dataquest Incorporated

Semiconductor Equipment, Manufacturing, and IVIaterials Woridwide

Silicon also is produced, to a lesser extent, by both merchant and captive semiconductor manufacturers. These semiconductor manufacturers collectively are referred to as captive silicon producers because they grow single-crystal silicon to produce wafers for their own internal consumption. Examples of captive producers with significant internal silicon production include AT&T, Motorola, and Texas Instruments in the

United States and Hitachi in Japan.

Merchant or Captive?

Some captive silicon producers have sold small amoimts of material on the merchant silicon market. These producers have sold wafers to ensure that internal production methods continue to produce material of competitive quality and cost. Dataquest estimates that merchant sales for these companies historically have represented a small percentage of their total captive silicon production; these companies are therefore identified as captive rather than merchant silicon producers.

Dataquest identifies Toshiba Ceramics, a subsidiary of Toshiba Corporation, as a merchant silicon company even though a substantial amount of its silicon production is consumed by its semiconductor parent.

However, because Toshiba Ceramics is actively marketing its material on the merchant market, it is considered a merchant rather than a captive silicon producer. Toshiba Corporation is considered a customer of

Toshiba Ceramics.

Merchant Silicon and Epitaxial Wafer Suppliers

Table 1-2 contains a list of merchant silicon manufacturers that were active in the worldwide market in 1994. This table, organized by region of corporate ownership, summarizes whether a company offers silicon and/or epitaxial wafers.

SEMM-WW-IVlS-9501 ©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer Marl<et Siiare Estimates

Table 1-2

Worldwide Merchant Silicon and Epitaxial Companies, 1994

Silicon Wafers Epitaxial Wafers

Companies

U.S. Companies

Crysteco Inc.

Epitaxy Inc.

General Instrument

Power Semiconductor Division

M/A-COM Semiconductor Products

Moore Technologies

Pure Sil (Formerly Pensilco)

Spire Corporation

Unisil

Virginia Semiconductor

Japanese Companies

Kawasaki Steel

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials

Mitsubishi Materials Silicon

Siltec Corporation

NSC Electron Corporation (Nittetsu Denshi)

Sumitomo SiTbc (formerly Osaka Titanium Company)

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

European Companies

Epitech

Huls

MEMC Electronic Materials

Okmetic

Siltronix SA

TopsU Semiconductor Materials A/S

Wacker

Rest of World Companies

Korea

Posco-Huls (incl in MEMC Electronic Mat'ls siimmary)

Siltron (formerly Lucky Advanced Materials Inc.)

Taiwan

Episil Technologies Inc. (Hermes Epitaxy affiliate)

Sino-America

Tatung Company

Source: Dataquest (June 1995)

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X'

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

SEMM-WW-lVlS-9501 ©1995 Dataquest Incorporated

June 19,1995

Semiconductor Equipment, iVIanufacturing, and Materials Worldwide

Section 2. Historicai IVISI by Region by Product, 1990-1994

Table 2-1

Shipments of Merchant and Captive Silicon Wafers* and Merchant Epitaxial Wafers to Each Region, 1990-1994 (Millions of Square Inches)

North America

Growth (%)

1990

640.2

10.0

1991

605.2

-5.5

1992

650.9

7.6

1993

720.0

10.6

1994

832.1

15.6

CAGR (%)

1990-1994

&8

Japan

Growth (%)

Europe

Growth (%)

Asia/Pacific-ROW

Growth (%)

991.9

8.7

235.4

1.8

180.8

39.2

1,038.5

4.7

212.2

-9.9

189.8

5.0

972.8

-6.3

235.0

10.7

238.0

25.4

1,127.3

15.9

291.2

23.9

311.2

30.8

1,279.7

13.5

351.2

20.6

456.0

46.5

6.6

10.5

26.0

Worldwide

Growth (%)

2,048.3

10.4

•Includes prime, virgin test, and monitor wafers

Source: Dataquest (June 1995)

2,045.7

-0.1

2,096.7

2.5

2,449.7

16.8

2,919.0

19.2

Table 2-2

Shipments of Merchant Epitaxial Wafers to Each Region, 1990-1994

(Millions of Square Inches)

9.3

CAGR (%)

1990-1994

21.2

North America

Growth (%)

Japan

Growth (%)

Europe

Growth (%)

Asia/Pacific-ROW

Growth (%)

Worldwide

Growth (%)

Source: Dataquest (June 1995)

204.5

9.5

1990

88.1

7.8

92.9

11.8

18.9

8.0

4.6

2.2

1991

84.2

-4.4

104.0

11.9

23.0

21.7

8.3

80.4

219.5

7.3

1992

128.1

52.1

87.0

-16.3

23.9

3.9

9.5

14.5

248.5

13.2

1993

154.5

20.6

89.0

2.3

36.1

51.0

13.0

36.8

292.6

17.7

1994

190.2

23.1

119.5

34.3

54.5

51.0

19.5

50.0

383.7

31.1

6.5

30.3

43.5

17.0

SEMM-WW-MS-9501

©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer l\/iarl<et Sliare Estimates

Table 2-3

Shipments of Merchant and Captive Silicon Wafers* to Each Region, 1990-1994

(Millions of Square Inches)

North America

Growth (%)

1990

552.1

10.4

1991

521.0

-5.6

1992

522.8

0.3

1993

565.5

8.2

1994

641.9

13.5

CAGR (%)

1990-1994

3.8

Japan

Growth (%)

899.0

8.4

934.5

3.9

885.8

-5.2

1,038.3

17.2

1,160.2

11.7

6.6

8.2 Europe

Growth (%)

Asia/Pacific-ROW

Growth (%)

216.5

1.3

176.2

40.5

189.2

-12.6

181.5

3.0

Worldwide

Growth (%)

1,843.8

10.5

'Includes prime, virgin test, and monitor wafers

Source: Dataquest (June 1995)

1,826.2

-1.0

211.1

11.6

228.5

25.9

1,848.2

1.2

255.1

20.8

298.2

30.5

2,157.1

16.7

Table 2-4

Shipments of Captive Silicon Wafers* to Each Region, 1990-1994

(Millions of Square Inches)

296.7

16.3

436.5

46.4

2,535.3

17.5

25.5

8.3

CAGR (%)

1990-1994

3.0

North America

Growth (%)

Japan

Growth (%)

Europe

Growth (%)

Asia/Pacific-

Growth (%)

ROW

1990

80.0

-2.4

46.0

24.3

8.0

60.0

0

N M

1991

70.0

-12.5

40.0

-13.0

5.0

-37.5

0

N M

Worldwide

Growth (%)

134.0

8.1

'Includes prime virgin test, and monitor wafers

NM = Not mean ngful

Source: Dataquest (June 1995)

115.0

-14.2

5.0

0

6.0

N M

120.0

4.3

1992

72.0

2.9

37.0

-7.5

1993

78.0

8.3

35.0

-5.4

5.0

0-

7.0

16.7

125.0

4.2

5.0

0

7.0

0

137.0

9.6

1994

90.0

15.4

35.0

0

-6.6

-11.1

N M

0.6

SEMM-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 2-5

Shipments of Merchant Silicon Wafers* to Each Region, 1990-1994

(Millions of Square Inches)

North America

Growth (%)

Japan

Growth (%)

Europe

Growth (%)

1990

472.1

12.9

853.0

7.6

208.5

-0.1

1991

451.0

-4.5

894.5

4.9

184.2

-11.7

Asia/Pacific-ROW

Growth (%)

176.2

40.5

181.5

3.0

Worldwide

Growth (%)

1709.8

10.7

•Includes prime, virgin test and monitor wafers

Source: Dataquest (June 1995)

1,711.2

0.1

1992

450.8

0

848.8

-5.1

206.1

11.9

222.5

22.6

1,728.2

1.0

1993

487.5

8.1

1,003.3

18.2

250.1

21.3

291.2

30.9

2,032.1

17.6

1994

551.9

13.2

1,125.2

12.1

291.7

16.6

429.5

47.5

2,398.3

18.0

CAGR (%)

1990-1994

4.0

7.2

8.8

25.0

8.8

Table 2-6

Shipments of Merchant Test and Monitor Wafers to Each Region, 1990-1994

(Millions of Square Inches)

North America

Japan

Europe

Asia/Pacific-ROW

Worldwide

Growth (%)

Source: Dataquest (June 1995)

1990

94.4

170.6

41.7

35.2

342.0

10.7

1991

90.2

178.9

36.8

36.3

342.2

0.1

1992

90.2

169.8

41.2

44.5

345.6

1.0

1993

102.4

203.2

50.0

59.7

415.3

20.1

1994

121.4

230.7

61.3

92.3

505.7

21.8

SEMM-WW-MS-9501

©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer Market Share Estimates

Section 3. Historical Wafer Size Distribution by Region, 1990-1994

SEM!V1-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

Semiconductor Equipment, iVIanufacturing, and l\/laterials Worldwide

T - H - ^ O q t v O O O O N C N I

O r H - < * T j i o d < - H O S o ^

T-H CM TJI T-H O ° l l-H

T-l CS

\0 rH 00 \0 SO

O O 00

\0 00 OS ro O

• ^ i 2 r H K i n o s c N s q p c j N

r H i r J i r J K r - i v O O C J ' " !

CO CO in CO in o so in

•<* 00 in in in CN tN •>*

CN in

CO CO -^

CM i ^ C ^ p c O C ^ O l N L O d

( N o 6 ( r i c d c N Q r ^ < N

T-t CO ^ O ^

T-H CM

cs o

i n

CN r H rH

o csT o

1

CO CO CO CN CN CO CN so CN so CO rH CD "^

CO CO CO rH T^ in CO rH in t>. p rH t>^ «* f,; Os rH

CO CO CN ii

C 0 O - * ' * C M C N , O 0 0 ' *

C D C O C O t N ^ r H o S d

CS CO CO O K rH rH CN|

•c

(Q

^ - 1

(X)

<H

M H

I

C O t ^ i - H O r H O S O S O t N

d c o s o d o s c S d ^ L o

CS| • * CM O " < rH

o

• ^ O O ' ^ - ^ O O O C O I D

d c o o d c o T j H o d S o d

CM • * (N O ^N rH

S

4 i

«-<

CA

• iM

D

»*

0^ rH

I N

D

TJ

oo

a\

rH

G

n

V

9

0) u

s

'C

Q

>.

D ^

01

N

1/5

u

G

ta

V

^

e

&

CO j 3

tn

01

MH (U

;-l

(S i n

o

i n

CO

r-l

O

CO rH i n

T-l

CO

C ^ i n

o

CN

CO

o

O •<*

O N

1-^

so

S t^ tN rH

(N rH

CN

r^

0 0

CO fv. fsj

C ^ so

0 0

0)

cr

^ T S « m S

<u2

c

0)

)H

-S ,0

4J

f 2 ^ 5 i

l l m

C/J

u

4i

1

>,

.c

c

•X!

{Ti

OJ ffl

^

V*-* o

(f> 'tH

UJ

Q

'S

D i

C ^ O O C O C O s O t V ' ^ C O

CO - ^ (N

S O O O O O O t > ; C O C N C O r H O s O s O v O s C D O ^

CO CO rH rH . ^

C O O S O S i n C O C O r H O O

C N s d s D s O ' * d C < " ^

CO CO 1-1

O N

" *

CO

CN t->

SO i n

CO

CO r H

CN

CO r H

O

r-l

r H

O

00 rH

^ "1

00 - ^

• ^ t v Cs CN 0 0 !>. rH O rH O CO SO

CO t ^ rs|

ON

C^ 0 0 rH rH CN rj<

D

O

CO

I

SEI\/IM-WW-IVIS-9501 ©1995 Dataquest Incorporated

June 19,1995

t

en

I

to

CJI

o

@

CO

so

O l

o

Bf

X3

c=

CD

o

a

Table 3-2

North American Wafer Size Distribution, 1987-1994

(Percent Square Inches by Diameter and Unit Distribution by Wafer Starts)

Area

(sq-in.) 1987

1988

1989 1990 Diameter

Percent Square lr\ches by Diameter

2 inches

3 inches lOOmm

125mm

ISOrtim

200mm

Total

Total - MSI

Growth (%)

3.14

7.07

12.17

19.02

27.38

48.67

0.1

3.3

36.3

40.1

19.4

0.7

100.0

442

8.9

0.1

2.2

33.4

41.9

20.7

1.7

100.0

546

23.7

0.1

3.2

31.6

36.7

26.9

1.5

100.0

582

6.5

0.1

1.8

28.8

35.4

30.5

3.4

100.0

640

10.0

Unit Distribution by V^^ Starts

(Millions of Wafers)

2 inches

3 inches

100mm

125mm

ISOmm

200mm

Total Wafers (M)

Avg Wafer Diam (")

Source: Dataquest (June 1995)

3.14

7.07

12.17

19.02

27.38

48.67

0.1

2.1

13.2

9.3

3.1

0.1

27.9

4.49

0.2

1.7

15.0

12.0

4.1

0.2

33.2

4.58

0.2

2.6

15.1

11.2

5.7

0.2

35.1

4.60

0.2

1.6

15.2

11.9

7.1

0.4

36.5

4.73

1991

0.1

1.5

26.8

32.5

35.6

3.5

100.0

605

-5.5

0.2

1.3

13.3

10.3

7.9

0.4

33.4

4.80

19

2

3

3

10

6

3

4

1

1

CO

CD

CJI

Semiconductor Equipment, Manufacturing, and Materials Woridwide

10 rH O

ON

i-H \ o

^ CN LD fS 00

CO

8 <H

CN csl

CO

CO

CO

m m l - H o o o

t x OS

' ^ T-H o tx es c>« OS Ln

^ CO s o s o i-H

T-i f O • *

O CO CO

8 ^ ^

i n •>* i-H T-H I^N 0 0 EN

CNJ T-H O •>* T-H ON ' ^

'-I CN4 cN m m

o

• *

CM

Cx

o

1—1

I—1

ON

r—1

OS

T—1

CM

CO

o

>*

• < * m

•<i<

I—1 m

O C O s O O O t v C O s o m csi CD 00 s o O 00 P

T—I 1—I 1—I • ^ m

T-H p sc CO so Ln

O ON Cx

C5 CN i n ON i-i r-J

8 o -^

I-H CO T f

C O O s c O m o O C O i - H - ^

C D C N C O i - H l r i o - * " ^ r-l CM i - l m • ^ r H CO ON OS

•>a; Lo p

O

CN t v r-i so o Q

0 0 t-H T t CO O

en

«

CD

u

MH

I

>>

4^ o

• P M

^rf

l - l s o c ^ ^ l r ) c o c o o c O l ^ l cs

^ s ° 8 s; 5^

0 0 - < * i - H - * c s l o t x O

c s i i - i o 6 K d o E > s o

CS -^l^ (N O "-^ <-!

*C

Q

«M

• f ^

C

;D

S-o

00 *f

OS OJ

^. s

.2

0

1-1

o cs r4 CO i-H

csi i n IN

o g ^ ^

Tt< Cx t x CM 00 Cx l-H O I-H O CO ^

CO t x CM ON

IN 0 0

CN •<*

U (A

w z

OS

4) HN

N oj

cn «

h S

(u cr

CO > C

SB^

• 2 c>H«

E2 S.^

SEMM-WW-MS-9501

3

rr

rn

4-> c

C I

u

U i cc.

<u

•6

.g

CM en tu

-6

,Fli

CO

C O C N S O ' ^ ( N , - H 0 0 m

c 5 c o - < ^ c s i c o o c o ° 9

T-H CM i-H m • ^

CO • * CN CO i-H

CO vo

o CO m (N o o T-i

i-H cs i-H m

C N | r H t N , S O O O O T } ( C S

C> CO CO OS t< •^ •

• *

o

IN

CN

CO s

IN

0 0 rH

CO m o •*

ON

CO so

SJJ

•^ IN tv tN 00 tv

T—I P T-H p CO SO

CO IN CM OS IN 00

i-l r-l C M • ^

©1995 Dataquest Incorporated cn

3

O

I

June 19,1995

Crt

Table 3-4

European Wafer Size Distribution, 1987-1994

(Percent Square Inches by Diameter and Unit Distribution by Wafer Starts)

^ Area

g Diameter (sq.in.) 1987 1988 1989 1990 1991

~^ Percent Square Inches by Diameter

2 inches

3 inches

3.14

7.07

0.4

5.9

0.4

5.0

0.4

4.3

0.3

3.7

0.3

3.1 lOOmjn

125mni

12.17

19.02

44.9

33.9

40.1

34.3

35.9

33.0

32.5

31.9

29.1

30.8

27.38 14.9 19.5 24.4 28.3 32.3 ISOrrim

200mm

Total

48.67

0

100.0

172

0.7

100.0

196

2.0

100.0

231

3.3

100.0

235

4.4

100.0

212

1

® Total - MSI g Growth (%) a

Sf

S3

Unit Dishibution by Wlldfei! Starts

§. (Millions of Wafers) g 2 inches

o

•3 3 inches

3

cff

100mm

CL

125mm

150mm

20Qmm

Total Wafers (M)

Avg Wafer Diam (")

3.14

7.07

12.17

19.02

27.38

48.67

11.0

0.2

1.4

6.3

3.1

0.9

0

12.0

4.27

14.0

0.2

1.4

6.5

3.5

1.4

0

13.1

4.37

17.9

0.3

1.4

6.8

4.0

2.1

0.1

14.7

4.48

1.7

0.2

1.2

6,3

3.9

2.4

0.2

14.3

4.58

-9.7

0.2

0.9

5.1

3.4

2.5

0.2

12.3

4.68

Source: Dataquest (June 1995)

CD

CO en

C/)

I

CO o i

o

@

CO

C D t n

O s.

( U

X 3

o o

CD

I

o.

Table 3-5

Asia/Facific-ROW Wafer Size Distribution, 1987-1994

(Percent Square Inches by Diameter and Unit Distribution by Wafer Starts)

Area

(sq.in.)

1987 1988 1989 1990 Diameter

Percent Square Inches by Diameter

2 inches

3 inches lOOmm

125inin

ISOmm

200min

Total

Total - MSI

Growth (%)

3.14

7.07

12.17

19.02

27.38

48.67

7.0

6.0

28.0

28.0

31.0

0

100.0

70

9.4

6.0

6.0

25.5

30.0

32.5

0

100.0

84

20.0

2.0

13.0

25.1

21.2

38.7

0

100.0

130

54.8

1.6

10.5

22.1

21.8

43.3

0.7

100.0

181

39.2

Unit Distribution b y V ^ e r Starts

(Millions of Wafer^J

2 inches

3 inches lOOmm

125mm

ISOmm

200mm

Total Wafers (M)

Avg Wafer Diam ("}

Source: Dataquest (June 1995)

3.14

7.07

12.17

19.02

27.38

48.67

1.6

0.6

1.6

1.0

0.8

0

5.6

3.99

1.0

0

6.4

4.09

1.6

0.7

1.8

1.3

0.8

2.4

2.7

1.4

1.8

0

9.2

4.24

2.9

0

11.9

4.41

0.9

2.7

3.3

2.1

3.5

0.1

11.2

4.65

0.7

2.2

2.4

2.2

1991

1.2

8.1

15.5

22.4

51.0

1.8

100.0

190

4.9

1

CO

CD

(SI

1994 Silicon Wafer i\/larl<et Share Estimates

13

Section 4. Historical Marlcet Share by Company, by Region, and by Product,

1990-1994

Table 4-1

Top 15 Merchant Silicon and Epitaxial Wafer Manufacturing Companies

Comparison of 1994 and 1993 Ranking by Worldwide Revenue

(End-User Revenue in Millions of U.S. Dollars)

1994

Rank

1993

Rank

All Companies

Shin-Etsu Handotai

MEMC Electronic Materials + Joint Ventures

Siimitomo SiTix

Wacker Chemitronic + SHtronic

Mitsubishi Materials Silicon

Komatsu Electronic Metals

Toshiba Ceramics

Nittetsu Denshi

Lucky Advanced Materials Inc.

UruSil Corp.

Showa Denko

Topsn

Okmetic

Crysteco Inc.

