AI summary by TwinMind
The Birth of Taiwan’s Semiconductor Industry
Asianometry
19m 34s
Taiwan's 1970s Economic and Political Challenges
- 1971: Republic of China lost UN seat to People's Republic of China
- September 1972: Taiwan broke diplomatic relations with Japan (critical ally)
- 1973: Oil crisis struck Taiwan's economy heavily dependent on plastics/heavy industry
- Growth fell drastically, opening government's eyes to needed changes
Early Electronics Industry Foundation
- Electronics industry was Taiwan's second largest exporter
- Mostly consisted of low-value manual assembly operations
- First factory built by General Instrument in 1966
- Followed by Phillips and Texas Instruments
- Local entrepreneurs founded Orient Semiconductor and Lingen Precision Industries
- Limited semiconductor engineering knowledge overall
National Chiao Tung University (NCTU) Role
- Founded in 1957 in Hsinchu with help from former Shanghai Jiao Tong University alumni
- Started Taiwan's first semiconductor lab in 1964
- Known for recruiting Chinese American professors from US
- Many Taiwanese EE graduates went abroad due to weak local semiconductor industrial base
- Later merged with National Yang Ming University → National Yang Ming Chiao Tung University (NYCU)
Government Policy Development
- Initial goals: computers and consumer electronics, not semiconductors
- 1965: Vice Minister Walter Fay received suggestion from RCA researcher Pan Wen-yen
- Funded tech forums with speakers from Bell Labs and IBM
- 1970: Clear semiconductors had big role in electronics
- Premier Chiang Ching-kuo (son of Chiang Kai-shek) initiated 10 major construction projects
- Ministers with common university ties met in breakfast shops planning semiconductor advancement
Industrial Technology Research Institute (ITRI) Formation
- 1973: Minister of Economic Affairs Sun Yun-suan established ITRI
- Consolidated three existing institutes: chemistry, metal, mining
- Mission: acquire, introduce, transfer new technologies to private industry
- September 1, 1974: Electronics Industrial Research Organization (ERSO) became ITRI's fourth lab
- Name emphasized industrial goal was electronics products, not semiconductors
Technical Advisory Committee (TAC) and Technology Selection
- Founded 1974, staffed with Chinese nationals working for American companies (IBM, Bell Labs)
- Led by Dr. Pan Wen-yen, many attended Princeton University
- Selected Complementary Metal Oxide Semiconductors (CMOS) technology
- Departure from American producers using bipolar junction transistors or NMOS
- CMOS had reputation for being tricky but offered low power consumption
- Focus on consumer electronics applications
RCA Technology Transfer Agreement
- 1974-1975: Down years for semiconductor industry, companies willing to teach for right price
- ERSO struck agreement with RCA for 7-7.5 micron CMOS technology
- RCA chosen over Texas Instruments (no CMOS) and others
- RCA was CMOS pioneer, had Taiwan TV factory, good reputation
- Technology not cutting-edge (US leading edge was 2-3 microns)
- Focus on completeness of information transfer, not advancement
- Contract required training 10 semiconductor designers
- RCA originally wanted $7.5 million, accepted $3 million after negotiation
Taiwan Team Training at RCA
- May 1976: 30-40 individuals sent to RCA plants for 6-12 months
- Well-educated team: bachelor's degrees minimum, many with master's/PhDs
- Many trained at NCTU semiconductor laboratory
- Three groups:
- Semiconductor design at RCA headquarters
- Manufacturing/fab/quality control at Findlay, Ohio and West Palm Beach, Florida
- Procurement and equipment testing
- RCA used "big brother/little brother" teaching method
- Unionized factory required month negotiation for Taiwanese to work on line
- Union rules: desks movable but experiment tables not, three people to change lightbulb
- Dr. Yang Ding-yuan instructed team to record everything, even toilet paper and cleaning materials
Demonstration Factory Construction
- 1977: ITRI started building first 3-inch wafer demonstration factory
- Complete copy of original RCA fab, same equipment from suppliers
- Apcon CAD system for integrated circuit design
- RCA resisted any deviations from contract
- Missing capability: photo mask production (templates containing chip designs)
- Found overseas Chinese to source masks for first product: high-altitude military propaganda balloon chip
- Heavy reliance on consultants for knowledge gaps
- RJ Seagull company designed clean room
- Former RCA supervisor brought on for additional color
- Final cost: ~$9 million
Demonstration Factory Performance
- RCA contract guaranteed 17% yield
- Demo fab started at 70% yield, reached 81% by December
- Surpassed original RCA factory performance
- Success attributed to newer equipment (ion implanter) and better workers
- Many workers overqualified, went on to great things in Taiwan semiconductor industry
- Taiwan already manufacturing powerhouse: 17% of world's TV sets made in Taiwan (1975)
- Minister Sun Yun-suan: RCA project was Taiwan's only successful comprehensive learning project
Commercialization and UMC Formation
- 80% yields correlated to net profit rate over 21%
- Ministry of Economic Affairs decided needed commercial operation
- ITRI's intent: transfer to private industry (much of Taiwan's economy state-owned)
- No existing private company wanted to run fab
- Late 1979: ITRI set up preparatory office for United Microelectronics Corporation (UMC)
- UMC received second factory with upgraded 3-micron process on 6-inch wafers
- ITRI invested millions in computer simulation programs for IC designs
- Twisted various Taiwanese companies to invest small amounts for stakes
- ITRI selected products: telephones, toys, timers, calculators, watches
- One chip played "Happy Birthday" melody for greeting cards
- Founded 1980, operations began 1982
- After initial struggle, became profitable producing telephone dialers for US market post-AT&T breakup
Hsinchu Science Park Development
- UMC among first occupants of Hsinchu Science Park (Taiwan's first industrial science park)
- Emulated Silicon Valley model: single company next to universities
- Minister without portfolio Dr. K.T. Lee implemented idea after successful export zone in Kaohsiung/Taichung
- Chose Hsinchu for two major universities: NCTU and National Tsing Hua University
- NCTU originally small, government expected to retake mainland and restore original campus
- When retaking mainland unlikely, NCTU explored relocation to Kaohsiung
- Hsinchu local council appealed to Premier Chiang Ching-kuo
- Chiang directed military to transfer 2,000 hectares for NCTU's new campus
- Park opened December 1980, struggled to find occupants initially
ITRI Spin-off Strategy and Design Houses
- ITRI acquired technologies via import/foreign partnerships, spun off as startups
- Since founding, ITRI spun off over 162 companies
- UMC and TSMC most famous, others include Vanguard International, Phison Electronics, Syntech
- As park filled, overseas Chinese returned with ideas founding design houses
- Design houses: designers with product ideas but no fab (fabless before term existed)
- Found no foundry capacity available, contracted with Japanese semiconductor makers
- Caught attention of ITRI chairman Morris Chang
- Chang proposed company to produce anyone's chip designs → became TSMC
Strategic Vision and Success Factors
- Taiwan didn't set out to build semiconductor industry as end goal
- Wanted complete IT industry: computers, tools, instruments with ICs at heart
- ICs were means to goal, not goal itself
- Major reason for success: comprehensive approach
- Closely studied and emulated American Silicon Valley model
- Extensively tapped Chinese American networks who brought learnings back home
Key Personnel Ages
- Hu Ding-hua led RCA technology transfer project at age 33
- Dr. Yang Ding-yuan selected and trained RCA team, turned 28 in 1976
- Young leaders motivated by opportunity, not money
- Could earn more elsewhere but believed in great opportunity ahead
- True pioneers blazing new journey for their nation
