AI summary by TwinMind
What Happened to Singapore's TSMC?
Asianometry
16m 42s
Company Origins and Formation (1987-1999)
- Singapore Technologies Group (STG) selected to spearhead Singapore's semiconductor entrance in 1987
- Technology transfer agreement with US fabless semiconductor maker Sierra Semiconductor
- Originally hoped for support from Sierra and National Semiconductor, but National dropped out
- Restructured to copy TSMC's independent foundry business model after seeing its success
- Spun off integrated circuit design house, moved forward as pure foundry
- Rebranded as Chartered Semiconductor Manufacturing (CSM) in 1999
- Went public on NASDAQ and Singapore stock markets in 1999
- Restructured in 2000 as majority government-backed entity under sovereign wealth fund Temasek
Technology Development and Early Growth (1987-1994)
- Sierra transferred 3.0 micron process technology and trained 100 Singaporean technical staff
- Process technology was outdated, needed rapid advancement
- Established in-house R&D lab immediately
- Rapid technology progression:
- 1990: 1.5 microns
- 1991: 1.2 microns
- 1992: 0.8 microns
- 1993: 0.6 microns
- 1994: Deal with Toshiba for 0.5 micron process technology
Expansion and Government Support (1990s)
- Rapidly expanded production facilities throughout 1990s
- Built 5 foundries on 60 hectares in northern Singapore
- Plans drawn for up to 25 foundries assuming success
- Expanded product assortment for complete turnkey solutions
- Singapore government provided:
- Tax incentives and subsidies for multinationals
- Training cost funding for companies
- State-run university programs for electrical engineers
- Achieved third place in foundry industry behind TSMC and UMC
Market Pressures and Competition (2000s)
- Caught in pincer movement that squeezed them out of market
- Chinese foundries entered market in 2000:
- Semiconductor Manufacturing International Corporation (SMIC) most notable
- Founded by former TSMC vice president
- Leveraged government subsidies and cheaper costs
- Same technology licenses available to them as CSM
- TSMC brought battle to Singapore in 1998:
- Joint venture with Phillips Electronics to start SSMC
- Invested 2 billion Singapore dollars over 3 years
- Trailing edge but good enough for many customers
Financial Performance Decline
- Could not differentiate product from cheaper Chinese and Taiwanese competitors
- Gross margin indicators:
- 2005: 11% gross margin
- 2003: -17.5% gross margin (down year)
- Same period TSMC: 44% gross margins
- Needed to reach leading edge to differentiate product
Samsung Entry and Technology Race (2005)
- Samsung announced standalone foundry business entry in 2005
- Leveraged immense resources and expertise
- Quickly rocketed to second place ahead of CSM and UMC
- Already biggest electronics and semiconductor company globally
- Owned largest share of digital memory oligopoly
- Had captive customer base despite claimed division separation
Escalating Costs and Technical Challenges
- Leading edge fab costs escalated dramatically:
- 2005: $3 billion (6x higher than 1995)
- 2020: TSMC's N5 fab cost $17 billion
- N3 process fab estimated at $19.5 billion
- CSM tried partnering with IBM to develop new edge processes
- Partnerships slowed product development, fell behind curve
- Technology partners licensed learnings to competitors within few years
Final Struggles and Acquisition (2009)
- Singapore government allocated hundreds of millions but needed billions
- 22 years of consecutive financial losses
- Abu Dhabi's Global Foundries acquisition strategy:
- Build global semiconductor champion using oil profits
- Buy multiple foundries to assemble global giant
- Teach UAE citizens STEM skills and business savvy
- September 2009: Global Foundries offered 2.68 Singapore dollars per share
- Price only 2 cents higher than closing price
- Total acquisition: 2.5 billion Singapore dollars (1.8 billion USD)
- Temasek backed deal, no premium needed
- Stock had risen 100% in previous months
Global Foundries Outcome
- Never executed successfully on business strategy
- Could not reach necessary scale for escape velocity
- TSMC and Samsung locked in premium customers at deep technical level
- Post-2010 advanced processes only available from TSMC or Samsung
- Required separate teams for each foundry
- 2018: Global Foundries ended 7nm development, focused on specialized processes
Strategic Assessment and Lessons
- Singapore had strong foundation:
- Established strategy
- Deep multinational connections
- Well-educated, disciplined workforce
- Competent, patient government
- Semiconductor market like "race with no finish line"
- Competitors keep pushing finish line further
- Questions about benefit to ordinary Singaporeans:
- Significant percentage of jobs went to highly paid expat foreigners
- Limited change throughout 22 years of losses
- Singapore semiconductor sector continues as significant part of economy
- Creates jobs, exports, profits for partners
- "Home runs rare and hard to make, doubles and singles can be perfectly fine"