EpisU

1994

4,591.9

1,174.3

757.6

582.4

500.4

475.2

452.9

275.0

105.8

89.8

40.0

20.7

20.1

19.0

17.9

15.7

8

9

10

11

12

13

14

15

4

5

6

7

1

2

3

84.5

61.0

22.3

16.4

18.5

16.7

14.8

13.1

1993

3,554.3

895.7

569.8

443.6

401.2

363.6

377.3

209.1

8

9

10

13

11

12

14

15

4

6

5

7

1

2

3

Percent

Change

29

31

33

31

25

47

79

26

25

31

20

32

9

14

21

20

All Other Companies

Source: Dataquest (June 1995)

45.1 46.7

-3

SEMM-WW-IVlS-9501

©1995 Dataquest incorporated June 19,1995

14 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 4-2

Each Company's Revenue from Shipments of Merchant Silicon Wafers* and Epitaxial

Wafers to the World, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991

1992

1993 1994

CAGR (%)

1990-1994

Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo SiTix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials

Okmetic

TopsU

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

11.0

279.0

317.4

32.0

348.1

691.5

10.2

129.7

1,818.9

410.9

5.0

15.5

313.6

1.4

746.4

3.0

0

26.2

10.0

39.2

14.1

10.7

12.5

22.6

59.9

11.3

332.6

326.4

56.4

382.1

768.5

11.9

148.2

2,037.4

446.6

4.5

19.5

295.9

2.0

768.5

3.2

0

35.4

12.5

51.1

16.8

10.7

16.6

11.9

56.0

13.9

317.8

306.5

72.0

370.9

780.4

13.9

140.7

2,016.1

16.3

10.6

17.6

13.1

57.6

483.7

5.9

17.3

337.3

2.0

846.2

11.2

0

48.1

12.2

71.5

14.8

12.5

22.3

16.5

66.1

8.7

377.3

363.6

84.5

443.6

895.7

16.4

209.1

2,398.9

552.5

16.7

18.5

401.2

2.2

991.1

13.1

17.3

61.0

6.8

98.2

17.9

14.8

40.0

20.0

92.7

0.0

452.9

475.2

105.8

582.4

1,174.3

20.7

275.0

3,086.3

660.8

19.0

20.1

500.4

2.6

1,202.9

15.7

96.8

89.8

in

210.0

11.5

14.1

12.7

52.1

Total Sales—World

2,664.4 2,913.0 2,991.4 3,554.3 4,591.9 14.6

'Includes prime, test, and monitor wafers

2

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEMM-WW-MS-9501 ©1995 Dataquest incorporated June 19,1995

1994 Silicon V\fafer l\/larl<et Share Estimates

15

Table 4-3

Each Company's Revenue from Shipments of Merchant Silicon Wafers* to the World,

1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993 1994

CAGR (%)

1990-1994

Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo SiTix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials^

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

320.9

5.0

15.5

220.3

1.4

563.1

14.1

0

12.5

14.1

40.7

10.7

225.2

227.2

32.0

272.7

499.9

10.2

80.8

1,358.7

0

0

26.2

10.0

36.2

10.3

261.9

233.0

56.4

283.0

554.6

11.9

90.7

1,501.8

346.5

4.5

19.5

217.4

2.0

589.9

0

0

35.4

12.5

47.9

16.8

0

16.6

5.6

39.0

11.3

258.5

218.9

72.0

265.4

562.3

13.9

91.9

1,494.2

373.8

5.9

17.3

214.8

2.0

613.8

0

0

48.1

12.2

60.3

16.3

0

17.6

4.5

38.4

5.5

295.4

282.1

84.5

307.5

643.1

16.2

151.2

1,785.5

403.1

16.7

18.5

224.1

2.2

664.6

14.8

0

22.3

4.9

42.0

0

17.3

61.0

6.8

85.1

0.0

343.1

342.2

105.8

373.0

882.0

19.6

203.3

2,269.0

440.5

19.0

20.1

270.3

2.6

752.5

17.9

0

40.0

5.9

63.8

0

96.8

89.8

in

194.3

11.9

13.7

7.5

52.2

Total Sales—World 1,998.7 2,178.6 2,206.7

2,577.2 3,279.6 13.2

•Includes prime, test, and monitor wafers

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron Included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec Included from 1994 onward

Source: Dataquest (June 1995)

SEMM-WW-i\/lS-9501

©1995 Dataquest Incorporated June 19,1995

16 Semiconductor Equipment, IVIanufacturing, and IVlaterials Worldwide

Table 4-4

Each Company's Revenue from Shipments of Merchant Epitaxial Wafers to the World,

1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990

1991

1992 1993

1994

CAGR (%)

1990-1994

Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials SOicon

NSC Electron Corporation

Sumitomo Sillx Silicon

Shin-Etsu Handotai

Showa Deriko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Comparues

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

0.3

53.8

90.2

0

75.4

191.6

0

48.9

460.2

90.0

0

0

93.3

0.0

183.3

3.0

0

0

0

3.0

0

10.7

0

8.5

19.2

100.1

0

0

78.5

0.0

178.6

1.0

70.7

93.4

0

99.1

213.9

0

57.5

535.6

3.2

0

0

0

3.2

0

10.7

0

6.3

17.0

2.6

59.3

87.6

0

105.5

218.1

0

48.8

521.9

109.9

0

0

122.5

0.0

232.4

11.2

0

0

0

11.2

0

10.6

0

8.6

19.2

149.4

0

0

177.1

0.0

326.5

3.2

81.9

81.5

0

136.1

252.6

0.2

57.9

613.4

0

12.5

0

11.6

24.1

13.1

0

0

0

13.1

220.3

0

0

230.1

0.0

450.4

0.0

109.8

133.0

0

209.4

292.3

1.1

71.7

817.3

15.7

0

0

0

15.7

0

14.8

0

14.1

28.9 10.8

29.1

25.2

51.2

Total Sales—World

665.7

734.4

784.7 977.1 1,312.3

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 ononic Materials Included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

18.5

SEiviivi-ww-ivis-gsol

©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer IVIarket Siiare Estimates 17

Table 4-5

Each Company's Revenue from Shipments of Merchant Silicon Wafers* and Epitaxial

Wafers to North America, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993 1994

CAGR (%)

1990-1994 Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Smnitomo SiTbc Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

5.7

7.0

55.0

0

66.9

138.5

0

0

273.1

199.0

0

2.8

152.0

0

353.8

0

0

0.5

0

0.5

11.0

4.7

12.5

21.2

49.4

5.8

7.5

55.7

0

69.5

161.9

0

0

300.4

201.6

0

2.8

126.8

0

331.2

0

0

0.5

0

0.5

13.6

4.7

13.9

10.5

42.7

9.8

8.4

69.6

0

96.3

193.0

0

0

377.1

202.6

0

2.2

143.0

0

347.8

0

0

3.0

0

3.0

11.4

5.2

14.8

11.8

43.2

215.8

0

2.4

180.0

0

398.2

4.3

20.7

81.9

0

127.4

224.8

0

0

459.1

0

0

4.5

0

4.5

11.3

5.9

15.6

13.9

46.7

0

33.7

101.2

0

182.8

280.0

0

0

597.7

264.7

1.1

2.6

205.0

0

473.4

0

0

5.8

0.1

5.9

13.6

7.0

29.9

16.5

67.0 7.9

21.6

7.6

85.3

Total Sales—^North America 676.8 674.8 777.1 908.5 1,144.0 14.0

'Includes prime, test, and monitor wafers

2

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials Included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEiViM-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

18

Semiconductor Equipment, IVianufactuiing, and Materials Woridwide

Table 4-6

Each Company's Revenue from Shipments of Merchant Silicon Wafers* to North

America, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992

1993 1994

CAGR (%)

1990-1994

Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo Silix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Comparues

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

11.0

0

12.5

13.7

37.2

5.4

5.6

47.8

0

18.0

86.7

0

0

163.5

137.0

0

2.8

90.0

0

229.8

0.5

0

0

0

0.5

135.4

0

2.8

88.0

0

226.2

4.8

4.7

44.6

0

19.4

98.8

0

0

172.3

0

0

0.5

0

0.5

13.6

0

13.9

5.2

32.7

7.4

5.2

49.8

0

30.8

115.5

0

0

208.7

129.8

0

2.2

67.0

0

199.0

0

0

3.0

0

3.0

11.4

0

14.8

4.1

30.3

120.2

0

2.4

74.0

0

196.6

1.3

18.5

58.5

0

43.2

134.9

0

0

256.4

0

0

4.5

0

4.5

11.3

0

15.6

4.5

31.4

130.8

1.1

2.6

82.0

0

216.5

0

29.1

56.6

0

69.7

167.0

0

0

322.4

0

0

5.8

0.1

5.9

13.6

0

29.9

5.4

48.9 7.1

18.5

-1.5

85.3

Total Sales—^North America 431.0 431.7 441.0 488.9 593.7 8.3

'Includes prime, test, and monitor wafers

2

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEi\/lM-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer l\/larket Share Estimates

19

Table 4-7

Each Company's Revenue from Shipments of Merchant Epitaxial Wafers to North

America, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993 1994

CAGR (%)

1990-1994 Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials SUicon

NSC Electron Corporation

Siimitomo Silix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials^

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

48.9

51.8

0

0

109.6

0.3

1.4

7.1

0

62.0

0

0

62.0

0

124.0

0

4.7

0

7.5

12.2

0

0

0

0

0

1.0

2.8

11.1

0

50.1

63.1

0

0

128.1

66.2

0

0

38.8

0

105.0

0

0

0

0

0

0

4.7

0

5.3

10.0

72.8

0

0

76.0

0

148.8

2.4

3.2

19.8

0

65.5

77.5

0

0

168.4

0

0

0

0

0

0

5.2

0

7.7

12.9

3.0

2.2

23.4

0

84.2

89.9

0

0

202.7

95.6

0

0

106.0

0

201.6

0

0

0

0

0

0

5.9

0

9.4

15.3

0

4.6

44.6

0

113.1

113.0

0

0

275.3

133.9

0

0

123.0

0

256.9

0

0

0

0

0

0

7.0

0

11.1

18.1 10.4

25.9

20.0

Total Sales—^North America 245.8 243.1 330.1 419.6

550.3 22.3

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; onic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEIVlM-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

20

Semiconductor Equipment, Manufacturing, and Materials Woridwide

Table 4-8

Each Company's Revenue from Shipments of Merchant Silicon Wafers* and Epitaxial

Wafers to Japan, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993

1994

CAGR (%)

1990-1994

Company

United States-Based Compardes

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo SiTix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Comparues

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

2.1

265.0

258.0

32.0

250.5

492.5

10.2

127.4

1,437.7

0

0

0

0

0

54.4

0

6.0

30.0

0

90.4

0.8

0

0

0

0.8

2.7

318.3

267.3

56.4

269.6

534.5

11.9

145.8

1,606.5

66.2

0

8.5

39.7

0

114.4

0

0

0

0

0

0.8

0

0.1

0

0.9

2.6

302.0

233.3

72.0

226.1

499.8

13.9

138.0

1,487.7

62.8

0

7.2

31.1

0

101.1

8.8

0

0

4.0

12.8

1.0

0

0.1

0

1.1

3.0

352.5

260.8

84.5

260.1

562.1

16.4

206.0

1,745.4

87.7

0

7.5

31.7

0

126.9

10.1

0

0

2.6

12.7

0.6

0

0.5

1.8

2.9

0.0

408.2

328.8

103.1

316.5

658.1

20.7

271.0

2,106.4

124.7

0

7.8

41.2

0

173.7

12.0

0

0

0.1

12.1

0.8

0

1.2

2.0

4.0 49.5

10.0

17.7

Total Sales—^Japan 1,528.9 1,721.8 1,602.7 1,887.9

2,296.2

'Includes prime, test, and monitor wafers

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 ononic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

10.7

SEMM-WW-MS-9501

©1995 Dataquest Incorporated

June 19,1995

1994 Silicon Water l\/larl<et Share Estimates

21

Table 4-9

Each Company's Revenue from Shipments of Merchant Silicon Wafers* to Japan,

1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992

1993 1994

CAGR (%)

1990-1994 Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Simiitomo SiTix SUicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episn

Posco-Huls

Siltron

Other ROW Companies

Total

2.1

212.6

176.0

32.0

227.0

355.5

10.2

80.0

1,095.4

51.5

0

6.0

25.8

0

83.3

0.8

0

0

0

0.8

0

0

0

0

0

2.7

250.4

185.6

56.4

236.1

387.1

11.9

89.9

1,220.1

63.3

0

8.5

30.4

0

102.2

0

0

0

0

0

0.8

0

0.1

0

0.9

2.4

245.9

166.1

72.0

200.8

363.8

13.9

89.6

1,154.5

59.4

0

7.2

23.1

0

89.7

0

0

0

4.0

4.0

1.0

0

0.1

0

1.1

2.8

272.9

203.5

84.5

224.6

404.7

16.2

148.2

1,357.4

76.4

0

7.5

23.5

0

107.4

0

0

0

2.6

2.6

0.6

0

0.5

0

1.1

0

303.0

241.2

103.1

250.9

486.2

19.6

199.3

1,603.3

108.4

0

7.8

29.1

0

145.3

0

0

0

0.1

0.1

0.8

0

1.2

0

2.0 25.7

10.0

14.9

Total Sales—^Japan

1,179.5 1,323.2 1,249.3 1,468.5 1,750.7 10.4

•Includes prime, test, and monitor wafers

2

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 ond; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEMI\/l-WW-!ViS-9501

©1995 Dataquest Incorporated

June 19,1995

22

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 4-10

Each Company's Revenue from Shipments of Merchant Epitaxial Wafers to Japan,

1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993

1994

CAGR (%)

1990-1994 Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo Silix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials^

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

SUtron

Other ROW Companies

Total

0

52.4

82.0

0

23.5

137.0

0

47.4

342.3

2.9

0

0

4.2

0

7.1

0

0

0

0

0

0

0

0

0

0

0

67.9

81.7

0

33.5

147.4

0

55.9

386.4

0

0

0

0

0

2.9

0

0

9.3

0

12.2

0

0

0

0

0

0.2

56.1

67.2

0

25.3

136.0

0

48.4

333.2

8.8

0

0

0

8.8

3.4

0

0

8.0

0

11.4

0

0

0

0

0

0.2

79.6

57.3

0

35.5

157.4

0.2

57.8

388.0

11.3

0

0

8.2

0

19.5

1.8

1.8

0

0

0

10.1

0

0

0

10.1

0

105.2

87.6

0

65.6

171.9

1.1

71.7

503.1

16.3

0

0

12.1

0

28.4

12.0

0

0

0

12.0

0

2.0

2.0

0

0

10.1

41.4

Total Sales—^Japan

349.4 398.6 353.4 419.4 545.5 11.8

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEMM-WW-MS-9501

©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer Market Share Estimates 23

Table 4-11

Each Company's Revenue from Shipments of Merchant Silicon Wafers* and Epitaxial

Wafers to Europe, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993 1994

CAGR (%)

1990-1994

Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials SUicon

NSC Electron Corporation

Siomitomo SiTix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials^

Okmetic

Topsn

Wacker

Other European Companies

Total

ROW-Based Companies

EpisU

Posco-Huls

SUtron

Other ROW Companies

Total

24.3

23.5

0

0

57.4

0

7.0

2.6

0

103.7

4.3

5.2

105.7

1.4

220.3

0

0

0

0

0

2.3

0.9

0

0.9

4.1

101.8

4.0

6.3

105.7

2.0

219.8

30.5

29.8

0

0

68.6

0

6.8

1.5

0

0

0

0

0

0

2.4

0.9

0.1

0.9

4.3

0

7.4

1.6

0

30.2

43.8

0

0

83.0

103.9

5.4

5.7

131.3

2.0

248.3

0

0

0

0

0

3.3

0.7

0.2

0.7

4.9

116.4

11.1

6.0

150.0

2.2

285.7

0

2.7

1.7

0

31.2

54.5

0

0

90.1

0

0

0

0

0

1.9

0.9

2.3

0.6

5.7

159.8

12.9

6.5

205.0

2.6

386.8

0

3.4

2.0

0

51.6

88.2

0

0

145.2

0.9

0

0

0

0.9

3.3

1.0

3.6

0.7

8.6 20.3

26.1

15.1

Total Sales—Europe 281.8

292.7

336.2

381.5

541.5 17.7

•Includes prime, test, and monitor wafers

2

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEI\/il\/l-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

24

Semiconductor Equipment Manufacturing, and Materials Woridwide

Table 4-12

Each Company's Revenue from Shipments of Merchant Silicon Wafers* to Europe,

1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990

1991

1992

1993 1994

CAGR (%)

1990-1994

Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo Silix Silicon

Shin-Etsu Handotai

Showa Deriko

Toshiba Ceramics

Total

Evirope-Based Companies

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

EpisU

Posco-Huls

Siltron

Other ROW Companies

Total

22.0

21.0

0

0

51.8

0

7.0

1.8

0

79.4

4.3

5.2

80.0

1.4

170.3

0

0

0

0

0

2.3

0

0

0.2

2.5

0

6.8

1.1

0

20.4

26.7

0

0

55.0

75.3

4.0

6.3

78.6

2.0

166.2

0

0

0

0

0

2.4

0

0.1

0.2

1.7

0

7.4

1.2

0

21.7

39.5

0

0

69.8

76.4

5.4

5.7

97.2

2.0

186.7

0

0

0

0

0

3.3

0

0.2

0.2

3.7

81.5

11.1

6.0

96.0

2.2

196.8

0

2.7

1.3

0

23.1

49.5

0

0

76.6

0

0

0

0

0

1.9

0

2.3

0.2

4.4

0

3.4

1.5

0

33.4

81.5

0

0

119.8

100.7

12.9

6.5

123.0

2.6

245.7

0

0

0

0

0

3.3

0

3.6

0.2

7.1

29.8

23.3

9.6

Total Sales—Europe 224.6 223.9

260.2 277.8 372.6 13.5

'Includes prime, test, and monitor wafers

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

^Monsanto Electronic Materials Included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEMM-WW-MS-9501

©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer Marl<et Share Estimates 25

Table 4-13

Each Company's Revenue from Shipments of Merchant Epitaxial Wafers to Europe,

1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993 1994

CAGR (%)

1990-1994 Company

United States-Based Companies

Crysteco

Epitaxy Inc.

UnisU

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo SiTix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Etirope-Based Companies

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

0

0

0.8

0

2.3

2.5

0

0

5.6

0

0

0

0

0

24.3

0

0

25.7

0

50.0

0

0.9

0

0.7

1.6

0

0

0.4

0

10.1

3.1

0

0

13.6

0

0

0

0

0

26.5

0

0

27.1

0

53.6

0

0.9

0

0.7

1.6

0

0

0.4

0

8.5

4.3

0

0

13.2

27.5

0

0

34.1

0

61.6

0

0

0

0

0

0

0.7

0

0.5

1.2

0

0

0.4

0

8.1

5.0

0

0

13.5

34.9

0

0

54.0

0

88.9

0

0

0

0

0

0

0.9

0

0.4

1.3

0.5

0

0

0

18.2

6.7

0

0

25.4

59.1

0

0

82.0

0

141.1

0.9

0

0

0

0.9

0

1.0

0

0.5

1.5

-1.6

45.9

29.6

Total Sales—Europe 57.2 68.8 76.0

103.7 168.9 31.1

2

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec Included from 1994 onward

Source: Dataquest (June 1995)

SEI\yiM-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

26

Semiconductor Equipment, Manufacturing, and IVIaterials Woridwide

Table 4-14

Each Company's Revenue from Shipments of Merchant Silicon Wafers" and Epitaxial Wafers to Asia/Pacific-ROW, 1990-1994

(End-User Revenue in Millions of U.S. Dollars)

1990

1991

1992 1993 1994

CAGR (%)

1990-1994 Company

United States-Based Companies

Crysteco

Epitaxy Inc.

UnisU

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sximitomo Silix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

EpisU

Posco-Huls

Siltron

Other ROW Companies

Total

6.4

37.0

0

2.3

50.7

3.2

0

1.8

0

3.0

0

25.7

10.0

38.7

53.8

0.7

1.5

25.9

0

81.9

0

5.1

0.0

0.5

5.6

2.8

0

1.9

0

12.5

42.3

0

2.4

61.9

77.0

0.5

1.9

23.7

0

103.1

0

5.1

2.5

0.5

8.1

3.2

0

34.9

12.5

50.6

114.4

0.5

2.2

31.9

0

149.0

2.4

0

45.1

8.2

55.7

18.3

43.8

0

2.7

68.3

1.5

0

2.0

0

0.6

4.7

2.5

0.6

8.4

1.4

1.4

19.2

0

24.9

54.3

0

3.1

104.3

132.6

5.6

2.6

39.5

0

180.3

3.0

17.3

56.5

4.2

81.0

1.0

5.7

3.9

0.2

10.8

0.0

7.6

43.2

2.7

31.5

148.0

0

4.0

237.0

111.6

5.0

3.2

49.2

0

169.0

2.8

96.8

84.0

7.5

191.1

0.2

6.8

5.3

0.8

13.1 23.7

47.0

19.9

49.1

Total Sales—Asia/Pacific-ROW 176.9 223.7 281.4 376.4 610.2

36.3

'Includes prime, test, and monitor wafers

2

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEMIVI-WW-IVIS-GSOl ©1995 Dataquest Incorporated

June 19,1995

1994 Silicon Wafer IVlarket Share Estimates

27

Table 4-15

Each Company's Revenue from Shipments of Merchant Silicon Wafers* to

Asia/Pacific-ROW, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992 1993 1994

CAGR (%)

1990-1994 Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Siomitomo SiTix Silicon

Shin-Etsu Handotai

Showa Deriko

Toshiba Ceramics

Total

Europe-Based Compaiues

MEMC Electronic Materials

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

5.7

36.7

0

0.8

48.0

3.2

0

1.6

0

0

0

25.7

10.0

35.7

53.0

0.7

1.5

24.5

0

79.7

0

0

0

0.2

0.2

2.8

0

1.7

0

7.1

42.0

0

0.8

54.4

72.5

0.5

1.9

20.4

0

95.3

0

0

34.9

12.5

47.4

0

0

2.5

0.2

2.7

108.2

0.5

2.2

27.5

0

138.4

12.1

43.5

0

2.3

61.2

1.5

0

1.8

0

0

0

45.1

8.2

53.3

0.6

0

2.5

0.2

3.3

1.4

1.3

18.8

0

16.6

54.0

0

3.0

95.1

125.0

5.6

2.6

30.6

0

163.8

0

17.3

56.5

4.2

78.0

1.0

0

3.9

0.2

5.1

0

7.6

42.9

2.7

19.0

147.3

0

4.0

223.5

100.6

5.0

3.2

36.2

0

145.0

0

96.8

84.0

7.5

188.3

0.2

0

5.3

0.3

5.8 132.1

46.9

16.1

51.5

Total Sales—Asia/Pacific-ROW 163.6 199.8 256.2 342.0 562.6 36.2

•Includes prime, test, and monitor wafers

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward; DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEIVIM-WW-MS-9501 ©1995 Dataquest incorporated June 19,1995

28

Semiconductor Equipment, IVIanufacturing, and Materials Worldwide

Table 4-16

Each Company's Revenue from Shipments of Merchant Epitaxial Wafers to

Asia/Pacific-ROW, 1990-1994 (End-User Revenue in Millions of U.S. Dollars)

1990 1991 1992

1993

1994

CAGR (%)

1990-1994

Company

United States-Based Companies

Crysteco

Epitaxy Inc.

