At a glance
WHAT IT’S REALLY ABOUT
Morris Chang on TSMC’s customer trust, nodes, and compounding scale
- Acquired interviews 93-year-old TSMC founder Morris Chang, focusing on pivotal episodes that shaped TSMC’s leadership: the early NVIDIA partnership, the 2009 40nm crisis, and the strategic “all-in” commitment to 28nm.
- Chang describes how he resolved a major NVIDIA dispute through direct executive engagement, a time-boxed settlement, and a long-term partnership mindset—turning a painful yield/quality period into enduring trust.
- He explains TSMC’s deliberate decision to institutionalize R&D at 8% of revenue and dramatically increase capital spending to capture the smartphone-era “sweet spot” at 28nm, despite board resistance.
- The conversation then traces how Apple became a customer (via Foxconn’s Terry Gou introducing Jeff Williams), why Intel failed to win Apple’s foundry business, and how node roadmaps (20nm detour, 16nm catch-up) created real strategic trade-offs.
IDEAS WORTH REMEMBERING
5 ideasCustomer trust is a strategic asset that can be “bought back” decisively.
In the 40nm dispute, Chang personally re-engaged Jensen Huang, quantified harm, and offered a >$100M settlement with no bargaining and a 48-hour deadline—avoiding arbitration and preserving a relationship that later produced many billions in business.
Operational crises compound when leadership avoids accountability.
Chang recounts that the prior CEO leaned on internal quality claims to deny fault (“zero”), prolonging conflict with NVIDIA; Chang’s fix began by re-opening facts, aligning incentives, and treating restitution as a trust reset, not a legal fight.
Avoid layoffs if you’ll need the talent back within a year.
Chang argues performance-based layoffs lack credibility due to supervisor subjectivity and destroy value when severance plus retraining roughly equals a year; in Moore’s Law businesses, capacity to execute future nodes depends on retaining trained teams.
Make R&D non-negotiable to sustain a Moore’s Law cadence.
After being denied incremental R&D at TI, Chang set TSMC’s R&D at a fixed 8% of revenue “regardless of recession,” removing annual bargaining and enabling teams to plan aggressively for successive nodes like 28nm (and later “sweet spots”).
Winning a node often requires synchronized R&D conviction and capex courage.
TSMC tripled capex (≈$2–2.5B to ≈$6B) to seize 28nm, driven by internal technical confidence (“sweet spot”) and strategic marketing inputs; the board resisted, but Chang insisted management must be accountable for the bet.
WORDS WORTH SAVING
5 quotesQuiet! Morris Chang is calling me.
— Dr. Morris Chang (recounting Jensen Huang’s reaction, 1997)
Our offer is effective forty-eight hours… We are not going to argue. We are not gonna bargain.
— Dr. Morris Chang
Intel just does not know how to be a foundry.
— Tim Cook (as relayed by Dr. Morris Chang)
There’s a tide in the affairs of man, which taken at its flood, leads on to fortune.
— Dr. Morris Chang (quoting Shakespeare on the 28nm bet)
China ruined it. They subsidized the hell out of it.
— Dr. Morris Chang (on TSMC’s solar expansion)
High quality AI-generated summary created from speaker-labeled transcript.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome