AI Summary
→ WHAT IT COVERS Morris Chang, 93-year-old TSMC founder, shares firsthand accounts of building relationships with Jensen Huang and Apple, resolving the 40-nanometer crisis, committing $6 billion to 28-nanometer production, and establishing Taiwan's pure-play foundry model that created a trillion-dollar semiconductor manufacturing company. → KEY INSIGHTS - **Customer relationship management:** Chang personally visited Jensen Huang's home in 2009 with a $100+ million settlement offer for 40-nanometer yield problems, giving 48 hours to accept without negotiation. This direct approach preserved TSMC's most important partnership and led to billions in future business with NVIDIA. - **R&D budget stability:** Chang set TSMC's research and development spending at exactly 8% of revenue regardless of economic conditions, eliminating annual budget negotiations. This guaranteed resource allocation enabled the R&D team to plan long-term technology development and contributed directly to achieving leadership at 28-nanometer and subsequent nodes. - **Capital allocation timing:** TSMC invested $6 billion in 28-nanometer capacity in 2010 despite board opposition, tripling previous capital spending. Chang convinced directors by citing inputs from market forecasting, R&D predictions of a "sweet spot" node, and business development analysis, betting correctly on smartphone demand driving semiconductor growth. - **Employment philosophy during downturns:** Chang refused to conduct layoffs based on performance reviews during the 2008 financial crisis, arguing that subjective ratings from 700 different supervisors lack credibility. He maintained that semiconductor companies always need people due to Moore's Law progression, making layoffs counterproductive when rehiring costs equal separation expenses. - **Pure-play foundry advantage:** TSMC's decision to never compete with customers in chip design proved decisive when Apple chose TSMC over Intel for iPhone manufacturing. Intel's inability to act as a true foundry partner, combined with customer trust issues from their historically superior attitude, created an insurmountable competitive moat for TSMC. → NOTABLE MOMENT When Apple requested 20-nanometer production instead of the planned 28-nanometer node, Chang faced a dilemma: invest billions in an unplanned technology detour or lose Apple as a customer. He chose to accept only half of Apple's requested capacity, requiring corporate bond issuance, which later proved prescient when Samsung temporarily won 16-nanometer business. 💼 SPONSORS [{"name": "Rippling", "url": "rippling.com/acquired"}, {"name": "Statsig", "url": "statsig.com/acquired"}] 🏷️ Semiconductor Manufacturing, TSMC History, Customer Relations, Capital Allocation, Technology Leadership, Taiwan Industry