Unisil

Other U.S. Companies

Total

Japan-Based Companies

Kawatec

Komatsu Electronic Metals

Mitsubishi Materials Silicon

NSC Electron Corporation

Sumitomo Silix Silicon

Shin-Etsu Handotai

Showa Denko

Toshiba Ceramics

Total

Europe-Based Companies

MEMC Electronic Materials^

Okmetic

Topsil

Wacker

Other European Companies

Total

ROW-Based Companies

Episil

Posco-Huls

Siltron

Other ROW Companies

Total

0.7

0.3

0

1.5

1.7

0

0

0.2

0

3.0

0

0

0

3.0

0.8

0

0

1.4

0

2.2

0

5.1

0

0.3

5.4

5.4

0.3

0

1.6

7.5

0

0

0.2

0

4.5

0

0

3.3

0

7.8

3.2

0

0

0

3.2

0

5.1

0

0.3

5.4

0

4.7

0

0.4

5.1

0

0

0.2

0

6.2

0.3

0

0.4

7.1

6.2

0

0

4.4

0

10.6

2.4

0

0

0

2.4

0

0.1

0.4

0

8.3

0.3

0

0.1

9.2

3.0

0

0

0

3.0

7.6

0

0

8.9

0

16.5

0

5.7

0

0

5.7

0

0

0.3

0

12.5

0.7

0

0

13.5

2.8

0

0

0

2.8

11.0

0

0

13.0

0

24.0

0

6.8

0

0.5

7.3

7.8

49.5

81.7

-1.7

Total Sales—Asia/Pacific-ROW 13.3 23.9

25.2 34.4 47.6 37.5

U.S. Semiconductor included from 1987 onward; Cincinnati Milacron included from 1989 onward

Monsanto Electronic Materials included from 1989 onward DNS Electronic Materials included from 1989 onward; Kawatec included from 1994 onward

Source: Dataquest (June 1995)

SEIVIIVI-WW-IVlS-9501

©1995 Dataquest Incorporated June 19,1995

1994 Silicon Wafer Market Share Estimates 29

Section 5. Wafer Pricing

Dataquest conducts a survey of the average polished and epitaxial wafer prices during the first calendar quarter each year. The survey covers pricing trends in the United States, Europe, Japan, Korea, and Taiwan and is a direct survey of wafer suppliers and brokers, both formal and informal. The raw data is averaged by region and provides part of the correlation among silicon wafer revenue, size distribution, and silicon area consumption.

Tables 5-1 and 5-2 represent the average prices used for regional prime, test, and epitaxial wafers for five different wafer sizes where available.

These prices should not be used as a benchmark for price comparisons among suppliers.

Wafer pricing is a complex process involving many specifications and variables. In addition, our survey methodology is strict enough only to gauge approximate values and trends for use as a research tool and guideline. In past publications of the market share estimates, Dataquest has published the restdts of the current year's survey (in this case first quarter, 1995). To prevent their use as a benchmark for price comparisons, we have decided to delay the publication of these results imtil next year's database update.

SEi\/ilVl-WW-MS-9501 ©1995 Dataquest Incorporated June 19,1995

30

Semiconductor Equipment, IVIanufacturing, and l\/laterials Worldwide

Table 5-1

1 9 9 0 - 1 9 9 4 P r i c e p e r W a f e r ( U . S . D o l l a r s a n d J a p a n e s e Y e n )

Area

(square inches) 1990

1991

Wafer Diameter

North America: polished CZ ($)

3-inch

100mm

125mm

150mm

200mm

North America: epi ($)

3-inch

100mm

125mm

150mm

200mm

Japai\: polished CZ (yen)

3-inch

100mm

125mm

150mm

200mm

Japan: epi (yen)

3-inch

100mm

125mm

150mm

200mm

Europe: polished CZ ($)

3-inch

100mm

125mm

150mm

200mm

Europe: epi ($)

3-inch

100mm

125mm

150mm

200mm

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.25

9.98

18.75

31.50

121.05

24.39

32.30

51.50

83.25

204.25

1,525

1,805

3,410

6,100

27,450

5,127

5,700

10,000

18,754

NA

8.05

11.75

19.03

32.45

123.39

25.01

35.07

57.02

88.31

NA.

7.10

9.96

18.80

32.44

109.80

24.39

32.30

51.50

83.25

195.75

1,525

1,870

3,320

6,100

22,470

5,127

5,700

10,000

18,754

NA

8.05

11.80

21.35

33.09

115.40

23.70

34.69

57.09

86.75

NA

1992

7.04

10.47

19.05

31.75

104.71

24.39

32.30

51.50

83.25

181.40

1,525

1,865

3,300

5,900

19,750

5,127

5,600

9,875

16,750

N A

8.05

11.40

21.30

33.15

115.75

24.01

33.40

55.41

87.40

NA

NA

27.90

48.85

81.54

192.50

1,310

1,510

3,100

5,400

19,600

1993

7.04

10.91

19.05

32.44

101.93

3,700

4,300

8,700

16,500

38,750

8.05

11.66

21.07

33.40

104.71

8.05

11.50

20.00

34.30

122.00

24.10

33.40

55.41

84.50

NA

24.10

33.00

53.50

83.00

213.00

(Continued)

1994

7.04

10.70

18.80

32.00

118.00

NA

26.50

47.75

81.00

205.00

1,310

1,400

2,740

4,700

15,200

3,700

4,000

8,500

16,500

28,500

SEl\/IM-WW-iVlS-9501

©19951 Dataquest Incorporated June 19,1995

1994 Silicon Vtefer Market Share Estimates

31

Table 5-1 (Continued)

Regional Average Selling Price of Polished and Epitaxial Wafers, at Start of Year

1990-1994 Price per Wafer (U.S. Dollars and Japanese Yen)

Area

(square inches) 1990 1991 1992

1993

Wafer Diameter

Asia/Pacific: polished CZ ($)

3-inch

100mm

125mm

150mm

200mm

Asia/Pacific: epi

($)

3-inch

100mm

125mm

150mm

200mm

NA = Not applicable or not available

Source: Dataquest (June 1995)

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.25

9.50

17.50

29.12

N A

25.10

34.71

53.87

86.23

NA

7.25

10.66

18.63

33.75

N A

25.10

35.05

53.87

85.70

N A

7.25

10.47

19.61

34.95

N A

25.10

35.00

51.90

83.20

NA

7.25

10.47

19.34

34.54

101.00

N A

35.50

52.00

82.00

N A

1994

7.25

10.50

19.22

34.90

115.00

NA

34.50

50.75

81.00

NA

Table 5-2

Regional Average Selling Price of Polished and Epitaxial Wafers, at Start of Year

1990-1994 Price per Square Inch (U.S. Dollars and Japanese Yen)

Area

(square inches) 1990 1991 1992 1993

Wafer Diameter

North America: polished CZ ($)

3-inch

100mm

125rrun

150mm

200mm

North America: epi ($)

3-inch

100mm

125mm

150mm

200mm

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

3.45

2.65

2.71

3.04

4.19

1.03

0.82

0.99

1.15

2.49

1.00

0.82

0.99

1.18

2.25

3.45

2.65

2.71

3.04

4.02

3.45

2.65

2.71

3.04

3.72

1.00

0.86

1.00

1.16

2.15

N A

2.29

2.57

2.98

3.95

1.00

0.90

1.00

1.18

2.09

1994

1.00

0.88

0.99

1.17

2.42

N A

2.18

2.51

2.96

4.21

(Continued)

SEMM-WW-MS-9501 ©1995 Dataquest incorporated

June 19,1995

32

Semiconductor Equipment, Manufecturing, and Materials Worldwide

Table 5-2 (Continued)

Regional Average Selling Price of Polished and Epitaxial Wafers, at Start of Year

1990-1994 Price per Square Inch (U.S. Dollars and Japanese Yen)

Area

(square inches) 1990 1991

1992 1993

Wafer Diameter

Japan: polished CZ (yen)

3-inch lOOmm

125mm

150mm

200mm

Japan: epi (yen)

3-inch

100mm

125mm

150mm

200mm

Europe: polished CZ ($)

3-inch

100mm

125mm

150mm

200mm

Europe: epi ($)

3-inch

100mm

125mm

150mm

200mm

Asia/Pacific: polished CZ ($)

3-inch

100mm

125mm

150mm

200mm

Asia/Pacific: epi ($)

3-inch

100mm

125mm

150mm

200mm

NA = Not applicable or not available

Source: Dataquest (June 1995)

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

7.07

12.17

19.02

27.39

48.70

1.03

0.78

0.92

1.06

NA

3.54

2.88

3.00

3.22

NA

1.14

0.97

1.00

1.18

2.53

725

468

526

685

N A

3.55

2.85

2.83

3.15

N A

216

148

179

223

564

3.35

2.85

3.00

3.17

NA

725

468

526

685

NA

1.03

0.88

0.98

1.23

NA

1.14

0.97

1.12

1.21

2.37

3.55

2.88

2.83

3.13

NA

216

154

175

223

461

3.40

2.74

2.91

3.19

N A

1.03

0.86

1.03

1.28

NA

1.14

0.94

1.12

1.21

2.38

3.55

2.88

2.73

3.04

NA

725

460

519

612

NA

216

153

174

215

406

N A

2.92

2.73

2.99

NA

1.03

0.86

1.02

1.26

2.07

523

353

457

602

796

1.14

0.96

1.11

1.22

2.15

3.41

2.74

2.91

3.09

NA

185

124

163

197

402

1994

1.14

0.94

1.05

1.25

2.51

3.41

2.71

2.81

3.03

4.37

185

115

144

172

312

523

329

447

602

585

1.03

0.86

1.01

1.27

2.36

NA

2.83

2.67

2.96

NA

SEMM-WW-MS-9501

©1995 Dataquest Incorporated

June 19,1995

13 o

•3

i

@

CO

CO

O l o

X3

en

CO

<o

CJI

C 3

3

CD

Table 6-1

Silicon Wafer Plant Expansions/New Lines Since 1990

C o m p a n y

Shin-Etsu Handotai

Sumitomo SiTix

Location

Shirakawa

Isobe

Nagano

Naoetsu

Mimasu

Shirakawa

Camus, OR

England

Vancouver, WA

Malaysia

Vancouver, WA

Japan

Imari

Imari

MainvilJe, O H

Mainvilie, OH

Fremont, CA

Albuquerque, NM

Imari

Mitsubishi Materials Silicon Noda

Yonezawa

Noda

Central Reseasdit'

Ikuno

Chitose

Ikuno

Status

R&D

Epi Expansion

New Volume Production Line

New Volume Production Line

Polishing Line

8-Inch Volume Production

S-Inch Volume Production

6-inch Volume Production

Epi Expansion

8-Inch Volume Production

8-Inch Expansion

8-Inch Expansion

No. 3 Volume Productiorv I>istS;

Expansion of No. 3 Line

Expansion

No. 2 Plant {Substrate Prffljuction)

Epi Expansion

8-lnch Epitaxial Wafer Line

Expansion of No, 3 Line

Pilot Line

Volume Production

R&D for 4M

R&D for 16M

8-Inch Volume Produclt^it

Epi Production Line

Expansion

6,8

6 , 8

6

6

6

8

8

8

6

6

Size

(inches)

Initial

Capacity

(K Wafers/mo.)

Sta

Da

3 /

30

10

200

100/200

50/90

40/100

2 /

3 /

4 /

4 /

2 /

Q l /

19

1994

1994

1994

8

8

8

6

8

8

8

130

170

20

60

19

1994

Q l /

3 /

19

1 /

19

5

250

20

20

9 /

Q l /

4 /

19

Q l /

7 /

19

CO

CO

CJI

CA)

CO

I

CO en

C3

@

CO

CO

en

O

S.

J3 c:

</>

3"

Table 6-1 (Continued)

Silicon Wafer Plant Expansions/New Lines Since 1990

Company

Mitsubishi Materials Silicon

(Continued)

Location

Ikuno

Salerri/ Oregon

To Be Announced

Komatsu Electronic Metals Nagasaki

Miyazaki

Hiratsuka

Portland, Oit:

Taiwan

Toshiba Ceramics

Showa Denko

NSC Electron

Yamagata

Central Resraidi

Niigata

Tokuyama

Niigata

Oguni

Chichibu

Chichibu

Hikari

Hikarl

Status

Expansion

Prime and Epi

Prime

8-Inch Expansion

8-lnch Expansion

Technical Center

Delayed

]V with Formosa Plasfiicgi(8 Inch

Line)

Expand at Okuni Plant

Pilot Line

Volume Production Line

Epi Expansion

Bxpansion - pulling/polishing

8-Inch Expansion

Expansion

Expansion

Expansion

Expansion

Initial

Size Capacity Start

(inches) (K Wafers/mo.) Date

10/35

60/100

20

1995/6

1996/7

1998

6,8

6

8

8

50

10

300

100

90

10

30

10

10

1995

1995

Q3/91

9/90

7/93

1994

1994

1995

7/91

1997

4/93

1996

MEMC

St. Peters, MO

St. Peters, MO

Utsonomiya

Expansion

Expansion

Volume Produ^Jtfen Line

30

30

30

11/91

1995

1995

CD

CD

CJl

i

CO

CO

CJl

CO

I

to

CJl

C 3

@

C D

C D

cn

5?

.o

< =

CD

= J o o

• 3 o

Table 6-1 (Continued)

Silicon Wafer Plant Expansions/New Lines Since 1990

Company

Posco-Huls

Wacker-Chemitronic

Location.

Korea

Korea

Korea

Korea

Status

Volume Production Line

Expansion

Expansion

Expansion

Wasserburg, Germany

Portland, OR

U.S. and Germany

Portland, OR

Wasserburg, Germany

Expansion of Epi

Upgrade Facility

Expansion of Production

Line 2

8-Inch Expansion

MEMC-China Steel JV Hsinchu Park, Taiwan Volume Production

MEMC Southwest (MEMC-

Texas Instruments JV)

Source: Dataquest (June 1995)

Sherman, TX

New 8-lnch Line and U;

:es

Initial

Size Capacity Start

(inches) (K Wafeis/mo.) Date

6,8 3/92

8

8

30

90

1994

1995

8 40 1996

6

6 , 8

8

8

8

75/125

50

6/90

1993

1994

1996/8

1995

6,8 125/215

100/200

1996/7

1997/8

For More information...

Clark Fiohs, Senior Industry Analyst (408) 437-8375

Internet address [email protected]

Via fax (408)437-0292

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our clients. Individual companies reported on and analyzed by Dataquest may be clients of this and/or other Dataquest services. This information is not furnished in connection with a sale or offer to sell securities or in connection with the solicitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their families may, from

0 3 t 3 0 l I C S l '™^ '° '™'^' ^*ve a long or short position in the securities mentioned and may sell or buy such securities.

©1995 Dataquest Incorporated—Reproduction Prohibited

TtKDunKWadstmtconwntion Dataquest is a registered trademark of A.C. Nielsen Company

DataQuest' a company of

The Dun & Bradstreet Corporation

Corporate Headquarters

Dataquest Incorporated

1290 Ridder Park Drive

San Jose, Galifomia 95131-2398

United States

Phone: 1408437-8000

Facsimile: 1408437-0292

Dataquest Incorporated

Nine Ttechnology Drive

Westborough, Massachusetts 01581-5093

United States

Phone: 1-508-871-5555

Facsimile: 1-508-871-6180

Dataquest Global Events

3990 Westerly Place, Suite 100

Newport Beach, Galifomia 92660

United States

Phone: 1-714476-9117

Facsimile: 1-714476-9969

European Headquarters

Dataquest Europe Limited

Holmers Farm Way

High Wycombe, Buckinghamshire

HP12 4XH

United Kingdom

Phone: 44-1494422722

Facsimile: 44-1494422742

Dataquest GmbH

Kronstadter Strasse 9

81677 Munchen

Germany

Phone: 49-89-930-9090

Facsimile: 49-89-930-3277

Dataquest Europe SA

Immeuble Defense Bergeres

345, avenue Georges Q6menceau

TSA 40002

92882 - Nanterre GTC Gedex 9

France

Phone: 33-141-35-13-00

Facsimile: 33-141-35-13-13

®1995 Dataquest Incorporated

SEMM-WW-MS-9501

Ms. Maria Valenzuela

Dataquest Incorporated

1-1100

--INTERNAL DIST.—

Japan Headquarters

Dataquest Japan K.K.

Shinkawa Sanko Building, 6th Floor

1-3-17, Shinkawa, Ghuo-ku

'Ibkyol04

Japan

Phone: 81-3-5566-0411

Facsimile: 81-3-5566-0425

Asia/Pacific Headquarters

7/F Ghina Underwriters Gentre

88 Gloucester Road

WanGhai

Hong Kong

Phone: 852-2824-6168

Facsimile: 852-2824-6138

Dataquest Korea

Suite 2407, Trade Tbwer

159 Samsung-dong, Kangnam-gu

Seoul 135-729

Korea

Phone: 822-551-1331

Facsimile: 822-551-1330

Dataquest Taiwan llF-2, No. 188, Section 5

Nan King East Road

Taipei 105

Taiwan, R.O.G

Phone: 8862-756-0389

Facsimile: 8862-756-2663

Other research offices in Beijing,

Singapore, and Thailand

Sales agents in Australia, Belgium,

Germany (Kronberg), Israel, Italy,

South Africa, and Spain

Qty:

DataQuest

_ co^^

DO Not Remove

l\/lidyear 1995 Forecast: Capital

Spending, Wafer Fab Equipment, and Silicon

Market Trends

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MT-9501

Publication Date: July 31,1995

Filing: Market Analysis

A/lidyear 1995 Forecast: Capitai

Spending, Wafer Fab Equipment, and Siilcon

Market Trends

Program: Semiconductor Equipment, Manufacturing, and Materials Worldwide

Product Code: SEMM-WW-MT-9501

Publication Date: July 31,1995

Filing: Market Analysis

Midyear 1995 Forecast: Capital Spending, Wafer Fab Equipment, and Silicon

Table of Contents

Page

1. Executive Summary 1995: Deja vu. It's 1994 All Over Again 1

Dataquest Perspective 3

2. Semiconductor Capital Spending Forecast 5

Highlights 5

Capital Spending Tables 6

And the Spending Binge Continues 7

When Will the Spending Boom End? 9

The North American Market Continues to Exhibit Strategic

Strength 10

Japan: DRAM Capacity Additions Drive Spending, but a Strong

Yen Subdues 11

Europe Sustains Presence as a Growth Market 12

Asia/Pacific Madly Investing in Two Distinct Ways 12

Who's Investing Where? 14

Dataquest Perspective 15

3. Wafer Fab Equipment Spending Forecast 17

Highlights 17

Annual Investment Themes for 1994 to 1998 18

When Will Capacity Expand to Meet Demand? An Update to the Over- or Underinvestment Model 18

Highlights of Key Equipment Segment Markets and Forecasts .... 24

Steppers and Track 24

Etch and Clean: Dry Etch and Chemical Mechanical

Polishing (CMP) 25

Deposition: CVD, PVD, and Silicon Epitaxy 26

Diffusion and RTP 27

Ion Implantation 27

Segments and Tools of Emerging Importance 28

Dataquest Perspective 29

4. SiUcon Wafer Forecast 31

SiUcon Forecast Tables 31

The 200mm Wafer Ramps Up 31

What about 300nMi Wafers? 31

Highlights of the North American Silicon Wafer Market and

Forecast 33

Highlights of the Japanese SiUcon Wafer Market and Forecast 35

Highlights of the European Silicon Wafer Market and Forecast... 38

Highlights of the Asia/Pacific-ROW Silicon Wafer Market and

Forecast 39

Dataquest's First Silicon Revenue Forecast 40

Dataquest Perspective 42

5. Semiconductor Consumption Forecast 43

Semiconductor Consumption 43

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide

Table of Contents (Continued)

Page

6. Semiconductor Production Forecast 45

Historical Semiconductor Production 45

Captive Semiconductor Production 45

The Start of a Trend: Europe a Production Mecca? 46

Semiconductor Production Trends: Accelerating Shift to Asia/

Pacific 47

Dataquest Perspective 48

Appendix A —Economic Assumptions, Second Quarter of 1995 51

Appendix B —Exchange Rates 55

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Midyear 1995 Forecast: Capital Spending, Wafer Fab Equipment, and Silicon ii[

List of Figures

Figure Page

3-1 Net Cumulative Over- and Underinvestment of Semiconductor

Wafer Fab Equipment 23

3-2 Net Cumulative Over- and Underinvestment of Semiconductor

Wafer Fab Equipment as a Percentage of Wafer Fab Equipment

Market 23

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

jv Semiconductor Equipment, Manufacturing, and Materials Worldwide

List of TaMoo

Table Page

2-1 Semiconductor Capital Spending — Top 20 Spenders, 1994 versus

1993 6

2-2 Semiconductor Capital Spending—Top 20 Spenders, 1995 versus

1994 7

2-3 Worldwide Capital Spending by Region, 1987 to 1994 8

2-4 Worldwide Capital Spending by Region, 1994 through 2000 8

2-5 Regional Investment Patterns of Semiconductor Manufacturers in 1994 14

2-6 Regional Investment Patterns of Semiconductor Manufacturers in 1995 15

3-1 Worldwide Wafer Fab Equipment Market by Region, 1988-1994... 19

3-2 Worldwide Wafer Fab Equipment Market by Region, 1994-2000... 19

3-3 Wafer Fab Equipment Revenue by Equipment Segment,

1988-1994 20

3-4 Wafer Fab Equipment Revenue Forecast by Equipment

Segment, 1994-2000 21

3-5 Annual Driving Forces and Investment Themes for Wafer Fab

Equipment, 1995 to 2000 22

4-1 Forecast of Captive and Merchant Silicon and Merchant

Epitaxial Wafers by Region 32

4-2 Forecast Growth Rates of Captive and Merchant Silicon and

Merchant Epitaxial Wafers by Region 33

4-3 Forecast of Captive and Merchant Silicon Wafers by Region 34

4-4 Forecast Growth Rates of Captive and Merchant Silicon by

Region (Percent of MSI) 35

4-5 Forecast of Merchant Prime and Test/Monitor Wafers by

Region 36

4-6 Worldwide Wafer Size Distribution Forecast, 1993-2000 37

4-7 North American Wafer Size Distribution Forecast, 1993-2000 38

4-8 Japanese Wafer Size Distribution Forecast, 1993-2000 39

4-9 European Wafer Size Distribution Forecast, 1993-2000 40

4-10 Asia/Pacific-ROW Wafer Size Distribution Forecast, 1993-2000.... 41

4-11 Worldwide Merchant Silicon Wafer Revenue Forecast,

1992-2000 42

5-1 Worldwide Semiconductor Consumption by Region 43

5-2 Worldwide Semiconductor Consumption by Region 44

6-1 Worldwide Semiconductor Production by Region, 1988-1994 46

6-2 Worldwide Semiconductor Production by Region, 1994-2000 48

A-1 G D P / G N P Growth Rates with Constant Prices and Exchange

Rates, Local Currencies 51

A-2 Consumer Price Index Growth Rates 52

B-1 Average 1994 and 1995 Exchange Rates per U.S. Dollar 55

SEMM-WW-MT-9501 ©1995 Dataquest incorporated July 31,1995

Chapter 1

Executive Summary 1995: Deja vu,

It's 1994 All Over Again

We knew a few months ago that the 1995 market would surprise us with its strength. A monthly leading indicator w e have been developing went crazy in April and has not rested since. Our May capital spending survey confirmed our indicator — that 1995 is basically 1994 all over again. Two major factors have contributed to this, both unanticipated six months ago.

