AI summary by TwinMind
How Japan Won Lithography (& Why America Lost)
Asianometry
21m 42s
American Lithography Dominance (1950s-1970s)
- 1978: 70% world lithography equipment from American suppliers
- 1982: Americans held 62% market share
- Origins in late 1950s: Fairchild, Texas Instruments created first silicon integrated circuits
- US government/military demand drove market (Apollo missions, Minuteman II ICBM)
- Equipment evolution: in-house production → specialized external vendors
- 1973: Perkin Elmer introduced Microline projection aligner
- Improved yield rates from 10% to 70%
- Emerged as industry leader
- 1975: GCA (Geophysics Corporation of America) acquired David Mann
- 1977: GCA shipped first commercial stepper to IBM
- Superior defect/yield rates vs projection aligners
- Weaknesses: slow throughput, $500k cost (30x more than Perkin Elmer)
- Major customers: IBM, AT&T, Fairchild, National Semiconductor
- GCA revenues: $62M in 1976
Japanese Technology Transfer Policy (1960s)
- Government policy: foreign companies must establish joint ventures with local players
- Local absorption of foreign knowledge with goal of eventual replacement
- Japanese stayed current with American practices but lacked indigenous innovation
- Rare Japanese successes: semiconductor assembly equipment (Shinkawa, Disco)
- Political restrictions forced automation vs offshoring to Southeast Asia
- Reality: Americans dominated computers, chips, lithography equipment
- Texas Instruments R&D spending > Fujitsu + Hitachi + NEC combined
Birth of VLSI Project (1976)
- Political catalyst: IBM one-chip computer news seen as "Commodore Perry black ships" event
- LDP politician Hashimoto Tumisoburo: "too many computer makers to cope with monster IBM"
- MITI pressured to reorganize Japan's computer industry
- VLSI project launched 1976 (Very Large Scale Integrated)
- Budget: $281M USD total, 50% government funded
- Duration: 4 years (ending 1980 - IBM's expected one-chip computer release)
- Goal: develop next-generation semiconductor device technology
VLSI Project Structure & Participants
- Five companies: Fujitsu, Hitachi, Mitsubishi Electric, NEC, Toshiba
- Excluded: consumer electronics (Sharp, Sony) - only R&D intensive computing companies
- Managing Director: Masato Nabashi (retired MITI bureaucrat)
- Organizational structure:
- Board of Directors (5 company presidents)
- Managing Director
- General Committee (company directors)
- Operating/Technical Committee (department heads)
- ~100 loaned employees (chosen by lab head, not companies)
- Lab types:
- Core cooperative lab (fundamental research, main office)
- Group labs (applied technology research)
- Location dispute: 6 months to choose cooperative lab location (eventually NEC Kanagawa)
Project Management & Culture
- Initial hostility between competing companies
- Nabashi's approach: "typical Japanese way" - frequent drinking sessions
- Goal: make researchers friends, eliminate problems
- Success: farewell party March 1980, equipment divided fairly, alumni newsletter created
- Results: >1000 new patents, 16% joint inventions between competitors
- Advanced indigenous semiconductor manufacturing knowledge beyond "fast followers"
Lithography Technology Development
- Six research goals: lithography, crystal, design, processing, testing, device creation
- Lithography achieved furthest progress
- Built three electron beam lithography types (≤1 micrometer resolution)
- Commissioned equipment from suppliers with three required expertise areas:
- High resolution optics
- Advanced electron beam/X-ray lithography techniques
- High precision mechanics
Canon & Nikon Partnership
- Canon background:
- 1970: began "Micron Project" for IC industry aligners
- 1973: PLA-300 introduced but not market competitive (accuracy issues)
- VLSI commissioned proximity aligner → PLA 500/600 series (massive hit)
- Nikon background:
- Better positioned: optics (medical microscopes since 1925) + mechanics (observatory telescopes)
- 1977: commissioned for lithography stepper development
- Never done steppers before but worked closely with VLSI members
- Neither company part of core VLSI consortium
- VLSI provided two critical supports:
- Foundation customer willing to pay for iterative development
- Clear requirements and close collaboration
Market Breakthrough & American Decline
- Nikon development method: purchased GCA stepper, reverse-engineered, built prototype
- 1980: Nikon released NSR-1010G stepper (first customers: NEC, Toshiba)
- 1980-1982: Japan lithography industry grew 66% vs US 11%
- Japanese advantages: optimized manufacturing workflows, better prices, older technology pushed farther
- Canon shocked industry: proximity aligner reaching 2-3 micron design rules (thought impossible)
- American vulnerabilities:
- 1981 recession year
- Poor market intelligence (relied on Japanese trading houses)
- GCA production constraints (Carl Zeiss lens delivery delays)
- Result: Japanese semiconductor makers defected, Nikon captured 20% market share in one year
- 1989: Japanese firms held 70% market (Canon, Nikon leading)
- American lithography industry collapse: one company lost $100M by 1986, other withdrew by 1989
Historical Context & Lessons
- Japanese victory temporary - couldn't maintain lithography lead
- Late 1980s-90s: Japanese semiconductor market collapsed
- American response: developed revolutionary technologies, created Sematech
- Next generation: EUV lithography → European ASML emergence
- Caution against direct parallels to current East Asian lithography situations