First, Intel's n e w aggressiveness in pricing for the Pentium has accelerated the transition toward Pentium PCs and the chips that support them. We have seen phenomenal semiconductor book-to-bill numbers in the first half of this year. Taking into consideration a revision of our PC tmit forecast to a 16 percent compound annual growth rate (CAGR), Dataquest made a change in its longer-term view of the semiconductor industry, raising the basic CAGR to over 16 percent (we had been in the 13 to 14 percent range).

Second, the continued lack of economic yields for the l M x l 6 version of the

16Mb DRAM has kept the industry "stuck" at producing a higher number of 4Mb units, which have lower bit-per-square-inch densities, and driven the industry to add square inch capacity (that is, capital equipment) at higher rates than anticipated. From what w e can see now, there will be plenty of equipment that could be brought to bear on 16Mb DRAM capacity by year-end 1996, and on line to answer demand through 1997. A marked downturn in the DRAM investment cycle will be triggered by the l M x l 6 configuration of the 16Mb DRAM achieving yield in the 60 to

65 percent area, expected to occur sometime in 1996.

Desktop connectivity products, telecommunications, and the PC market will lead to stable growth in microcomponents and logic devices, giving strategic strength to the North American region. Japan wiU be concentrating on ramping memories to try and hold its market share against the

Korean and now the Taiwanese. A struggling economy will keep capitcd investment muted once the DRAM ramp is satisfied. Globalization strategies will benefit investment in both Europe and Asia/Pacific, the two fastest-growing regions for the next six years.

Dataquest has been bullish on the prospects for Europe, and 1995 is no different. Europe will actually rival Asia/Pacific for tiie fastest-growing region in 1995. European companies are a large part of this expansion, aided by strong domestic economies. Major projects by the multinational manufacturers are also contributing. In the long term, we stiU see Europe as the second fastest-growing region for spending through the decade.

The Asia/Pacific-ROW region will grow at an astronomical pace in 1995, as Korean DRAM expansion accelerates (even more), foundry expansion in Taiwan, Singapore, and others continues to grow, and new DRAM players enter the scene in Taiwan (PowerChip, Vanguard, and Netn Ya).

As these new projects being started in 1995 will continue to consume capital fimds in 1996, we are looking for continued modest growth in 1996.

Asia/Pacific-ROW will continue to be the fastest-growing region through

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated

Semiconductor Equipment, Manufacturing, and Materials Worldwide

this decade. The year 1995 wiU be a watershed year for capital spending, as the Asia/ Pacific region becomes the largest capital spending region in the world, surpassing both Japan and North America.

Dataquest believes that the relatively large capacity expansion phase of

1993 to 1995 (three-year growth of 200 percent) has now exceeded the three-year growth recorded in the 1987-to-1989 expansion. It should be noted, however, that the two periods are different in one key respect: The current period is also experiencing an accelerated long-term growth for the underlying semiconductor industry, driven by a productivity-related

PC boom. This PC boom is expected to continue, so we are not overly alarmed about the magnitude of this cycle. Momentum of investments will make 1996 a healthy growth year in capital spending.

However, we also believe that in 1996 spending will decelerate, causing a flat spending pattern through 1997 and 1998. Although we continue to beUeve the cyclical nature of investment in semiconductor capacity will diminish, it will now require that the PC boom continue to drive the underlying semiconductor growth large enough to dampen the memory component of the cycle. After a two-year flat period, investments should pick u p again in 1999.

The year 1995 will rival 1994 as the peak growth year in the industry, with wafer fab equipment growing 52 percent, and we are forecasting the spending momentum will carry over, resulting in a 22 percent growth in the front-end equipment market for 1996 to $19.9 billion. Segment growth in 1994 and 1995 is being led by DRAM or capital spending-sensitive equipment, with steppers, high-current and high-voltage implant, wafer inspection, and polysilicon etch exhibiting significantly stronger growth than market growth. We expect no major segment declines in 1995, as capacity additions are broad-based and worldwide.

After four strong expansion years from 1993 through 1996, equipment purchases in 1997 should decHne modestly, followed by a relatively flat growth in 1998. Investment in DRAM capacity wHl be curtailed as producers elect to convert their 4Mb DRAM capacity to 16Mb, which adds bit capacity through the instant increase in bits per square inch. Also, many

Japanese DRAM facilities now running 150mm wafers wiU convert to

200mm wafers, further delaying the need for new equipment. DRAMsensitive equipment technologies or capital-intensive segments such as steppers, implantation, diffusion, and polysihcon etch will be affected more than logic-sensitive technologies such as sputtering, epitaxial reactors, RTP, nontube CVD, or metal etch. The next expansion should kick in by 1999, driven by 0.35- to 0.4-micron capacity expansion. The wafer fab equipment market is forecast to be $29.7 billion in the year 2000.

We have factored in an infrastructure investment in equipment for late

1997 through 1999, which will affect the forecast size of the markets positively. This additional investment will be for initial equipment to fiU a number of 300mm fabs to run silicon by 1998 and 1999. However, our outlook for a significant 300mm equipment market will wait until cifter 2000.

Yield enhancement is the trend of the time. Any system technology that can be relatively low-priced and has a direct impact on yield will gain

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Executive Summary 1995: D6j^ vu, It's 1994 All Over Again

immediate acceptance in volume. Areas particularly important that are emerging today are in cleaning technology, photostabilizers, and process control metrology.

The siUcon wafer market, driven by a stronger long-term picture for semiconductors in general, will grow faster over the next six years than the recent past. As the industry transforms into a 200mm baseline, the outlook for silicon wafer manufacturers becomes brighter. Silicon manufacturers have answered the call for 200mm capacity, and the semiconductor market has again responded with the cry "More, more!" We believe silicon manufacturers' ramp plans in 200mm have been strategically and smartly measured since the over-capacity situations of 1985, and after, are still remembered. While there is going to be activity with 300mm wafers, this is expected to be R&D-focused and low volumes until after the turn of the decade.

Dataquest Perspective

Our forecast for capital spending and wafer fab equipment sales during the next six years assumes that the explosive growth in 1995 will carry over into 1996. These are being driven by the PC market, with telecommunications and networking spurring demand for semiconductor chips across a broad spectrum —along with continued tight capacity convincing companies to expand. Our outlook for the future includes a modest decline in equipment spending in 1997, and a relatively flat 1998, before a resumption of double-digit growth in 1999.

The semiconductor industry has a global manufacturing business. Production of semiconductors is constantly shifting among regions, as new capital money is flowing toward areas with relative lower capital cost and higher-growth areas of consumption. Where the PC goes, so go senuconductors. This is true from the perspective of the business forecast, as well as the production line. Europe and Asia/Pacific, with very large capital

\ spending upticks over the last several years and continuing in 1995, wiU t continue to gain share in world production over the next several years.

The shifts and currents in semiconductor production trends mean that equipment and material suppliers will absolutely need a global presence in every sense of the word to remain competitive in the market. Product supply and support can no longer concentrate on local trends, as all major semiconductor companies have made it clear that they are investing on a worldwide basis. We see evidence of these trends as Lam Research and

Watkins-Johnson set u p direct offices in Japan, as Tokyo Electron Ltd.

(TEL) acquires the Varian half of Varian/TEL to go direct into Europe and

North America, and as ASM Lithography accelerates penetration into

Asia/Pacific. Silicon plants are now being strategically placed, such as

SEH's Malaysian plant, Komatsu's joint venture with Formosa Plastics in

Taiwan, and MEMC's joint ventures in both Korea (Posco-Huls) and

Taiwan (Taisil).

Furthermore, the concept of contract manufacturing in semiconductors is clearly here to stay. Equipment and material suppliers could find themselves selling their technical products to an international team from several companies, including the manvif acturer and the designer. However,

SEMI\/I-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

the emergence of the dedicated foundry company, taking ownership of the process and manufacturing flow, will tend to centralize this activity.

Although w e continue to believe the cycHcal nature of investment in semiconductor capacity will diminish, it will take a few more boom-and-bust cycles before the underlying semiconductor growth is large enough to dampen the memory component of the cycle, provided the PC boom is sustained as forecast.

Project Manager: Clark Fuhs

Contributing Analysts: Calvin Chang, Mike Glennon, Daniel Heyler, Sue

Kapoor, Ben Lee, Mario Morales, Nader Pakdaman, Beth Sargent, George

Shiffler, Yoshihiro Shimada, Yasumoto Shimizu, and J.H. Son

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Chapter 2

Semiconductor Capital Spending Forecast

This chapter presents data on worldwide semiconductor capital spending by region. Capital spending in a region includes spending by all semiconductor producers with plants in that region. Components of capital spending are property, plant, and equipment expenditure for front- and backend semiconductor operations.

Chapter Highlights

This chapter will discuss the following highHghts:

• After a strong 53.4 percent growth in 1994, Dataquest has upgraded the

1995 forecast for global semiconductor capital spending to 60 percent growth at $34.8 biUion. Anticipated continued tight capacity, and a strong semiconductor market in 1995, means continued growth continuing into 1996, currently forecast at about 15 percent.

• North America is showing consistent strength with a 44.7 percent growth in 1995. Worldwide demand for desktop connectivity products and telecommunications equipment continue to fuel the investment strategies in U.S.-manufactured semiconductor products, heavily weighted toward logic and ASIC capacity. North American capital spending is expected to moderate in 1996 and 1997 as these investments are absorbed and the U.S. PC market exhibits more normal historic growth patterns, although worldwide PC unit sales will continue at a

16 percent CAGR.

• Japanese companies are continuing to invest in semiconductor capacity to preserve their market share position in memories, although the recent strength in the yen is keeping a Hd on spending enthusiasm.

Japan as a region kept pace with the world in investment in 1994 but will lag the market in 1995 as Japanese companies invest more outside

Japan. Healthy but subdued growth of 34.2 percent is anticipated in

1995, but only 15 percent on a yen basis. Lagging investment patterns in

Japan is expected to continue throughout the decade.

• Dataquest has been bullish on the prospects for Europe, and 1995 is no different. Europe will actually rival Asia/Pacific for the fastest-growing region, with capital spending growing a remarkable 94 percent. European companies are a large part of this expansion, aided by strong domestic economies; major projects by the multinational manufacturers are also contributing. In the long term, we still see Europe as the second fastest-growing region for spending through the decade.

• Following a very strong capital investment growth year of T7 percent in

1994, the Asia/Pacific-ROW region will grow at an astronomical 94 percent in 1995, as Korean DRAM expansion accelerates (even more), foundry expansion in Taiwan, Singapore, and others continue to grow, and new DRAM players enter the scene in Taiwan (PowerChip,

Vanguard, and Nan Ya). As these new projects being started in 1995 will continue to consume capital funds in 1996, w e are looking for about

23 percent growth in 1996. Asia/Pacific-ROW wiU continue to be the fastest-growing region through this decade.

SEMI\/1-WW-MT-9501 ©1995 Dataquest Incorporated

Semiconductor Equipment, Manufacturing, and Materials Worldwide

• 1995 will be a watershed year for capital spending, as the Asia/Pacific region becomes the largest capital spending region in the world, surpassing both Japan and North America.

Capital Spending Tables

A projected Ust of the top 20 semiconductor capital spending companies in

1994 is presented in Table 2-1, with the same list for projected 1995 spending shown in Table 2-2. Capital spending details by region are provided in the next two tables. Table 2-3 shows historical semiconductor capital spending by region for the years 1987 through 1994. Table 2-4 shows the capital spending forecast by region for the years 1994 through 2000. Yearly exchange rate variations can have a significant effect on the interpretation of the 1987 through 1994 data. For more information about the exchange rates used and their effects, refer to Appendix B.

1993

Rank

6

7

1

2

5

3

4

8

12

9

21

17

14

20

23

16

13

10

15

11

Table 2-1

Semiconductor Capital S p e n d i n g - T o p 20 Spenders, 1994 versus 1993

(MilHons of U.S. DoUars)

1994

Rank

9

10

11

12

13

5

6

7

8

3

4

1

2

14

15

16

17

18

19

20

Company

Intel

Motorola

NEC

Samsung

Fujitsu

Hitachi

Toshiba

Texas Instruments

LG Semicon

SGS-Thomson

Hyundai

Mitsubishi

Advanced Micro Devices

IBM Microelectronics

Matsushita

Siemens AG

Sony

Micron Technology

Philips

Sanyo

1994

2,419.0

1,640.0

1,117.3

1,000.0

988.6

969.9

933.1

825.0

800.0

780.0

700.0

690.0

625.0

525.0

513.2

410.0

392.9

387.0

385.0

356.1

1993

1,700.0

1,100.0

696.9

630.0

700.1

775.6

719.4

545.0

400.0

480.0

400.0

449.6

357.0

412.0

168.6

299.0

359.7

184.6

134.0

347.0

Percent

Change

42.3

49.1

60.3

58.7

41.2

25.1

29.7

51.4

100.0

62.5

75.0

53.5

75.1

27.4

204.4

37.1

9.2

109.6

187.3

2.6

Source: Dataquest (July 1995)

Total Top 20 Companies

Total Worldwide Capital Spending

Top 20 Companies' Percentage of Total

16,457.1

21,775.9

75.6

10,858.6

14,194.3

76.5

51.6

53.4

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated

July 31,1995

Semiconductor Capital Spending Forecast

Table 2-2

Semiconductor Capital Spending—Top 20 Spenders, 1995 versus 1994

(MiUions of U.S. Dollars)

11

5

6

7

16

14

8

1994

Rank

1

2

9

4

3

19

13

20

23

12

15

10

18

10

11

12

13

7

8

9

1995

Rank

1

2

3

4

5

6

14

15

16

17

18

19

20

Company

Intel

Motorola

LG Semicon

Samsung

NEC

Hyundai

Fujitsu

Hitachi

Toshiba

Siemens AG

IBM Microelectronics

Texas Instruments

Mitsubishi

Matsushita

SGS-Thomson

Micron Technology

Philips

Advanced Micro Devices

Sanyo

TSMC

1,357.8

1,229.1

1,060.0

1,000.0

975.0

949.0

909.0

850.0

1995

3,538.0

2,250.0

2,125.0

1,975.0

1,643.6

1,500.0

1,419.5

800.0

750.0

696.0

642.2

528.0

1994

2,419.0

1,640.0

800.0

1,000

1,117.3

700.0

988.6

969.9

933.1

410.0

525.0

825.0

690.0

513.2

780.0

387.0

385.0

625.0

356.1

307.0

Total Top 20 Companies

Total Worldwide Capital Spending

Top 20 Companies' Percentage of Total

Note: 1995 numbers are projections.

Source; Dataquest (July 1995)

26,197.2

34,844.7

75.2

16,371.2

21,775.9

75.2

60.0

60.0

The historical numbers presented in Table 2-3 have been restated sUghtly to coincide with our roll-up of the company-by-company database of capital spending. About two years ago, Dataquest began a process to improve the capital spending data collection and methodology. Now that this process is complete, we wish to make our published figures agree with our database. In almost all cases, the changes are very minor.

And the Spending Binge Continues...

After a three-year rest, 1993 started a growth cycle in a big way. After a

22.4 percent growth in senuconductor capital spending in 1993, an acceleration growth year of 53.4 percent followed in 1994. Dataquest has upgraded our forecast for 1995 to peak at 60 percent growth worldwide, based on our most recent capital spending survey.

Percent

Change

46.3

37.2

165.6

97.5

47.1

9.0

106.7

94.8

11.4

80.4

72.0

114.3

43.6

40.0

31.7

158.5

90.5

18.2

37.5

77.1

SEMI\/1-WW-I\/1T-9501 ©1995 Dataquest Incorporated July 31,1995

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 2-3

Worldwide Capital Spending by Region, 1987 to 1994 (Millions of U.S. Dollars)

North America

Percent Growth

Japan

Percent Growth

Japan (Billions of Yen)

Percent Growth

Europe

Percent Growth

Asia/Pacific-ROW

Percent Growth

Worldwide

Percent Growth

1987

2,622

25.9

2,458

33.2

356

15.4

885

15.7

540

23.6

6,505

26.8

1988

3,349

27.7

4,495

82.9

584

64.0

960

8.5

1,033

91.3

9,837

51.2

1989

3,794

13.3

5,360

19.2

740

26.7

1,186

23.5

1,865

80.5

12,205

24.1

1990

4,217

11.1

5,596

4.4

806

8.9

1,560

31.5

1,542

-17.3

12,915

5.8

Note; The data in this table includes nnerchant and captive semiconductor companies.

Source: Dataquest (July 1995)

1991

3,895

-7.6

5,702

1.9

768

-4.7

1,248

-20.0

2,300

49.2

13,145

1.8

1992

4,135

6.2

3,958

-30.6

500

-34.9

1,188

-4.8

2,318

0.8

11,599

-11.8

1993

4,943

19.5

4,315

9.0

480

-4.0

1,730

45.6

3,206

38.3

14,194

22.4

1994

7,178

45.2

6,460

49.7

658

37.0

2,479

43.3

5,659

76.5

21,776

53.4

Table 2-4

Worldwide Capital Spending by Region, 1994 through 2000 (Millions of U.S. Dollars)

North America

Percent Growth

Japan

Percent Growth

Japan (Billions of Yen)

Percent Growth

Europe

Percent Growth

Asia/Pacific-ROW

Percent Growth

1994

7,178

45.2

6,460

49.7

658

37.0

2,479

43.3

5,659

76.5

1995

10,386

44.7

8,669

34.2

758

15.3

4,811

94.1

10,979

94.0

1996

11,460

10.3

9,387

8.3

821

8.3

5,609

16.6

13,481

22.8

1997

11,677

1.9

8,905

-5.1

779

-5.1

5,937

5.8

13,274

-1.5

14,395

8.4

Worldwide

Percent Growth

21,776

53.4

34,845

60.0

39,937

14.6

39,793

-0.4

42,086

5.8

Note: The data in this table includes merchant and captive semiconductor companies,

Source: Dataquest (July 1995)

1998

12,275

5.1

9,164

2.9

801

2.9

6,252

5.3

1999

14,335

16.8

10,251

11.9

897

11.9

7,075

13.2

16,079

11.7

47,740

13.4

2000

17,633

23.0

12,906

25.9

1,129

25.9

8,188

15.7

22,441

39.6

61,168

28.1

CAGR (%)

1994-2000

16.2

12.2

9.4

22.0

25.8

18.8

The continued growth in personal computer unit sales, with increased growth in telecommunications and networking products, has created tremendous demand for a variety of semiconductor components. The wafer fab capacity crimch has continued into all regions, and most semiconductor products, most notably DRAMs and advanced ASICs. The capacity shortage has given rise to a sharp acceleration in capital spending in all areas, with the strongest-growth occurring in DRAM expansion in

Asia/Pacific and now Europe.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Semiconductor Capital Spending Forecast

The big three Korean companies wiU increase spending by an unbelievable 124 percent to a combined $5.6 biUion in 1995, and a mostly new crowd of Taiwanese companies are now entering the DRAM manufacturing business, spending over $1 billion collectively in 1995. Japanese suppliers of memory are increasing investment as well, collectively increased nearly 45 percent to $10.6 bUlion, an equivalent dollar increase as Korean companies' investment. Intel and Motorola still head the list for 1994, as the microprocessor and microcontroller demand continues to be strong.

Equipping new and acquired facilities (in the case of Motorola) wiU continue to drive these companies in 1995. The three Korean companies, with the aforementioned increase in DRAM capacity spending, now occupy the

No. 3, No. 4, and No. 6 spots in the billion doUar spending club, which currently consists of 11 companies, with four near or over $2 billion. NEC,

Fujitsu, Hitachi, Toshiba, and Siemens all make this top 10 as the memory capacity keeps roUing in. In fact, the top spenders from No. 3 through No.

14 are all heavily concentrated in DRAM spending. With the general health of the industry, smaller semiconductor companies in all regions are participating in the capital spending boom in 1995, keeping the concentration of capital spending by the top 20 flat at 75 percent.

TSMC debuts on the top 20 list for 1995, as foundry capacity expansion has now evolved into a major trend. In fact, foundry spending in Asia/

Pacific wiU nearly double in 1995 to an estimated $1.6 bilHon. This industry has transformed into a dedicated niche business and is no longer a specialized way to use excess capacity.

When Will the Spending Boom End?

Our longer-term forecast projects that significant growth in capital spending will spill into 1996 from the sheer momentum with a moderated growth of 14.6 percent, concentrated toward filling new fabs with equipment as 29 (yes, 29) new 200mm fabs are planning to come on line in 1996, as well as five 150mm fabs for a total of 34. This compares with a total of

28 for 1995 (19 of which are 200mm). Dataquest beheves that in mid-1996, a decelerating capital spending picture wiU begin to emerge, as the capacity additions of 1994 and 1995 are ramped. From what we can see now, there is plenty of equipment that could be brought to bear on 16Mb

DRAM capacity by midyear 1996 and on line to answer demeind through

1997.

Overall semiconductor product demand is expected to remain strong through 2000 with a CAGR of 16.3 percent (see Chapter 5), so w e expect that microcomponent capacity wiU start to ramp u p again in 1998, with the next major investment cycle picking u p in 1999. Our model does not currently include significantly more 16Mb DRAM capacity expansion (over what is already being put in place in 1995 and 1996) until 1999. In the two

"pause" years of 1997 emd 1998, w e beHeve DRAM manufacturers will concentrate on converting capacity away from 4Mb toward 16Mb, which increases bits per square inch processed, and then concentrate on shrinks to squeeze value out per square inch before a capital cycle starts again. In addition, in Japan, w e expect that many 4Mb/16Mb fabs currently running 150mm wafers will convert to 200mm wafers, further gaining efficiency and productivity from the sunk capital investment.

SEMM-WW-I\/IT-9501 ©1995 Dataquest Incorporated July 31,1995

10 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Through 2000, we project a five-year worldwide capital spending CAGR of 18.8 percent, shghtly ahead of the semiconductor consumption growth.

We believe that capital spending may be influenced in 1997 and 1998 positively with the facihty construction and purchase of equipment toward the world's first 300mm wafer fab. We have buUt this infrastructure investment in our model.

About a year ago, w e introduced a model that quantifies the over/underinvestment picture for wafer fab equipment and semiconductor capacity.

While the last several years of activity have created and sustained a net underinvestment, and not fully corrected in 1994, this picture is now corrected to create about a 32 percent overinvestment by the end of the year

(see Chapter 3 and Figures 3-1 and 3-2).

The North American iVIarlcet Continues to Exhibit Strategic Strength

Capital spending in North America wiU grow a consistent 44.7 percent in

1995, with most of the investment growth in 1995 coming from U.S. companies connected with ASIC and logic products. We expect capital spending to decelerate gradually through 1997 with a resumed acceleration in

1998, resulting in a CAGR of 16.2 percent for 1994 through 2000. This somewhat lower than the market expansion is because of an emerging trend to shift production into the foundry market, which is concentrated in Asia (see Chapter 6).

The relatively strong growth in capital spending has been driven by the ongoing growth in PCs, teleconuntmications, and networking. These products have seen increasing use as tools to increase the productivity of the workplace. These electronic products, with their increased semiconductor content, have created enormous demand for microprocessors, microcontrollers, SRAM, programmable logic and memory, standard logic, and peripheral controllers. The U.S. companies dominate many of these market segments. These segments combined are expected to maintain fairly stable growth rates over the next few years, with PC growth slowing (yet still maintaining a 16 percent CAGR) and networking and telecommunications expanding. The near-term market for PCs has reaccelerated from Intel's new aggressive Pentium pricing strategy, which has hastened the conversion to the Pentium.

New products and services, such as personal communicators, interactive television, and video-on-demand, provide the potential for enormous growth in semiconductor sales, especially for highly integrated complex logic and signal processing chips that will be the core engines of future systems.

While the strategic strength of the core logic products enables a healthy and flourishing semiconductor production environment, it is also one that is less volatile in capital spending. In these boom years of 1994 and 1995, the North American region grew at somewhat lower than the market rates. This trait will also enable the North American market to grow in capital spending faster than market rates in the slower years such as 1997.

We beUeve companies will strategically invest in capacity to preserve competitive advantage, and any cutbacks are likely to occur in the smaller companies rather than the first-tier manufacturers.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Semiconductor Capital Spending Forecast 11

Capital investments in North America for 1994 mainly focused on equipping new fabs coming on line by both major and small companies. TTie major projects include Intel's Fab 11.2 in New Mexico and expected orders for Fab 12 in Arizona, AMD's Fab 25 in Austin, Texas, Motorola's MOS-13 in Texas and its continued ramp of MOS-12 in Arizona, Cypress' Fab 4 in

Minnesota, SGS-Thomson's new Arizona facility, IBM's expansion in

Burlington (yes, IBM is back!), and the ongoing purchases for Texas

Instruments' DMOS-5 fab in Dallas. Smaller companies are also spurring the growth this year, with Integrated Device Technology, VLSI Technology, Zilog, Atmel, International Rectifier, AMI, and National Semiconductor all bringing on new capacity. Samsung, which has stated an intention to invest $1 bilHon to build a fab in the United States, has yet to pick a site.

We imderstand that it has narrowed the selection to two sites: c3regon or

Texas. We would expect an announcement within the next few months, and we have factored this into our 1998 forecast.

Japan: DRAM Capacity Additions Drive Spending, but a Strong Yen Subdues

Japan's 34.2 percent increase in capital spending in 1995 is only 15.3 percent on a yen basis, as Japanese companies look to invest outside Japan to optimize buying power. We are forecasting a subdued 8 percent growth in capital spending for 1996, factoring in a slight decline in 1997 as the mission will have been accomplished in DRAMs in the near term.

Some of the Japanese electronics giants are experiencing good profit growth, driven by semiconductor operations. The demand for world memory capacity presented an opportunity to grow profits from semiconductors. Investments by Japemese companies will grow by nearly 44 percent in 1995, with an increased amoimt going overseas. However, as long as the Japanese economy is under pressure, Japanese companies will feel a

"patriotic" dedication to invest in Japan, and we see no company spending more than 30 percent of committed investment outside Japan. With the strength of the yen, multinationals are reducing their investment proportion inside Japan as weU.

Although new facilities by Japanese companies wiU come on hne outside

Japan throughout the rest of tiiis decade, DRAM investments inside Japan are really the only driving force today. Beyond 1995, investment increases in Japan will need to come from growing the domestic economy.

Dataquest beheves ein economic recovery in Japan started in 1994, but with slow acceleration, and could stall if the yen does not stabilize. The degree at which companies wiU invest wiU be affected by the strength of this recovery. As we see no real fvmdamental change coming in Japan until very late in the decade, we are forecasting a below-average CAGR of

9.4 percent in Japan for the years 1994 through 2000.

One bright spot could be that a PC boom could emerge in Japan over the next year or two, spawned by the networking infrastructure that is currently being built. This would breathe new hfe into the Japanese semiconductor market, and our forecast would be brightened a bit. However, we believe that even a PC boom would still lead to a growth forecast that is several percentage points below the world average. The fundamentals of a

Japanese production capacity is still too heavily concentrated in DRAMs, with no clear future direction emerged as yet, keeping us from being more optimistic for capital activities in Japan.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

12 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Europe Sustains Presence as a Growth Market

After a higher-than-expected growth bubble of 45.6 percent in 1993, European spending moderated to a slower-than-market growth rate in 1994 as multinationals (such as Intel) substantially completed the majority of their expansions in 1993. The growth of 43.3 percent in 1994 is nonetheless extremely healthy, primarily being fueled by the Europeans themselves, with the ever-present SGS-Thomson, PhiUps expanding in Nijmegen, and

Ericsson equipping its expansion.

Europe continues to attract the capital in 1995, with a true explosion of spending, forecast to grow a whopping 94 percent. Multinationals have reentered, with Motorola upgrading the Scotland fab bought from Digital, the new IBM/Philips venture in Germany, Analog Devices expanding in

Ireland, Texas Instruments expanding again in Italy, and the IBM/

Siemens fab in France continuing to ramp 16Mb DRAMs. In addition,

Siemens' new fab in Dresden, Germany, is a key expansion, pulling

Siemens into the top 10 in capital spending worldwide. NEC and other

Japanese companies' continued commitment to Europe has given the latest boost in investment momentum, and we are looking for continued growth in 1996 of 16.6 percent.

Europe is being viewed as a strategic location for production to take better advantage of European and 16Mb DRAM growth in the future, driven by the PC production boom (see Chapter 6) without the import tariffs. Samsung has announced a fab to come on Hne in Europe during 1998, but as of the writing undecided about the exact location.

With increased multinational presence starting again in 1995 as the economies pick up, and with recent trends of PC production and foundry providers (Newport WaferFab and Tower Semiconductor), w e now expect

Europe to be an above-average investment region in the long term, with a six-year CAGR of 22 percent, the second fastest region in the world.

Asia/Pacific Madly Investing in Two Distinct Ways

The often erratic but sustained semiconductor capital spending growth in the Asia/Pacific region continued at the explosive rate of 76.5 percent in

1994. And if one thought this market could not accelerate from that level, think again —for 1995 becomes the year that the Asia/Pacific region is the largest expansion region in the world, surpassing both Japan and North

America in terms of dollars spent. Our forecast is for 94 percent growth in

1995, and a moderated growth of about 23 percent in 1996 as several new fab projects continue to be built and equipped, and the number of new projects growing. Longer term, we expect Asia/Pacific to exhibit the most aggressive growth in capital spending of any region. Dataquest forecasts a

1994-through-2000 CAGR of 25.8 percent.

Spending in 1995 is coming primarily from two areas, DRAMs and foundry capacity. The Korean conglomerates are continuing their relentless DRAM capacity expansion plans in 1995. Hyundai is installing equipment for its new E-3 project, the third phase of a 200mm wafer, 16Mb

DRAM fab started in 1992. LG Semicon (formerly Goldstar) is expanding its equivalent C2 line, as the agreement with Hitachi for the 16Mb DRAM ramps. LG Semicon is also bringing on line the G-FAB for non-DRAM memory products. Samsung is continuing to spend, ramping Line 6, also a

200mm, 16Mb DRAM facility.

SEMM-WW-MT-9501 ©1995 Dataquest incorporated July 31,1995

Semiconductor Capital Spending Forecast 13

The real story in 1995 are the new Taiwan players. Vanguard International will be bringing on its new DRAM fab later this year, and PowerChip and

Nan Ya have broken ground on DRAM fabs coming on line in 1996. All of these are targeted at 16Mb DRAM running 200inm wafers. Current players such as TI/Acer and MOSel-Vitelic are also increasing their spending with new projects.

Taiwan chip companies TSMC, Macronix, Winbond, and UMC, along with Chartered Semiconductor in Singapore, have added major projects in expansion of foundry capacity; and a new foundry player has emerged in

Thailand — Sub Micron Technology (STI). STI has $1.6 biUion funding for two separate fab lines, the first to come on line in 1996. Separate funding to establish a technology park similar to Hsin-chu in Taiwan exists, with the location to be in or near Bangkok. The combined spending of these companies for foundry (which excludes spending associated with their own products in the case of UMC, Macronix, and Winbond) increased from about $900 milHon in 1994 to about $1.6 bilHon in 1995, continuing at probably higher levels into 1996. The driving force is the changing face of contract manufacturing in semiconductors. Gone are the days when excess fab capacity could support the fotindry business of fabless companies (as well as integrated device manvifacturers (IDMs), which are companies with fabs).

Dataquest estimates that only about 40 percent of the contracted manufacturing of semiconductors originates from fabless companies. The remainder is from IDMs that wish to place manufacturing lower value-added products away from their own facilities to maximize resources and cost.

The last few years have seen the flourishing of the dedicated fotmdry, most of this type of capacity in Asia/Pacific. However, it is still beUeved that the largest concentration of foimdry capacity in the world resides in

Japan, with companies like Rohm, Seiko-Epson, Sharp, and other large integrated companies.

However, the appetite for leading-edge foundries has caused another transformation to occur. With the cost of capital increasing and at a higher level for leading-edge equipment, fotmdry companies such as Chartered have established longer-term contracts with design companies, often with capital infusions toward production equipment. On the buyer's side, companies are now receiving guaranteed capacity for their products. It should be noted that TSMC, the largest f otmdry company in the world, does not prescribe to the concept of capital infusion. This is a similar business model to Solectron's in the electronic equipment marketplace.

The foundry industry is now a strategic industry rather than simply a tactical one. With this transformation nearly complete, we are starting to see the dedicated investment to build new foundry capacity.

In addition to the established semiconductor-producing countries, huge long-term opportunities exist in developing countries Hke China and the Commonwealth of Independent States (formerly the Soviet Union).

Ultimate demand for semiconductor products in those countries could approach demand in superconsumer countries Uke the United States and

Japan. China, in particular, generates a GDP comparable to that of Japan if evaluated on a purchasing power parity basis. U.S., European, and

SEI\/II\/I-WW-I\/IT-9501 ©1995 Dataquest Incorporated July 31,1995

14 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Japanese telecommunications companies are working with the Chinese government to install telephone exchange equipment and digital lines.

Several hurdles must be overcome before either China or Russia becomes a viable market for advanced front-end semiconductor manufacturing.

Technology export restrictions must be eased to allow the construction of relatively advanced fabrication facilities. Foreign suppHers must establish local sales and service centers and need to define their market access.

Financing capability must be estabHshed by the host countries. Solidification of international trade relationships through participation in the

General Agreement of Tariffs and Trade (GATT) must also be established.

China's internal poHtical structure poses a potential barrier to achieving continued most favored nation status with the United States. It will likely take a few years to sort out these issues. Dataquest assumes that semiconductor investment in China could begin to expand in the 1997 time frame

(today, NEC is leading the investment charge), accelerating into the later half of the decade.

Who's Investing Where?

In our capital spending survey recently completed, Dataquest gathered a picture of how money is being spent. Table 2-5 summarizes how companies based in different regions are spending their money abroad for 1994, and Table 2-6 summarizes the same for 1995. About 78 percent of money spent goes into the domestic economy worldwide, and that ratio holds steady for 1994 and 1995.

Asia/Pacific companies have historically placed aU of their investments domestically, but 1994 saw diversification for the first time, and this should continue in 1995. Asia/Pacific companies spent about 96 percent of their money domestically in 1994, and this percentage is expected to drop slightly to 92 percent in 1995. Europeans have been the most aggressive capital exporters historically, placing only 63 percent of their investment inside of Europe. This figure has grown to 68 percent in 1995 as the domestic economies have seen the emergence of some companies that invest mainly in Europe.

Table 2-5

Regional Investment Patterns of Semiconductor Manufacturers i n 1994

(Millions of U.S. Dollars)

North American Companies

Japanese Companies

European Companies

Asia/Pacific-ROW Companies

All Companies

Percent Growth from 1993 (%)

Source: Dataquest (July 1995)

World

8,628.3

7,326.9

1,866.0

3,954.7

21,775.9

53.4

North

America

6,223.7

541.9

277.6

135.0

7,178.1

45.2

Japan

566.3

5,892.9

0.4

0

6,459.6

49.7

Exirope

957.4

308.7

Asia/

Pacific-ROW

880.9

583.5

1,182.7

30.0

2,478.8

43.3

405.3

3,789.7

5,659.4

76.5

Percent of World

Spending (%)

39.6

33.6

8.6

18.2

100.0

'

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Semiconductor Capital Spending Forecast

15

Table 2-6

Regional I n v e s t m e n t Patterns of Semiconductor Manufacturers i n 1995

(Millions of U.S. Dollars)

North Americein Companies

Japanese Companies

European Companies

Asia/Pacific-ROW Companies

All Companies

Percent Growth from 1994 (%)

World

12,294.6

10,579.5

3,112.0

8,858.6

34,844.7

60.0

Source: Dataquest (July 1995)

North

America

8,859.1

918.5

339.4

268.5

10,385.6

44.7

Japan

491.0

8,177.4

0.6

0.0

8,669.1

34.2

Etirope

1,567.0

717.7

2,111.6

A s i ^

Pacific-ROW

1,377.5

765.8

660.4

414.8

4,811.0

94.1

8,175.3

10,979.0

94.0

Percent of World

Spending (%)

35.3

30.4

8.9

25.4

100.0

-

Japanese companies' domestic investment in 1994 is very close to the worldwide average with about 80 percent, dropping to 77 percent in 1995.

North American companies are also high domestic spenders, with about

72 percent staying at home for both years.

The North American and Japanese regions are net investors, while European and Asia/Pacific regions are net beneficiaries of that investment, paralleling the net exporters of semiconductors.

While all regions are spending in Asia/Pacific, and all multinational regions investing in Europe, only North American companies have the strategic vision to invest in Japan. Japanese companies are also investing on a worldwide basis. We believe it is one of the key elements necessary in a strategic plan for a semiconductor company to be competitive on a global basis.

Dataquest Perspective

Capital spending exploded in 1994 and is accelerating in 1995. The major reason for the acceleration is the surprisingly persistent growth in xmit PC shipments, with the aggressiveness of Intel's Pentium pricing strategy.

Major DRAM expansion accelerated in the second half of 1993 and will continue through the first half of 1996. From what we can see now, there is plenty of equipment that could be brought to bear on 16Mb DRAM capacity by year-end 1996 and on line to answer demand through 1997. A marked downturn in the DRAM investment cycle will be triggered by the l M x l 6 configuration of the 16Mb DRAM, achieving yield in the 60 to

65 percent area, expected to occur sometime in 1996.

Desktop connectivity products, telecommunications, and the PC market will lead to stable growth in microcomponents and logic devices giving strategic strength to the North American region. Japan will be concentrating on a ramping memories to try and hold its market share against the

Korean companies, and now the Taiwanese. A struggling economy will keep capital investment muted once the DRAM ramp is satisfied. Globalization strategies wiU benefit both European and Asia/Pacific investment, the latter being the fastest-growing region during the next five years.

SEIVIIVI-WW-IVlT-gSOl ©1995 Dataquest Incorporated July 31,1995

16 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Dataquest believes that the relatively large capacity expansion phase of

1993 to 1995 (three-year growth of 200 percent) has now exceeded the three-year growth recorded in the 1987-to-1989 expansion. It should be noted, however, that the two periods are different in one key respect: The current period is also experiencing an accelerated long-term growth for the underlying semiconductor industry, driven by a productivity-related

PC boom. This PC boom is expected to continue, so w e are not overly alarmed about the magnitude of this cycle. Momentum of investments will make 1996 a healthy growth year in capital spending.

However, we also believe that in 1996 spending will decelerate, causing a flat spending pattern through 1997 and 1998. Although w e continue to beheve the cyclical nature of investment in semiconductor capacity will diminish, it will now require that the PC boom continue to drive the underlying semiconductor growth large enough to dampen the memory component of the cycle. After a two-year flat period, investments should pick u p again in 1999.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Chapter 3

Wafer Fab Equipment Spending Forecast

This chapter presents data on worldwide spending by region for wafer fabrication equipment. Wafer fab equipment spending in a region includes spending by all semiconductor producers with plants in that region. Included are all classifications of equipment for front-end semiconductor operations.

Chapter Highlights

Highlights of this chapter are as follows:

• Wafer fab equipment spending in 1995 will continue at a neck-breaking

52 percent growth worldwide, after a 56 percent growth year in 1994.

• This growth will be driven by Asia/Pacific region (79 percent growth) and a PC production boom in Europe (69 percent wafer fab equipment growth). Continued investment in North America and a DRAMsensitive Japan (40 and 37 percent growth, respectively) rotmd out the third year of the boom.

• Momentum from 1995 and 29 new 200mm fabs coming on line in 1996 will keep growth intact. Dataquest is now forecasting nearly 22 percent growth in wafer fab equipment in 1996.

• Segment growth in 1994 and continued growth in 1995 are being led by

DRAM and capital-spending-sensitive equipment, with steppers, highcurrent and high-voltage implant, wafer inspection, and polysiHcon etch exhibiting significantly stronger growth than market growth. We expect no major segment declines in 1995, as capacity additions are broad-based and worldwide.

• Our model that measures the net cumulative under- or overinvestment indicates that by the end of 1995, the semiconductor manufacturing world will be overinvested in wafer fab equipment to the time of

$5.2 billion, or 31.7 percent of the market. This is above the peaks exhibited in 1984 and 1989, so excess capacity should emerge in 1996, probably in the DRAM market where capacity has been added recently.

• After four strong expansion years in 1993 through 1996, equipment purchases in 1997 should decline modestly, followed by a relatively flat growth in 1998. The next expansion should kick in by 1999, driven by a

0.35- to 0.4-micron capacity expansion. The worldwide wafer fab equipment market is forecast to grow at an 18.4 percent CAGR between 1994 and 2000.

• We have factored in an infrastructure investment in equipment for 1997 through 1999, which will affect the forecast size of the markets positively. This additional investment will be for initial equipment to fill a few 300mm fabs to run silicon by 1998 and 1999. The bulk of this

"300mm. equipment bubble" will occur in 1998. However, our outlook for a significant 300mm equipment market will wait until 2000 and

2001.

This chapter presents historical and forecast data on the worldwide wafer fabrication equipment market, by region emd by key equipment segment.

SEI\/IM-WW-MT-9501 ©1995 Dataquest Incorporated 17

18 Semiconductor Equipment, Manufacturing, and Materials Worldwide

In this midyear forecast outlook on wafer fab equipment, w e have chosen to focus our forecast of equipment categories on specific segments and issues, namely:

• The annual investment theme for 1995 to 2000

• Steppers and automatic photoresist processing equipment (track) in hthography

• Dry etch and chemical mechanical polishing (CMP) in etch and clean

• Silicon epitaxy, CVD, and PVD in deposition

• Diffusion and RTF

• Ion implantation (medium current, high current, and high voltage)

• Segments of emerging importance

These segments of the equipment market represent not only the majority of all wafer fab equipment expenditure in the world today, but also embody the key technological capability for advanced device production.

Highlights of some of the factors affecting individual equipment segment forecasts also are presented.

Equipment spending in a region refers to spending by all companies — both domestic and foreign—within the region. In addition, w e note that annual exchange rate variations can have a significant effect on 1988 through 1994 data appearing in the tables in this chapter. Appendix B details the exchange rates used in this document.

Tables in this chapter provide details on the following:

• Table 3-1 — Historical market data by geographic region for the years

1988 through 1994

• Table 3-2—Forecast market data by geographic region for the years

1994 through 2000

• Table 3-3 —Historical data for key equipment segments for the years

1988 through 1994

• Table 3-4 — Forecast data for key equipment segments for the years 1994 through 2000

Annual Investment Themes for 1994 to 1998

Behind our equipment and segment forecasts are assumptions about how semiconductor producers wiU perform and invest. These are summarized in Table 3-5 for the years 1995 through 2000. The following areas are considered: the availability of profits for reinvestment, memory versus logic growth, technology shifts, and brick and mortar versus equipnient purchases.

When Will Capacity Expand to Meet Demand? An Update to the Over- or

Underinvestment Model

In our forecasts last year, w e introduced a model that provided a measure of the net cumulative over- or underinvestment in wafer fab equipment to support capacity needs in the industry. Since equipment purchases

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Wafer Fab Equipment Spending Forecast

19

Table 3-1

Worldwide Wafer Fab Equipment Market by Region, 1988-1994

(MilUons of U.S. Dollars)

North America

Percent Change

1988

1,534

38.9

1989

1,657

8.0

1990

1,589

-4.1

1991

1,524

-4.1

1992

1,570

3.0

1993

2,129

35.6

1994

3,141

47.5

CAGR (%)

1988-1994

12.7

^ •

Japan

Percent Change

Europe

Percent Change

Asia/Pacific

Percent Change

2,270

77.8

2,813

23.9

2,992

6.4

3,007

0.5

2,096

-30.3

2,460

17.4

3,668

49.1

662

25.9

519

126.6

721

8.9

764

6.0

641

-16.1

820

58.0

522

-36.3

843

61.5

634

-1.1

978

54.3

1,385

41.6

789

-6.4

1,309

65.9

2,562

95.7

8.3

-

13.1

-

30.5

-

Total Wafer Fab Equipment

Percent Change

4,984

58.9

6,011

20.6

Note; Some columns may not add because of rounding.

Source: Dataquest (July 1995)

5,867

-2.4

6,014

2.5

5,089

-15.4

Table 3-2

Worldwide Wafer Fab Equipment Market by Region, 1994-2000

(Millions of U.S. Dollars)

6,876

35.1

10,755

56.4

North America

Percent Change

1994

3,141

47.5

1995

4,409

40.4

1996

5,040

14.3

1997

4,966

-1.5

1998

5,160

3.9

1999

6,065

17.5

13.7

•r-

2000

7,801

28.6

CAGR (%)

1994-2000

16.4

-;

Japan

Percent Change

3,668

49.1

5,008

36.5

5,459

9.0

4,931

-9.7

4,953

0.4

5,643

13.9

7,146

26.6

11.8

-

Europe

Percent Change

1,385

41.6

2,341

69.0

2,740

17.0

2,696

-1.6

2,842

5.4

3,317

16.7

4,002

20.7

Asia/Pacific

Percent Change

2,562

95.7

4,582

78.8

6,615

44.4

6,296

-4.8

6,369

1.2

7,470

17.3

10,751

43.9

Total Wafer Fab Equipment

Percent Change

10,755

56.4

16,340

51.9

Note: Some columns may not add because of rounding.

Source: Dataquest (July 1995)

19,854

21.5

18,888

^ . 9

19,323

2.3

22,495

16.4

29,701

32.0

19.3

-

27.0

-

18.4

-

SEMIVI-WW-I\/IT-9501 ©1995 Dataquest Incorporated July 31,1995

20

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 3-3

Wafer Fab Equipment Revenue by Equipment Segment, 1988-1994

(Millions of U.S. Dollars)

Equipment Segment

Worldwide Fab Equipment ($M)

Percent Change

1988

4,984

58.9

1989

6,011

20.6

1990

5,867

-2.4

1991

6,014

2.5

1992

5,089

-15.4

1993

6,876

35.1

1994

10,755

56.4

CAGR (%)

1988-1994

13.7

-

Steppers

Track

Other Lithography^

Total Lithography/Track

921

253

307

1,481

1,181

322

261

1,764

1,052

317

242

1,612

979

364

205

1,549

646

353

147

1,147

1,014

507

162

1,683

1,838

711

190

2,739

12.2

18.8

-7.7

10.8

Automated Wet Stations

Other Clean Process^

Dry Etch

Dry Strip

Chemical Mechanical Polishing

Total Etch and Clean

Tube CVD

Non-Tube Reactor CVD

Sputtering

Silicon Epitaxy

Other Deposition^

Total Deposition

Diffusion

RTP

Total Diffusion/RTP

Medium Current Implant

High Current Implant

High Voltage Implant

Total Ion Implantation

Total Process Control^

118

241

18

377

608

144

133

533

100

NS

911

243

134

670

121

NS

1,168

268

132

690

118

NS

1,208

291

143

717

119

11

1,281

286

103

682

123

20

1,212

285

198

1,096

138

44

1,761

186

275

260

86

165

972

220

388

320

15

170

1,173

259

457

359

68

153

1,2%

268

474

425

89

147

1,403

213

437

446

84

119

1,300

283

585

584

83

115

1,650

296

22

318

332

25

357

325

33

358

326

46

372

246

36

283

342

45

388

131

301

25

457

676

114

250

7

370

605

108

228

18

353

643

83

164

16

263

544

108

233

18

359

634

15.7

21.9

25.7

2.7

-7.8

17.6

8.9

23.6

10.3

23.8

9.8

19.5

12.4

NA

18.9

12.1

8.4

8.1

9.6

8.9

492

80

572

234

391

29

654

1,017

519

233

1,555

202

65

2,573

446

903

1,025

100

101

2,575

Factory Automation

Other Equipment

Total FA/Other Equipment

130

187

317

195

222

417

216

202

418

227

185

412

194

146

340

250

151

401

412

213

625

21.2

2.2

12.0

Total Wafer Fab Equipment 4,984 6,011 5,867 6,014 5,089 6,876 10,755 13.7

NS = Not surveyed

NA = Not applicable includes contact/proximity, projection aligners, direct-write, maskmaking, and x-ray lithography

2 Includes spray processors, post-CMP clean, vapor phase clean, and other clean processes as defined in our Market Statistics book

3 Includes Evaporation, MOCVD, and MBE

^ Includes optical CD, CD SEM, wafer inspection, and other process control equipment

Note; Some columns do not add to totals shown because of rounding.

Source: Dataquest (July 1995)

SEMM-WW-MT-9501

©1995 Dataquest Incorporated July 31,1995

Wafer Fab Equipment Spending Forecast

21

Table 3-4

Wafer Fab Equipment Revenue Forecast by Equipment Segment, 1994-2000

(MiUions of U.S. DoUars)

Equipment Segment

Worldwide Fab Equipment ($M)

Percent Change

1994

10,755

56.4

1995

16,340

51.9

1996

19,854

21.5

1997

18,888

-4.9

1998

19,325

2.3

1999

22,495

16.4

2000

29,701

32.0

CAGR (%)

1994-2000

18.4

-

Steppers

Track

Other Lithography^

Total Lithography/Track

1,838

711

190

2,739

2,990

1,029

260

4,279

3,306

1,360

316

4,981

2,833

1,366

310

4,509

2,638

1,411

321

4,369

3,217

1,665

353

5,235

4,752

2,109

416

7,277

17.2

19.9

14.0

17.7

Automated Wet Stations

Other Clean Process^

Dry Etch

Dry Strip

Chemical Mechanical Polishing

Total Etch and Clean

519

233

1,555

202

65

2,573

778

341

2,271

297

95

3,783

999

428

2,750

347

141

4,666

999

438

2,663

340

142

4,582

1,034

454

2,822

367

164

4,841

1,102

484

3,399

416

225

5,626

1,393

597

4,529

529

342

7,390

Tube CVD

Non-Tube Reactor CVD

Sputtering

Silicon Epitaxy

Other Deposition^

Total Deposition

Difhision

RTP

Total Diffusion/RTP

Medium-Current Implant

High-Current Implant

High-Voltage Implant

Total Ion Implantation

Total Process Control*

446

903

1,025

100

101

2,575

662

1,389

1,536

134

118

3,838

838

1,702

1,827

169

131

4,666

812

1,662

1,794

185

113

4,567

812

1,706

1,875

213

110

4,715

922

1,980

2,227

241

115

5,484

1,191

2,569

2,851

273

131

7,015

492

80

572

735

123

858

921

165

1,086

861

183

1,045

864

234

1,098

990

256

1,246

1,292

312

1,604

234

391

29

654

324

631

87

1,041

375

784

113

1,273

317

680

93

1,090

300

628

104

1,032

355

742

140

1,237

487

1,054

202

1,744

1,017 1,551 1,954 1,942 2,027 2,250 2,819 18.5

Factory Automation

Other Equipment

Total FA/Other Equipment

412

213

625

650

340

990

824

405

1,229

812

342

1,154

860

383

1,243

990

427

1,417

1,337

517

1,853

21.7

15.9

19.9

Total Wafer Fab Equipment 10,755 16,340 19,854 18,888

19,325

22,495 29,701 18.4

^ Includes contact/proximity, projection aligners, direct-write, maskmaking, and x-ray lithography

2 Includes spray processors, post-CMP clean, vapor phase clean, and other clean processes as defined in our Market Statistics book

3 Includes Evaporation, MOCVD, and MBE

* Includes optical CD, CD SEM, wafer inspection, and other process control equipment

Note: Some columns do not add to totals shown because of rounding.

Source: Dataquest (July 1995)

SEMM-WW-MT-9501

©1995 Dataquest Incorporated

July 31,1995

17.8

19.0

18.6

18.2

4.4

18.2

17.9

17.0

19.5

17.4

31.9

19.2

17.5

25.5

18.8

13.0

18.0

38.0

17.7

22

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 3-5

A n n u a l D r i v i n g Forces a n d I n v e s t m e n t T h e m e s for Wafer Fab Equipment, 1995 to 2000

Logic Semiconductor Unit

Growth*

Investment in Logic Capacity*

Memory Semiconductor Unit

Growth*

Investment in Memory Capacity*

1995

Sohd

Sohd

Sohd

Strong

Front-End Equipment versus

Facihties Loading of Capital

Facihties

Primary Technologies Invested 0.35-0.6 micron

*Scale: Strong > Solid > Moderate > Weak > Dead

Source: Dataquest (July 1995)

1996

Sohd

Sohd

Weak

Moderate

Equipment

0.35-0.5 micron

1997

Moderate

Moderate

Moderate

Weak

Facilities

0.35-0.5 micron

1998

Moderate

Moderate

Moderate

Weak

Balanced

0.3-0.5 micron

1999

Moderate

Moderate

Solid

Moderate

Balanced

0.25-0.4 micron

2000

Solid

Solid

Strong

Strong

Equipment

0.25-0.4 micron precede actual capacity on line by a number of months or quarters, this model could be viewed as a "leading indicator" to capacity shortages and excesses. The results of this model are close to a 1- to 2-year indicator of turning points in the equipment industry. The methodology of the net cumulative investment (NCI) model is linked to our forecast model.

To review, our methodology starts with a few key assumptions and baselines:

• Long-term growth rates for semiconductors and wafer fab equipment are correlated. In other words, semiconductor revenue and profits are needed before money can be spent on equipment, and vise versa.

• Also, NCI equals zero over time, meaning that in a noncyclical environment where annual growth rates are constant, investment and capacity are at equihbrium at aU times. Of course, our industry cycles through over- and imderinvestment.

• The output is a tangible number and is in dollars of over- or underinvestment at year-end. However, the more useful output of the model divides this gross dollar number by the wafer fab equipment market size. The result is a percentage of market figure that is repeatable in levels from cycle to cycle.

• To take into consideration the long-term growth of the semiconductor and equipment industries changing over time, the model has a factor allowing the fundamentals of the industry to change over time.

A net positive or negative investment is calculated relative to the longterm growth baseline annually and then added to the prior year. The calculation resulted in a doUar value net cumulative over- or underinvestment and has correlated well with historical patterns.

Figures 3-1 and 3-2 show the most recent results of the model. In absolute dollar terms, by the end of 1995 the industry will be $5.2 bilhon overinvested, or 31.7 percent of the wafer fab equipment market, exceeding levels witnessed during the 1984 and 1989 peaks. These levels are being driven by two basic causes. First, the PC unit demand is continuing to

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Wafer Fab Equipment Spending Forecast

Figure 3-1

Net Cumulative Over- and Underinvestment of Semiconductor Wafer Fab Equipment

Billions of Dollars

4

3

6

5

m

2

1

0

-1

^

r

^s^m

-^3,

I I

m

P

-2

-3

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

9503948

Source: Dataquest (July 1995)

Figure 3-2

Net Cumulative Over- and Underinvestment of Semiconductor Wafer Fab Equipment as a Percentage of Wafer Fab Equipment Market

Percentage of Wafer Fab Equipment Market

40

30

m

20

10

-10

0

m

W/ m m

^

W'

SS2X

-20

-30

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Source: Dataquesi (July 1995)

23

SEIV11\/I-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

24 Semiconductor Equipment, Manufacturing, and Materials Worldwide

grow at high-teen annual growth. About one-third of the semiconductor industry, and over one-half of the capital spending on new capacity, is to support this demand.

Second, the DRAM market has not yet converted to run the more siliconefficient 16Mb DRAM, placing this investment cycle about seven years behind the last cycle. DRAM bit demand generally runs at 50 percent annually, and DRAM manufacturing has depended upon increasing unit densities (increasing the bits per square inch) to meet this demand.

Shrinks of existing generations alone only bring 15 to 25 percent annual bit-per-square-inch efficiencies, while converting a fab running 4Mb

DRAMs to 16Mb DRAMs would increase bits per square inch by two to three times. As low yields are holding back the economic conversion of the market, top line bit demand is translating to square inch demand (silicon) and the equipment to run it. Because the equipment being installed is fully

"convertible" to run 16Mb DRAMs, w e can think of these fabs building

"pent-up supply" in bits. Once 16Mb DRAM yields (for the l M x l 6 configuration) exceed 60 to 65 percent, it will be more economical to run these hnes, and DRAM prices wiU erode. The current view is that this will not occur until well into 1996.

We have factored into the model an investment in a few 300mm fabs starting in 1997 through 1999, with the bulk in 1998. This is considered an equipment "bubble demand" because the equipment will be shipped into a nonproductive fab (meaning no semiconductor revenue will be initially generated).

With our forecast for a momentum style growth in 1996 and two pause years in 1997 and 1998, the model indicates a reacceleration of equipment spending starting in 1999.

Highlights of Key Equipment Segment IVIarlcets and Forecasts

Steppers and Tracic

From 1989, the peak year of stepper shipments at more than 950 units, the market tumbled to less than 400 tools in 1992 before recovering. During this DRAM-sensitive ramp, we believe the industry will see its first 1,000stepper year. In fact, w e believe over 1,150 steppers wiU be shipped in

1995. Shifts in the product mix toward higher-priced i-hne systems and wide field lenses have also driven u p average selling prices (ASPs). This, along with the strong yen, yields a forecast revenue increase of 63 percent on a dollar basis.

Stepper revenue is forecast to grow at a 17.2 percent CAGR, slightly below the market average for 1994 through 2000. Our forecast for stepper unit growth over the five-year forecast horizon remains modest but higher than what w e have seen in the past, about 7 percent in a CAGR between

1994 and 2000.

With the adoption of phase shift mask technology, off-axis illumination techniques as well as conventional i-line tools with variable NA, i-hne is clearly a viable technology down to the 0.35-micron regime and will continue to dominate the overall stepper technology mix through the 2000.

Excimer/ deep-UV steppers will begin to represent a more significant

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Wafer Fab Equipment Spending Forecast 25

portion of the product mix from 1997 onward for use in 0.25-micron (and below) devices, and ICs with large chip areas such as advanced microprocessors, which require large field size capability. Dataquest beheves that field size pressures accompanied by shrinking geometry will drive the industry toward step-and-scan (or step-and-stitch) technologies for the majority of excimer/deep-UV shipments, beginning in 1997.

Track equipment is forecast to grow at a 19.9 percent CAGR between 1994 and 2000, sKghtly ahead of the industry growth of 18.4 percent. While we believe that the rapid shift in the product mix toward higher priced systems has been recently completed, we do expect another product shift to occur in the track market, and higher ASPs, associated with the ramp of deep-UV steppers, which require more sophisticated environmental control systems.

Etch and Clean: Dry Etch and Chemical Mechanical Polishing

(CMP)

Dataquest began covering the chemical mechanical poHshing (CMP) market in 1993, a year that saw sales increase 121 percent to over $44 million.

Dataquest includes the post-CMP clean system, usually sold in conjunction with a CMP tool as part of the cleaning segment and not in the CMP segment. The year 1994 was expected to be another high-growth year, but the market has really been disappointing—growing at a rate slower than the market average to approach $65 milHon (or 49 percent growth).

The year 1995 holds much the same for the CMP systems, growing about

46 percent to $95 million. We beheve that the early adopters are continuing to buy but, risk averse, have been slow to purchase initial or follow-on systems. Also, the application appears to be Umited to devices with at least three levels of metal and tends to exclude the DRAM market.

These systems are used to remove material from the surface of the wafer to result in a flat surface over the entire wafer. This global planarization, primarily of dielectric layers, is required to achieve high yields in devices where three or more levels of metal are used. Today's advanced logic and

ASIC devices are fueling this market growth, but a key to the near-term success of the market will be how dedicated foundry companies will be to handling their 0.5-micron triple-level metal standard processes. Until recently, most foundries we have talked with will not put CMP into that standard process. With TSMC's recent acceptance of the AMD 486 microprocessors emd its disclosure of a CMP tool purchase, this may change; then again TSMC may only n m it for the AMD wafers.

Dataquest beheves that this technology and market will become a major part of semiconductor manufacturing in the long run but, as with most new technologies, take a bit longer than people would wish. If w e were to make a forecast based purely on technology driving the market, we would not be slowing the CMP market forecast in 1997 to 1998. However, we believe those years wiU see some holding back of capital investment, and history has shown us that even advanced, emerging technologies are rarely spared in a capital slowdown. Nonetheless, CMP is our fastestgrowing segment with a 31.9 percent CAGR for the years 1994 through

2000.

SEI\/1I\/1-WW-MT-9501 ©1995 Dataquest incorporated July 31,1995

26 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Dry etch systems continue to exhibit strong revenue growth, with a CAGR of 19.5 percent forecast for the years 1994 through 2000. Unit shipments are expected to grow as greater multilevel interconnect process capacity is brought on line, increasing the need for dielectric eind metal capacity.

Relatively strong ASP growth will lend additional momentum to dry etch revenue growth as new high-density plasma systems for 0.35-micron applications enter the market, and multichamber cluster tools continue to increase their presence.

Wet process equipment is forecast to show revenue growth of 17.6 percent annually from 1994 through 2000, basically on par with overall equipment market growth. This market is more heavily dependent on the brick and mortar part of the capital investment dollar, as wafer cleaning is a fundamental in aU parts of the fab. Automated wet stations dominate this segment, now accounting for nearly 70 percent of the clean process systems.

Emerging areas of cleaning tools include (in order of size) spray processors, post-CMP clean, and vapor phase cleaning. As manufacturers look to novel ways to increase yields, we would expect these and other new techniques to evolve and emerge in importance.

Deposition: CVD, PVD, and Silicon Epitaxy

CVD equipment sales are predicted to grow at an 18.6 percent annualized rate from 1994 through 2000, on par with overall equipment growth. The steady growth in multilevel interconnect capacity will continue to generate demand for dielectric and metal CVD systems. ASP growth will also contribute to revenue growth, as more highly integrated systems with improved productivity and particle control appear in the marketplace.

Advanced dielectric deposition systems utilizing HDP sources will be introduced for intermetal dielectric (IMD) applications for sub-0.5-micron processes. Most systems will be introduced on multichamber cluster platforms. Metal CVD wiU continue to exhibit strong growth, driven by blanket tungsten for contact and via plugs, CVD barrier metals such as

CVD titanium nitride, and dichlorosllane (DCS) tungsten siHcide for shrink 16Mb and 64Mb DRAMs. For these reasons, the forecast for nontube CVD systems outperforms tube furnaces.

Sputter deposition systems are also forecast to grow at an annualized rate of 18.6 percent for the years 1994 through 2000. As in the case of dry etch and CVD equipment, continued expansion of multilevel interconnect process capacity is the primary driver behind sputter system growth. Rapid growth in average system ASP has helped drive total revenue growth in

1994, primarily from the quick and expanding dominaince of Applied

Materials in the market. With AppHed now accounting for more than

50 percent of the market, the bulk of the ASP increases are behind us.

Revolutionary changes in system architecture, pioneered by the Applied

Materials Endura system, wiU continue to yield improvements in film properties, equipment productivity, and defect density. This is a market segment that will be somewhat buffered from a slowdown in DRAM investment, as the fundamental growth in the number of metal layers in

ASICs and logic devices drive a more stable outlook.

The shift from batch- to single-wafer epitaxial systems has been the primary driver on epitaxial deposition systems, as a result of the need for

200mm epitaxial wafer capacity. In 1994 we saw the Apphed Materials

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Wafer Fab Equipment Spending Forecast 27

system become an "Intel accepted supplier," so silicon companies are investing to add this capacity. However, this capacity is more expensive than wafer suppliers would like, so w e expect the concept of "minibatches" to emerge as a viable production strategy, as it has in CVD.

Moore Technologies is known to be in the processing of releasing such a new product. A strong automotive and discrete market has increased demand for the specialty batch units. An increased product mix of logic semiconductors, sustained demand for discretes, and 200mm wafer capacity addition will be the primary drivers for epitaxial deposition equipment growth beyond 1995.

Diffusion and RTP

Diffusion systems are expected to demonstrate a CAGR of 17.5 percent for

1994 through 2000. This segment is DRAM- and capital spending-sensitive and therefore tends to be cyclic. The years 1993 through 1995 are dramatic growth years, while 1997 will actually witness some decline.

The displacement of horizontal tube systems by vertical tube reactors will continue. Newer vertical systems will be configurable as multitube clusters with integrated dry clean capability to compete with single wafer cluster tools. Tube systems will also incorporate small batch capabilities to offer greater flexibility for custom and semicustom circuit manufacturers.

RTP will grow at an annualized rate of 25.5 percent for 1994 through 2000.

This market grew much faster than we anticipated, nearly doubling in

1994. The real growth for this segment will come from the transition of the thermal "nondepositing" processes away from diffusion tubes and into single-wafer RTP systems for 300mm wafers. We have factored in a large complement of systems into initial 300mm facilities starting in 1997, largely contributing to the higher-than-the-market growth. RTP systems are primarily used today for saUcide anneals and are primarily driven by logic and ASIC capacity. Dataquest believes that batch tube systems will continue to resist penetration by RTP in areas such as well drive, BPSG reflow, and thermal oxidation because of the demonstrated cost-of-ownership benefits in these areas, at least through 200mm wafers.

Ion Implantation

Overall ion implantation system revenue is forecast to grow at a CAGR of

17.7 percent for the years 1994 through 2000. This market segment wiU continue to be the most volatile, because of the highly device-specific nature of the implant segments. The fastest-growing segment is expected to be high-energy implantation, which is evoking intense interest because of its potential for process simplification and mantifacturing cost reduction. The first year of true production ramp is expected to occur in 1998 as

0.4-micron technologies become mainstream, although early adopters such as Samsung have placed high-voltage implant into 16Mb DRAMs.

The market strength in 1995 has resulted from two impacts. First, Japanese companies are beginning to take interest in applying the technology, and second, the first penetration into logic application is under way.

New implant systems wiU continue to offer improvement in uniformity, particle control, charging, and wafer throughput. The number of implant steps requiring medium-current ion sources is expected to grow faster than high-dose implemt steps, again driven by the higher proportion of the

SEI\/IM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

28 Semiconductor Equipment, Manufacturing, and Materials Worldwide

worldwide semiconductor market that is logic, with the shallow junctions preferentially driving the trend toward medium-current implant system sales.

Segments and Tools of Emerging Importance

Yield enhancement is the talk of the moment. If a piece of equipment enhances yield directly in any way, particularly if it is an incremental concept and on the lower end of the price scale, it will gain immediate acceptance. This appears to be the case in three specific areas: photostabihzers, a tool we refer to currently as a pull dryer, and emerging issues in process control/ overlay.

Photostabilizers are relatively small systems, selling for perhaps $200,000, that employ a UV-curing technique to treat exposed photoresist on the wafer before placing the wafer into the etcher. Fusion Systems in the

United States has been one of the leaders in this area. The system improves the quaUty of the photoresist in preparation for etch and thus dramatically improves the etch system performance. Samsung is known to have installed many units, and the product segment is starting to attract major interest.

A visit to Steag AG in Germany recently introduced us to the concept of what we currently refer to as a pull dryer. These systems are replacing the spin dryers currently employed within the automated wet stations today.

Wafers are placed in the system after a wet cleaning process to dry, but they do not spin. Instead, they are pulled from a bath of IPA/water in a controlled fashion into an IPA vapor atmosphere. As the wafer is pulled out of the liquid sheets off the surface in such a way as to not leave any water spots, which often hold kiUer residue defects. Micron Technology is known to have purchased many of these systems, and w e expect this technology to be introduced by several companies in the wafer cleaning area.

Process control systems are the core of the yield enhancement movement, and it is estimated that approximately 80 percent of the failures in ICs

(killer defects) may be attributed to particulate contamination. Therefore, pre- and postprocess particle and defect monitoring and characterization are key to increasing yield. There are several key developments to watch currently in the equipment segments of process control that we beUeve will experience very strong growth in the years ahead. These segments include wafer inspection and review, CD metrology, and thin-film measurement.

Patterned wafer inspection stations have led the way to in-Hne use in fabs.

Demand for unpattemed wafer inspection is also gaining momentum in the applications such as post-CMP inspection. Defect review technology has evolved from microscope-based systems to automated stations that characterize defects at the coordinates provided by wafer inspection systems. Laser-based systems, such as Ultrapointe's recently introduced product, address the throughput issue along with increased resolution.

With higher throughput and automated capabilities, defect review stations should follow the inspection segment to be used in-line.

The CD-SEM market has grown tremendously in the past several years, with the major thrust being the introduction of high-throughput

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Water Fab Equipment Spending Forecast 29

automated SEM systems introduced several years ago. Operator "interpretation" of SEM data and measurement has always been a problem in CD metrology, and today's SEM systems have made quaUtative improvement to the electron emission source and are equipped with pattern recognition software and hardware, thus automating the interpretation function. In turn, the ASPs have increased twofold in a range of $1.2 milHon. ASPs wiU continue to grow along with demand for these automated high-throughput systems.

Thin-film measurement is key to intra- and interlevel metal interconnect and storage capacitance applications. Thin-film measurement systems are used in-line to monitor the in etch, lithography (photoresist), deposition, and diffusion steps. Although this market was driven by logic apphcations more than memory production, in the past several years DRAM manufacturers have begun integrating thin-film measurement stations into their process lines.

Dataquest Perspective

Wafer fab equipment spending in 1995 wiU grow 52 percent worldwide, driven by massive spending in the Asia/ Pacific region and a PC production boom in Europe. A DRAM-sensitive investment in Japan and continued healthy investment in North America round out this third boom year.

Segment growth in 1994 and 1995 is being led by DRAM- or capital spending-sensitive equipment, with steppers, high-current and high-voltage implant, wafer inspection, and polysilicon etch exhibiting significantly stronger growth than market average. We expect no major segment declines in 1995, as capacity additions are broad-based and worldwide.

We have increased our forecast growth in 1996 to 21.5 percent in the worldwide wafer fab equipment market, based on backlog and bookings momentum from the 1995 surge and several new Asian company projects coming on line in 1996.

The NCI model that measures the net cumulative under- or overinvestment indicates that by the end of 1995, the semiconductor manufacturing world will be overtnvested in wafer fab equipment to the tune of $5.2 bilhon, or 31.7 percent, exceeding peaks exhibited in 1984 and 1989.

After four strong expansion years in 1993 through 1996, equipment purchases in 1997 should decline modestly, followed by a relatively flat growth in 1998. Investment in DRAM capacity will be curtailed as producers elect to convert their 4Mb DRAM capacity to 16Mb, which adds bit capacity through the instant increase in bits per square inch. Also, many

Japanese DRAM facilities now running 150inm weifers will convert to

200mm wafers, further delaying the need for new equipment. DRAMsensitive equipment technologies or capital-intensive segments such as steppers, implantation, diffusion, and polysiUcon etch will be affected more than logic-sensitive technologies such as sputtering, epitaxial reactors, RTP, nontube CVD, or metal etch. The next expansion should kick in by 1999, driven by 0.35- to 0.4-micron capacity expansion.

We have factored in an infrastructure investment in equipment for late

1997 through 1999, which will affect the forecast size of the markets positively. This additional investment will be for initial equipment to fill a few

300mm fabs to run silicon by 1998 and 1999. However, our outlook for a significant 300mm equipment market wiU wait until after 2000.

SEMM-WW-MT-gbOl ©1995 Dataquest Incorporated July 31,1995

30 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Yield enhancement is the trend of the time. Any system technology that can be relatively low-priced and has a direct impact on yield will gain immediate acceptance in volume. Areas particularly important today are cleaning technology, photostabilizers, and process control metrology.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Chapter 4

Silicon Wafer Forecast

Dataquest's current forecast and the underlying assumptions behind our expectations for regional silicon wafer demand reflect significant silicon wafer growth in 1995 and 1996, with upwetrd revisions to the 1997 through

2000 forecast, in Une with an increased semiconductor consumption forecasts worldwide. Our latest forecast, along with highlights of some of the key factors affecting the regional markets, are presented here.

Silicon Forecast Tables

Tables in this chapter include Dataquest's most recent forecasts of regional unit silicon wafer consumption. Five tables (4-1 through 4-5) detail unit consumption by region. Individual forecasts of major product segments such as prime, epitaxial, and test and monitor wafers are included. In addition, five tables (4-6 through 4-10) of regional forecasts for wafer size distribution are presented.

The 200mm Wafer Ramps Up

Dataquest has been studying the subject of 200nun wafers and their ramp rate closely over the last year, particularly in light of the massive announcements in fab capacity over the next few years. There are currently 58 new 200nnm f abs announced, which will come on line in the three years from 1995 through 1997, and there probably will be several more to come. Half of those aimotmced will come on line in 1996 alone.

Wafer suppliers have answered the need with several new wafer plants and bilhons of dollars of committed investment. We continue to beheve that the ramp of 200mm wafers will be somewhat supply-constrained through 1997, that is to say that the industry cannot convert from 150mm at wiU. In this forecast update, w e have revised the forecast —again upward —to reflect expected fab activity in each region and supply increases in place.

While we are stopping short of calling a shortage of 200mm wafers, buyers of wafers will experience firm to rising prices and will be placed on allocation from time to time over the next several years. In an upcoming report,

Dataquest will take a closer look at potential shortage issues in the wafer market. In fact, the 200mm wafer may be in the best supply situation compared to other areas in a few years.

The demand for test and monitor wafers will explode with the ramping of the 200mm wafer. Test wafers account for about 15 to 20 percent of wafers at 150mm, but this number is expected to exceed 50 percent for 200mm wafers during the early stages of ramping (estimates for 1994 are about

60 percent). We do expect the percent usage to subside to the one-third level as 200mm processing becomes mature, yet the near-term crunch exacerbates the supply issues.

What about 300mm Wafers?

Now that the wafer size has been settled, and the time horizon for the first

300m plant(s) has been proposed, w e are initiating a forecast for consumption of 300mm wafers. The level is stiU commercially zero before the year

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated 31

(73

CO cn o

CD

CO o o

-a

o

01

I — ^

CD

O .

@

to

CO tn

O

»>

Table 4-1

Forecast of Captive and Metchant Silicon* and Merchant Epitaxial Wafers by Region

(Millions of Square Inches)

Worldwide Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

1993

2,449.7

2,157.1

292.6

1994

2,919.0

2,535.3

383.7

1995

3,409.0

2,951.4

457.6

1996

3,987.2

3,468.1

519.1

1997

4,478.1

3,917.2

560.9

1998

4,742.9

4,126.7

616.2

1999

5,132.5

4,453.9

678.6

North Amorica Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

Japan Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

Europe Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

Asia/Pacific-ROW Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

•Includes prime, test, and monitor wafers

Source: Dataquest (July 1995)

720.0

565.5

154.5

1,127.3

1,038.3

89.0

291.2

255.1

36.1

311.2

298.2

13.0

832.1

641.9

190.2

1,279.7

1,160.2

119.5

351.2

296.7

54.5

456.0

436.5

19.5

952.3

725.9

226.4

1,457.2

1,317.7

139.5

414.0

347.2

66.8

585.5

560.6

24.9

1,050.1

794.1

256.0

1,686.2

1,530.5

155.7

488.7

411.1

77.6

762.2

732.4

29.8

1,126.8

849.7

277.1

1,892.4

1,726.3

166.1

561.4

476.7

84.7

897.5

864.5

33.0

1,204.0

896.2

307.8

1,947.7

1,768.5

179.2

609.5

516.4

93.1

981.7

945.6

36.1

1,334.6

992.3

342.3

2,054.4

1,860.2

194.2

646.9

545.6

101.3

1,096.6

1,055.8

40.8

0 0

CO

CO cn

Silicon Water Forecast 33

Table 4-2

Forecast Growth Rates of Captive and Merchant Silicon* and Merchant Epitaxial

Wafers by Region (Percent of MSI)

Worldwide Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

1993

16.8

16.7

17.7

1994

19.2

17.5

31.1

1995

16.8

16.4

19.3

1996

17.0

17.5

13.4

1997

12.3

12.9

8.1

1998

5.9

5.3

9.9

1999

8.2

7.9

10.1

2000

10.7

10.5

12.1

North America Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

Japan Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

Europe Total Silicon + Epi

Merchant and Captive Silicon*

Epitaxial Silicon

Asia/Pacific-ROW Total Sihcon + Epi

Merchant & Captive Silicon*

Epitaxial Silicon

'Includes prime, test, and monitor wafers

Source: Dataquest (July 1995)

2.3

23.9

20.8

51.0

30.8

30.5

36.8

10.6

8.2

20.6

15.9

17.2

15.6

13.5

23.1

13.5

11.7

34.3

20.6

16.3

51.0

46.5

46.4

50.0

14.4

13.1

19.0

13.9

13.6

16.7

17.9

17.0

22.6

28.4

28.4

27.7

18.4

16.2

30.2

30.6

19.7

10.3

9.4

13.1

15.7

16.1

11.6

18.0

6.7

14.9

16.0

9.1

17.8

18.0

10.7

7.3

7.0

8.2

12.2

12.8

6.9

5.5

11.1

2.9

2.4

7.9

8.6

8.3

9.9

9.4

9.4

9.4

10.8

10.7

11.2

5.5

5.2

8.4

6.1

5.7

8.8

11.7

11.7

13.0

2000; however, the recent goal of a fab to come on line by 1998 or 1999 means that 300mm wafers will be made and that significant activity will be occurring in R&D. Although no company has stated a firm commitment, we would expect something to be announced among Samsung,

Motorola, Intel, or perhaps one Japanese company (NEC?) by the end of this year toward a 300mm fab. We are assuming that there will be at least two plants, perhaps three, based on who decides to cooperate with whom

(we are not forecasting the partnerships).

10.7

10.2

12.3

8.0

7.8

10.2

8.6

7.9

12.0

16.9

16.8

18.9

Further details and issues regarding the move toward 300mm wafers are included in a recent Market Analysis newsletter (SEMM-WW-MA-9503).

Please refer to that newsletter for a more detailed discussion.

We beHeve the first commercially productive plant will be started in the

2000-to-2002 time frame (after the feasibility noted before), with serious volume ramp-up in the years 2002 to 2004. This would be consistent with

0.25-micron technology being primarily produced on 300mm wafers. This means that 200mm waijfers represent a two-technology-generation wafer size, and f abs being built today may have longer Hves than history would indicate.

Highlights of the North American Silicon Wafer IViarlcet and Forecast

Sihcon consumption in North America is forecast to grow 14.4 percent in

1995 to 952 million square inches (MSI), followed by a mild 10.3 percent growth in 1996 to 1,050 MSI. Microprocessor and other logic chip demand have been and will continue to be key drivers behind increased silicon

SEiVilVl-WW-MT-gSOl ©1995 Dataquest Incorporated July 31,1995

CO

^

^

CO

O i

o

o

•3

o

CD

O .

@

-^

CO

CO tn

O

&>

r-t-

M

X3

Table 4-3

Forecast of Captive and Merchant Silicon* Wafers by Region

(Millions of Square Inches)

Worldwide Total Silicon*

Merchant Silicon

Captive Silicon

1993

2,157.1

2,032.1

125.0

North America Total Silicon*

Merchant Silicon

Captive Silicon

Japan Total Silicon*

Merchant Silicon

Captive Silicon

Europe Total Silicon*

Merchant Silicon

Captive Silicon

Asia/Pacific-ROW Total Silicon*

Merchant Silicon

Captive Silicon

'Includes prime, test, and monitor wafers

Source: Dataquest (July 1995)

565.5

487.5

78.0

1,038.3

1,003.3

35.0

255.1

250.1

5.0

298.2

291.2

7.0

1994

2,535.3

2,398.3

137.0

1995

2,951.4

2,814.4

137.0

1996

3,468.1

3,331.1

137.0

1997

3,917.2

3,780.2

137.0

1998

4,126.7

3,989.7

137.0

199

4,453.

4,316.

137.

641.9

551.9

90.0

1,160.2

1,125.2

35.0

296.7

291.7

5.0

436.5

429.5

7.0

725.9

635.9

90.0

1,317.7

1,282.7

35.0

347.2

342.2

5.0

560.6

553.6

7.0

794.1

704.1

90.0

1,530.5

1,495.5

35.0

411.1

406.1

5.0

732.4

725.4

7.0

849.7

759.7

90.0

1,726.3

1,691.3

35.0

476.7

471.7

5.0

864.5

857.5

7.0

896.2

806.2

90.0

1,768.5

1,733.5

35.0

516.4

511.4

5.0

945.6

938.6

7.0

992.

902.

90.

1,860.

1825.

35.

545.

540.

5.

1,055.

1,048.

7.

CO

CO

CO tn

Silicon Wafer Forecast 35

Table 4r4

Forecast Growth Rates of Captive and Merchant Silicon* by Region (Percent of MSI)

Worldwide Total Silicon*

Merchant Silicon

Captive Silicon

1993

16.7

17.6

4.2

1994

17.5

18.0

9.6

1995

16.4

17.3

0

1996

17.5

18.4

0

1997

12.9

13.5

0

1998

5.3

5.5

0

1999

7.9

8.2

0

2000

10.5

10.8

0

North America Total Silicon*

Merchant Silicon

Captive Silicon

Japan Total Silicon*

Merchant Silicon

Captive Silicon

Europe Total Silicon*

Merchant Silicon

Captive Silicon

Asia/Pacific-ROW Total Silicon*

Merchant Silicon

Captive Silicon

'Includes prime, test, and nnonitor wafers

Source: Dataquest (July 1995)

21.3

0

30.5

30.9

16.7

8.2

8.1

8.3

17.2

18.2

-5.4

20.8

16.6

0

46.4

47.5

0

13.5

13.2

15.4

11.7

12.1

0

16.3

13.1

15.2

0

13.6

14.0

0

17.0

17.3

0

28.4

28.9

0

9.4

10.7

0

16.1

16.6

0

18.4

18.7

0

30.6

31.0

0

16.2

0

18.0

18.2

0

7.0

7.9

0

12.8

13.1

0

16.0

0

8.3

8.4

0

9.4

9.5

0

5.5

6.1

0

2.4

2.5

10.7

11.9

0

5.2

5.3

0

5.7

5.7

0

11.7

11.7

0

demand in North America, and epitaxial wafer demand is expected to grow faster than the overall market throughout this decade.

10.2

11.2

0

7.8

7.9

0

7.9

8.0

0

16.8

16.9

0

Merchant epi wafer consumption will increase 19 percent to 226 MSI, driven in large part by microprocessor manufacturers, such as Intel and

AMD, which build their microprocessors on epi wafers. Strength in the automotive and discrete segments of the chip market have also been a bit stronger than expected. By 1998, epitaxial silicon will account for 25 percent of the square inches consumed in North America—the highest concentration in any region.

Dataquest's longer-term forecast for North American silicon consumption has remained steady despite the overall semiconductor market upward revision since the last update. The primary reason for the conservative forecast comes from the belief that most of the incremental growth in production wiU be targeted in the European and Asia/Pacific regions

(see Chapter 6). We are projecting that total silicon MSI will grow at a

10 percent CAGR for the period of 1994 through 2000.

Highlights of the Japanese Silicon Wafer Market and Forecast

Our forecast for Japan's silicon consumption remains basically unchanged from our last update, with the silicon market growing 13.9 percent to 1,457

MSI in 1995 with continued strong growth in 1996 and softening in 1997 and 1998. The persistent high price of 4Mb DRAMs through 1995, with low yields not edlowing the more siUcon-efficient 16Mb density to be produced economically, is the reason for the continued near-term optimism.

As the conversion to the 16Mb density occurs, silicon square inch demand will ease.

SEIVlM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

CO

C O

C71

o

©

CO

CO

(SI

O

&3.

su

.a

c

CD

</>

O

O

— T

T3

—^

0>

CD

Table 4-5

Forecast of Merchant Prime a n d Tes^Monitor Wafers b y Region

(Millions of Square Inches)

Worldwide Merchant Silicon

Growth Rate (%)

Prime

Test and Monitor

North America Merchant Silicon

Growth Rate (%)

Prime

Test and Monitor

Japan Merchant Silicon

Growth Rate {%)

Prime

Test and Monitor

Europe Merchant Silicon

Growth Rate (%)

Prime

Test and Monitor

Asia/Pacific-ROW M ^ ^ i a i i l Silicon

Growth Rale {%)

Prime

Test and Monitor

Source; Dataquest (July 1995)

1993

2,032.1

17.6

1,616.8

415.3

1994

2,398.3

18.0

1,892.6

505.7

1995

2,814.4

17.3

2,187.1

627.3

1996

3,331.1

18.4

2,560.5

770.6

1997

3,780.2

13.5

2,903.1

877.1

1998

3,989.7

5.5

3,062.9

926.8

1

4,31

3,32

98

487.5

8.1

385.1

102.4

1,003.3

18.2

800.1

203.2

250.1

21.3

200.1

50.0

291.2

30.9

231.5

59.7

291.7

16.6

230.4

61.3

429.5

47.5

337.2

92.3

551.9

13.2

430.5

121.4

1,125.2

12.1

894.5

230.7

635.9

15.2

466.8

169.1

1,282.7

14.0

1,024.5

258.2

342.2

17.3

270.6

71.6

553.6

28.9

425.2

128.4

704.1

10.7

512.5

191.6

1,495.5

16.6

1,184.9

310.6

406.1

18.7

316.4

89.7

725.4

31.0

546.7

178.7

759.7

7.9

558.1

201.6

1,691.3

13.1

1,338.7

352.6

471.7

16.2

366.1

105.6

857.5

18.2

640.2

217.3

511.4

8.4

3%.0

115.4

938.6

9.5

700.7

237.9

806.2

6.1

589.8

216.4

1,733.5

2.5

1,376.4

357.1

90

1

66

24

1,82

1,45

36

54

42

12

1,04

1

78

25

CO

CO

CO on

Silicon Wafer Forecast 37

Table 4-6

Worldwide Wafer Size Distribution Forecast, 1993-2000

(Percent Square Inches by Diameter and Unit EHstribution by Wafer Starts)

Diameter

Percent Square Inches by Diameter

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total

Total MSI

Growth (%)

Area

(sq .in.)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

1993

3.6

0

100.0

2,450

16.8

0.1

1.6

16.1

28.8

49.6

1994 1995

0.1

1.4

14.8

24.7

48.0

11.0

0

100.0

2,919

19.2

16.4

0

100.0

3,409

16.8

0.1

1.2

13.2

22.6

46.5

1996 1997 1998

0.1

1.0

11.6

19.8

46.2

21.4

0

100.0

3,987

16.9

0

0.8

10.3

17.8

45.0

26.1

0

100.0

4,478

12.3

0

0.7

9.5

16.6

43.8

29.3

0

100.0

4,743

5.9

1999

0

0.6

9.0

15.4

42.5

32.3

0.1

100.0

5,133

8.2

2000

35.5

0.1

100.0

5,681

10.7

0

0.5

8.4

14.3

41.2

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total Wafers

Average Wafer Diameter

(Inches)

Source: Dataquest (July 1995)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

1.0

5.6

32.5

37.1

44.4

1.8

0

122.5

5.05

1.1

5n

35.5

37.9

51.2

6.6

0

138.0

5.19

1.2

5.9

36.8

40.5

58.0

11.5

0

153.9

5.31

0.7

5.7

37.9

41.4

67.3

17.5

0

170.6

5.45

0.6

5.3

37.9

41.9

73.5

24.0

0

183.2

5.58

0.6

4.7

37.1

41.4

75.9

28.6

0.02

188.3

5.66

0.3

4.6

38.0

41.5

79.7

34.0

0.06

198.3

5.74

0

4.0

39.1

42.6

85.5

41.5

0.07

212.8

5.83

Unlike North America, with its sizable CMOS epi wafer market, Japan's merchant epitaxial wafer market is more focused on discrete and bipolar applications. Therefore, a recovery in the economy will have more of an effect on the growth of Japan's epi wafer market. Epitaxial demand is expected to grow nearly 17 percent in 1995, after a 34 percent increase in

1994. Another good growth year is expected in 1996.

Dataquest rennains moderately conservative with its longer-term growth scenario for silicon wafer demand in Japan, as the country continues to work through economic and semiconductor production infrastructure

(too heavily dependent on the DRAM) difficulties. Recent investment patterns, however, indicate that the Japanese semiconductor manufacturer is willing to come to the table and invest, preserving their stake in the memory business against the Koreans. The desired shift of the Japanese product mix to higher value-added semiconductors is apparently on the back burner until the current memory cycle subsides, but will again come to the

•SEMI\/I-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

38 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 4-7

North American Wafer Size Distribution Forecast, 1993-2000

(Percent Square Inches by Diameter and Unit Distribution by Wafer Starts)

Diameter

Percent Square Inches by Diameter

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total

Total MSI

Growth (%)

Area

(sq. in.)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

1993

0.1

1.2

22.4

26.5

44.0

5.8

0

100.0

720

10.6

1994 1995 1996 1997

0.1

1.1

20.1

21.7

42.2

14.8

0

100.0

832

15.6

0.1

1.0

17.2

19.1

39.6

23.0

0

100.0

952

14.4

0

0.9

15.3

17.2

39.6

27.0

0

100.0

1,050

10.3

0

0.7

14.1

16.0

39.4

29.8

0

100.0

1,127

7.3

1998

1999

0

0.6

13.2

14.9

38.2

33.0

0.1

100.0

1,204

6.9

0

0.5

12.7

13.7

35.9

37.0

0.2

100.0

1,335

10.8

2000

0

0.3

11.9

12.5

35.6

39.5

0.2

100.0

1,478

10.7

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total Wafers

Average Wafer Diameter

(Inches)

Source: Dataquest (July 1995)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

0.2

1.2

13.3

10

11.6

0.9

0

37.2

4.97

0.3

1.3

13.7

9.5

12.8

2.5

0

40.2

5.14

0.3

1.3

13.5

9.6

13.8

4.5

0

42.9

5.31

0

1.3

13.2

9.5

15.2

5.8

0

45.0

5.45

0

1.1

13.1

9.5

16.2

6.9

0

46.8

5.54

0

1.0

13.1

9.4

16.8

8.2

0.1

48.5

5.62

0

0.9

13.9

9.6

17.5

10.1

0.2

52.2

5.71

0

0.6

14.5

9.7

19.2

12.0

0.3

56.0

5.79 forefront soon. We are estimating that silicon demand wiU grow at a

9.6 percent CAGR for the years 1994 through 2000, the slowest growth of all regions.

Highlights of the European Silicon Wafer IVIarket and Forecast

Demand for silicon wafers in Europe, as well as wafer fabrication equipment, remains heavily dependent on the fab activities of foreign semiconductor companies. With increased presence from European companies, the outlook for siHcon consumption has brightened.

European silicon demand is forecast to be 414 MSI in 1995, up just under

18 percent from 1994 levels. Siemens' Dresden and other DRAM production and U.S. multinationals Intel, Motorola, and Texas Instruments will continue to ramp to answer the PC production boom in Europe, helping silicon consumption grow another 18 percent in 1996. These driving forces have caused us to call Europe the second fastest-growing region for silicon consumption, growing at a CAGR of 12.2 percent through the year 2000.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Silicon Wafer Forecast 39

Table 4-8

Japanese Wafer Size Distribution Forecast, 1993-2000

(Percent Square Inches by Diameter and Unit Distribution by Wafer Starts)

Diameter

Percent Square Inches by Diameter

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total

Total MSI

Growth (%)

Area

(sq. in.)

1993

3.14

7.07

12.17

19.02

27.38

48.67

109.56

0

1.5

11.5

32.2

53.3

1.5

0

100.0

1,127

15.9

1994 1995

1996

0

1.4

10.8

29.9

51.6

6.3

0

100.0

1,280

13.5

0

1.3

10

28.3

51.6

8.8

0

100.0

1,457

13.9

0

1.0

9.3

25.2

51.5

13.0

0

100.0

1,686

15.7

1997

1998

1999

0

0.9

8.2

22.7

51.5

16.7

0

100.0

1,892

12.2

0

0.8

7.6

21.7

51.2

18.7

0

100.0

1,948

2.9

0

0.8

7.1

20.5

50.7

20.8

0.1

100.0

2,054

5.5

2000

0

0.7

6.8

19.5

48.9

24.0

0.1

100.0

2,219

8.0

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total Wafers

Average Wafer Diameter

(Inches)

Source: Dataquest (July 1995)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

0

2.4

10.7

19.1

21.9

0.3

0

54.4

5.14

0

2.5

11.4

20.1

24.1

1.7

0

59.8

5.22

0

2.7

12.0

21.7

27.5

2.6

0

66.4

5.28

0

2.4

12.9

22.3

31.7

4.5

0

73.8

5.39

0

2.4

12.8

22.6

35.6

6.5

0

79.8

5.49

0

2.2

12.2

22.2

36.4

7.5

0

80.5

5.55

0

2.3

12.0

22.1

38.0

8.8

0.2

83.3

5.60

0

2.2

12.4

22.8

39.6

10.9

0.2

87.9

5.67

Epitaxial wafer demand in the region will increase from 36 MSI in 1993 to

67 MSI in 1995, a nearly doubling in consumption in two years, a direct result of Intel in Ireland. Beyond 1995, epitaxial wafer demand will come from European producers, leading to a CAGR just 1 percentage point above polished bare wafer consumption through the decade.

Highiights of the Asia/Pacif ic-ROW Silicon Wafer IVIarlcet and Forecast

Sihcon consumption is growing at a 28.4 percent pace in 1995, the largest growth rate of any other region in the world. Production is expanding at a fierce pace, and not expected to ease through 1997, as many large 200mm fab projects are in various stages of construction and start-up. As the

Asia/Pacific region coUects the largest proportion of 1995 capital moneys, the trend for high silicon consumption growth will continue unabated.

The phenomenal growth in sihcon consumption Asia/Pacific in 1995 is tied directly to the manufacturing activities of the Korean DRAM producers. Future growth will be seen here, but increased production in Taiwan,

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

40 Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide

Table 4-9

European Wafer S i z e Distribution Forecast, 1993-2000

(Percent Square Inches b y Diameter a n d U n i t Distribution b y Wafer Starts)

Diameter

Percent Square Inches by Diameter

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total

Total MSI

Growth (%)

Area

(sq. in.)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

1993

0.2

1.7

22.0

30.1

42.3

3.7

0

100.0

291

23.9

1994

0.1

0.7

20.5

21.7

43.0

14.0

0

100.0

351

20.6

1995

1996 1997 1998 1999

0.1

0.5

19.7

20.8

44.1

14.8

0

100.0

414

17.9

0

0.4

17.0

18.0

45.1

19.5

0

100.0

489

18.0

0

0.2

15.0

15.7

45.2

23.9

0

100.0

561

14.9

0

0.1

13.5

14.2

44.7

27.5

0

100.0

610

8.6

0

0.1

12.5

13.2

44.2

30

0

100.0

647

6.1

2000

0

0.1

11.8

12.5

42.9

32.7

0

100.0

702

8.6

Unit Distribution by Wafer Starts

(Milhons of Wafers)

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total Wafers

Average Wafer Diameter

(Inches)

Source: Dataquest (July 1995)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

0.2

0.7

5.3

4.6

4.5

0.2

0

15.5

4.89

0.1

0.3

5.9

4.0

5.5

1.0

0

16.9

5.14

1.3

0

19.6

5.19

0.1

0.3

6.7

4.5

6.7

2.0

0

21.7

5.35

0

0.3

6.8

4.6

8.0

0

0.2

6.9

4.6

9.3

2.8

0

23.7

5.49

0

0.1

6.8

4.6

10

3.4

0

24.8

5.59

0

0.1

6.6

4.5

10.4

4.0

0

25.7

5.67

Singapore, and most recently armounced Thailand foundries will also contribute to the growth. Foundry-related capital spending in these Asia/

Pacific countries has exploded, increasing from about $900 million in 1994 to over $1.6 bilUon in 1995. As these fabs come on hne in 1995 and beyond, they will provide some regional consumption stability as memory-related silicon consumption cools in 1997 and 1998. Taiwan, with its many new

DRAM producers coming on hne in 1996, will cause silicon consumption to grow by more than 30 percent again in 1996. Asia/Pacific-ROW remains the fastest-growing silicon consumer, with a five-year CAGR forecast of 16.9 percent.

0

0.1

6.8

4.6

11.0

4.7

0

27.3

5.73

Dataquest's First Silicon Revenue Forecast

Dataquest has been tracking silicon wafer revenue and market share since

1985 but has always provided forecast information in terms of square inch area and unit wafer size distributions. With the announcement of the initial public offering of MEMC Electronic Materials (during July 1995),

SEiVIM-WW-IVIT-gSOl ©1995 Dataquest Incorporated July 31,1995

Silicon Wafer Forecast 41

Table 4-10

Asi^Pacific-ROW Wafer Size Distribution Forecast, 1993-2000

(Percent Square Inches by Diameter and Unit EHstribution by Wafer Starts)

Diameter

Percent Square Inches by Diameter

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total

Total MSI

Growth (%)

Area

(sq. in.) 1993 1994 1995 1996 1997

3.14

7.07

12.17

19.02

27.38

48.67

109.56

0.6

2.9

13.0

20.8

56.3

6.4

0

100.0

311

30.8

0.5

2.4

12.0

17.8

52.3

15.0

0

100.0

456

46.5

0.4

1.9

9.8

15.4

47.0

25.5

0

100.0

586

28.4

0.3

1.6

7.9

12.4

44.3

33.5

0

100.0

762

30.2

0.2

1.3

7.0

11.0

38.0

42.5

0

100.0

898

17.8

1998 1999

0.2

1.0

6.4

10

35.4

46.9

0.1

100.0

982

9.4

0.1

0.8

6.0

9.2

34.3

49.4

0.2

100.0

1,097

11.7

2000

33.4

52.4

0.2

100.0

1,281

16.9

0

0.6

5.2

8.2

Unit Distribution by Wafer Starts

(Millions of Wafers)

2 Inches

3 Inches

100mm

125mm

150mm

200mm

300mm

Total Wafers

Average Wafer Diameter

(Inches)

Source: Dataquest (July 1995)

3.14

7.07

12.17

19.02

27.38

48.67

109.56

0.6

1.3

3.3

3.4

6.4

0.4

0

15.4

5.07

0.7

1.5

4.5

4.3

8.7

1.4

0

21.2

5.24

0.7

1.6

4.7

4.7

10.1

3.1

0

24.9

5.47

0.7

1.7

4.9

5.0

12.3

5.2

0

29.9

5.69

0.6

1.7

5.2

5.2

12.5

7.8

0

32.9

5.90

12.7

9.5

0.1

34.5

6.02

0.6

1.4

5.2

5.2

0.3

1.2

5.4

5.3

13.7

11.1

0.2

37.2

6.13

15.6

13.8

0.2

41.5

6.27

0

1.1

5.5

5.5 we believe it is now important to consider the revenue-generating capability of the industry.

We believe a revenue forecast wUl benefit our clients, gaining increased visibility to the capital markets for the industry. Table 4-11 contains the revenue forecast for sihcon wafers worldwide. In our analysis, we have concluded that the revenue forecast would resemble the semiconductor industry more closely than the capital spending markets. The concept of semiconductor revenue per square inch is more closely tied to silicon consumption than raw wafer capacity of the industry. The six-year CAGR of

15.3 percent is about 1 percentage point below the semiconductor forecast of 16.2 percent, consistent with the model that semiconductor manufacturers will try to control the costs associated with manufacturing, which includes using silicon more efficiently in the future.

SEMI\/I-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

42 Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 4-11

W o r l d w i d e Merchant S i l i c o n Wafer R e v e n u e Forecast, 1992-2000

(MilUons of U.S. Dollars)

Worldwide

Percent Growth

1992

2,991

2.7

1993

3,554

18.8

1994

4,592

29.2

1995

5,591

21.8

1996

6,778

21.2

Note: This table's data includes polished, virgin test, and epitaxial silicon.

Source: Dataquest (July 1995)

1997

7,837

15.6

1998

8,537

8.9

1999

9,444

10.6

2000

10,793

14.3

CAGR (%)

1994-2000

15.3

Dataquest Perspective

The siHcon market, driven by a stronger long-term picture for semiconductors in general, wiU grow faster over the next six years than the recent past. As the industry transforms into a 200mm baseline, the outlook for silicon wafer manufacturers becomes brighter. SiUcon manufacturers have answered the call for 200mm capacity, and the semiconductor market has again responded with the cry "More, more!" We beHeve siHcon manufacturers ramp plans in 200mm have been strategically and smartly measured, because the overcapacity situations of 1985 and after are still remembered.

While there will be activity with 300mm wafers, this is expected to be

R&D-focused and low volumes until after the turn of the decade.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated July 31,1995

Chapter 5

Semiconductor Consumption Forecast ^ ^ ^ ^ ^ . ^ ^ ^ ^

This chapter presents data on the worldwide semiconductor market by region. The regional semiconductor market, or regional semiconductor consumption, deals with where chips are consumed; this contrasts with regional semiconductor production, which deals with where chips are manufactured. The data presented here is for the merchant market and does not include the value of chips made by captive semiconductor manufacturers for internal use.

This is an excerpt from the Semiconductor Five-Year Forecast, published by

Dataquest in April (SCND-WW-MT-9501). Further details regarding this forecast can be found in that publication.

Yearly exchange rate variations can have a significant effect on the 1988 through 1994 data in the following tables. For more information about the exchange rates used and their effects, refer to Appendix B.

Semiconductor Consumption

Table 5-1 shows revenue and growth from semiconductor shipments for the years 1988 through 1998, broken down by region. Table 5-2 shows revenue and growth from semiconductor shipments for the yeeirs 1994 through 2000, broken down by region.

Table 5-1

W o r l d w i d e Semiconductor C o n s u m p t i o n * b y R e g i o n ( M i l l i o n s of U.S. Dollars)

North America

Percent Growth

Japan

1988

15,844

23.2

20,772

1989

17,070

7.7

21,491

1990

16,540

-3.1

20,257

Percent Growth

Europe

Percent Growth

Asia/Pacific-ROW

Percent Growth

Worldwide

39.2

8,491

30.7

5,752

45.0

50,859

3.5

9,498

11.9

6,280

9.2

54,339

Percent Growth 33.0 6.8

* Includes merchant semiconductor companies only

NA = Not applicable

Source: Dataquest (July 1995)

-5.7

10,415

9.7

7,333

16.8

54,545

0.4

1991

16,990

2.7

22,496

11.1

11,014

5.8

9,194

25.4

59,694

9.4

1992

20,430

20.2

20,579

-8.5

12,218

10.9

12,034

30.9

65,261

9.3

1993

27,926

36.7

24,645

19.8

15,461

26.5

17,486

45.3

85,518

31.0

1994

35,939

28.7

31,010

25.8

20,819

34.7

22,812

30.5

110,580

29.3

CAGR (%)

1989-1994

16.1

-

7.6

-

17.0

-

29.4

-

15.3

-

SEMM-WW-MT-9501

©1995 Dataquest Incorporated

43

44

Semiconductor Equipment, Manufacturing, and IVIaterials Worldwide

Table 5-2

Worldwide Semiconductor Consumption* by Region (Millions of U.S. Dollars)

1994

North America

Percent Growth

Japan

35,939

28.7

Percent Growth

31,010

25.8

Europe

Percent Growth

Asia/Pacific-ROW

20,819

34.7

Percent Growth

Worldwide

Percent Growth

22,812

30.5

110,580

29.3

* Merchant semiconductor sales only

Source: Dataquest (July 1995)

1995

45,307

26.1

35,897

15.8

26,088

25.3

28,525

25.0

135,817

22.8

1996

53,570

18.2

39,599

10.3

28,932

10.9

33,879

18.8

155,980

14.8

1997

61,770

15.3

41,945

5.9

32,860

13.6

39,481

16.5

176,056

12.9

1998

72,163

16.8

45,087

7.5

37,671

14.6

47,923

21.4

202,844

15.2

1999

83,191

15.3

49,905

10.7

42,607

13.1

58,543

22.2

234,246

15.5

2000

98,414

18.3

54,896

10.0

49,168

15.4

70,661

20.7

273,139

16.6

CAGR (%)

1994-2000

18.3

-

10.0

-

15.4:

-

20.7

-

16.3

-

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated

July 31,1995

Chapter 6

Semiconductor Production Forecast

This chapter presents data on the worldwide semiconductor production by region. Semiconductor production is defined by the place where the wafers are fabricated. Regional semiconductor production includes all production in the region, including merchant and captive producers and all foreign producers. For instance. North American semiconductor production includes Digital Equipment and Delco fabs as well as Japanese company and European company fabs in North America.

Yearly exchange rate variations can have a significant effect on the 1988 through 1994 data in the following tables. For more information about the exchange rates used and their effects, refer to Appendix B.

The semiconductor industry has a global manvifacturing business. Production of semiconductors is constantly shifting among regions, as new capital money is flowing toward areas of relative lower capital cost, and higher-growth areas of consumption. Dataquest reviews some of the trends and potential impacts for the future.

Historical Semiconductor Production

Table 6-1 shows the historical semiconductor production for the years

1988 through 1994 broken down by region. Dataquest follows a niethodology that employs the use of our fab database, estimating the memory, microcomponent, and logic production components separately, and estimating net production among regions for foundry activity. This approach provides insight in observing and forecasting production trends.

Because of the reclassification of the MOS portion of IBM Microelectronics' business as merchant, the captive production figures changed dramatically in 1993. However, IBM's bipolar production, which is consumed internally, is stiU classified as captive.

Captive Semiconductor Production

Semiconductor production from captive manufacturers is estimated to be

$1.98 billion in 1994, down from just over $2 billion in 1993. IBM restructured and entered the merchant semiconductor market as of 1993.

Dataquest has reclassified IBM's MOS semiconductor production to merchant, but the bipolar products (exclusively used internally) are still reported as captive. This part of IBM's business, which will be converted to MOS over the next four to five years, resulted in a lower production figure for captive production in 1994 and for future years.

Many captive producers may consider the move to merchant to take better advantage of the worldwide growth of semiconductors, leveraging their sunk costs in plant and equipment for higher return in a larger end-user base. Still others may elect to take advantage of the now evolving and maturing fovmdry business, electing to contract out their nianufacturing rather than invest in expensive new facilities for their relatively small production base. We have not, however, included any such movement to merchant or fabless in our captive production forecast.

SEMM-WW-MT-9501 ©1995 Dataquest Incorporated 45

46

Semiconductor Equipment, Manufacturing, and Materials Worldwide

Table 6-1

Worldwide Semiconductor Production* by Region, 1988-1994 (Millions of U.S. Dollars)

1988 1991

Total North America

Percent Growth

Percent Worldwide

Merchant

Captive

Total Japan

Percent Growth

Percent Worldwide

Merchant

Captive

Total Europe

Percent Growth

Percent Worldwide

Merchant

Captive

Total Asia/Pacific-ROW

Percent Growth

Percent Worldwide

Merchant

Captive

Worldwide

Percent Growth

Merchant

Percent Growth

Captive

Percent Growth

26,388

344

5,854

23.9

10.6

5,277

577

1,868

20,533

20.8

37.3

17,326

3,207

26,732

40.5

48.6

71.8

3.4

1,868

NA

54,987

31.4

50,859

33.0

4,128

15.2

* Includes only merchant and captive semiconductor company sales

NA = Not applicable

Source: Dataquest (July 1995)

26,039

7.6

40.4

22,275

3,764

28,338

7.4

44.0

27,925

413

6,979

9.9

59,184

7.6

54,339

6.8

4,845

17.4

669

1,974

5.7

3.3

1,974

NA

1989

22,232

8.3

37.6

18,464

3,768

28,527

6.7

48.2

28,119

408

6,451

10.2

10.9

5,782

-1.6

10.7

5,723

627

2,392

21.2

4.0

2,392

NA

59,328

0.2

54,545

0.4

4,783

-1.3

1990

24,202

8.9

40.8

20,453

3,749

26,384

-7.5

44.5

25,977

407

6,350

10.8

6,396

583

3,097

29.5

4.8

3,097

NA

64,453

8.6

59,693

9.4

4,760

-0.5

12.2

7,957

632

4,391

41.8

6.2

4,391

NA

70,460

9.3

65,260

9.3

5,200

9.2

1992

29,457

13.1

41.8

25,248

4,209

28,023

-1.1

39.8

27,664

359

8,589

23.1

13.4

11,452

320

7,577

72.6

8.7

7,577

NA

87,539

24.2

85,518

31.0

2,021

-61.1

1993

33,446

13.5

38.2

31,745

1,701

34,744

24.0

39.7

34,744

0

11,772

37.1

1994

40,268

20.4

35.8