Skip to content
AcquiredAcquired

TSMC founder Morris Chang

We flew to Taiwan to interview TSMC Founder Morris Chang in a rare English interview. In fact, the last long-form video interview we could find was 17 years ago at the Computer History Museum… conducted by the one-and-only Jensen Huang! This episode came about after asking ourselves a version of the Jeff Bezos “regret minimization” question: what conversations would we *most* regret not having if the chance passed Acquired by? Dr. Chang was number one on our list, and thanks to a little help from Jensen himself, we’re so happy to make it happen. Dr. Chang shares the stories of a few crucial moments from TSMC’s history which have only been written about in his (currently Chinese-only) memoirs, including how TSMC won Apple’s iPhone and Mac chip business and a 2009 discrepancy with NVIDIA that almost jeopardized their relationship, and the lessons he took from them. We can’t think of a better way to kick off 2025. Please enjoy! Sponsors: Many thanks to our fantastic Spring ‘25 Season partners: J.P. Morgan Payments https://bit.ly/acquiredJPMPmorrisyt ServiceNow https://bit.ly/acquiredsn Fundrise https://bit.ly/acquiredfundrise25 Links: Worldly Partners’ Multi-Decade TSMC Study https://worldlypartners.com/businesshistory Karina Bao's writing https://karinabao.substack.com/ Carve Outs: AAA https://www.aaa.com Defunctland https://www.youtube.com/defunctland Everything Everywhere all at Once https://www.imdb.com/title/tt6710474/ Asianometry https://www.youtube.com/asianometry More Acquired: Get email updates with hints on next episode and follow-ups from recent episodes https://www.acquired.fm/email Join the Slack http://acquired.fm/slack Subscribe to ACQ2 https://pod.link/acquiredlp Check out the latest swag in the ACQ Merch Store! https://www.acquired.fm/store 0:00 - Intro 3:22 - TSMC and Nvidia's relationship 9:50 - The 40nm node development problem in 2009 20:55 - The TSMC Laid Off Employee Protest 26:16 - TSMC's Incredible Core Business vs. Expansions 29:08 - Resolving the 2009 NVIDIA dispute 35:59 - TSMC's commitment to 28nm 43:04 - Going all-in on the 28nm node 52:05 - Meeting Apple and Jeff Williams 1:23:09 - Goldman Sachs 1:34:57 - ServiceNow 1:36:01 - Intel's and Apple 1:46:45 - Apple’s and the 20nm node 1:47:58 - Pricing 1:50:24 - Apple and trade-offs 1:55:22 - The IBM-Qualcomm story 2:02:18 - The Learning Curve 2:10:45 - The Flywheel of Returns 2:14:18 - TSMC’s Unexpected(?) Success 2:15:57 - The early days of TSMC 2:21:10 - Ben and David's Reflections 2:27:43 - Taiwanese Science Park 2:47:08 - Carve Outs Notes: This episode contains a paid endorsement for Fundrise. All investments can lead to loss. Acquired hosts and guests may hold assets discussed in this episode. This podcast is not investment advice, and is intended for informational and entertainment purposes only. You should do your own research and make your own independent decisions when considering any financial transactions.

Ben GilberthostDavid RosenthalhostDr. Morris Changguest
Jan 27, 20252h 54mWatch on YouTube ↗

CHAPTERS

  1. 0:00 – 3:22

    Cold open and why this Morris Chang interview happened (Taipei trip + memoir translation)

    Ben and David set the stage for a rare in-office interview with TSMC founder Morris Chang, arranged via Jensen Huang. They explain how Chang’s newly released (but not yet English-published) autobiography volume guided the conversation toward pivotal stories about NVIDIA, Apple, and the fabless revolution.

    • Acquired’s semiconductor “tour” since the original TSMC episode (NVIDIA, Qualcomm, Arm, Synopsys)
    • How Jensen Huang helped secure the interview with Morris Chang
    • Taipei whirlwind visit and time at Hsinchu Science Park / TSMC HQ
    • Memoir context: unpublished English translation and what stories they’ll focus on
  2. 3:22 – 9:33

    TSMC’s early relationship with Jensen Huang and a near-missed NVIDIA opportunity (1997)

    Chang recounts receiving a physical letter from Jensen Huang when NVIDIA was tiny and in trouble—and being irritated that TSMC’s US office hadn’t responded. He personally calls Jensen, meets him, and is impressed by his clarity and optimism, setting the foundation for a decades-long partnership.

    • Jensen’s 1997 letter: NVIDIA is small, needs a foundry, and can’t get traction with TSMC sales
    • Chang’s principle: never ignore small prospective customers
    • First call and meeting: Jensen’s frankness about financial distress + bold revenue promise
    • NVIDIA’s chip succeeds; within a few years they become a top-5 TSMC customer
  3. 9:33 – 15:18

    The 40nm crisis (2009): yield and quality disputes under successor leadership

    Ben tees up the 40nm node problems that hurt NVIDIA, and Chang explains they occurred while he’d handed the CEO role to a would-be successor. The dispute centered on manufacturing yield issues and a contentious quality-fault argument that left NVIDIA uncompensated and frustrated.

    • 40nm problems: yield challenges plus a quality controversy
    • CEO-era decision: quality leadership claims TSMC not at fault → no offer to NVIDIA
    • Why 40nm mattered: critical step before moving to 28nm
    • Chang grows impatient and prepares to retake the CEO role amid multiple compounding issues
  4. 15:18 – 25:29

    The layoff backlash: protests at Chang’s home and a leadership turning point

    Chang describes a culturally and operationally shocking episode: 600–700 employees were effectively laid off via performance reviews during the financial crisis. Protests escalated to demonstrations at his home, culminating in a personal, human moment when his wife brought breakfast to protesters—helping precipitate Chang’s return as CEO.

    • Why “layoff by performance rating” is subjective and corrosive to trust
    • Economic logic: separation costs + retraining makes short-cycle layoffs self-defeating
    • Protests at Chang’s home; neighbors impacted; police presence
    • Sophie Chang’s breakfast gesture de-escalates and influences events
    • Chang’s frustration: CEO bypassed board oversight by rebranding layoffs as “poor performance”
  5. 25:29 – 29:27

    Core foundry dominance vs. ‘new businesses’ (solar + LED) and what went wrong

    Chang keeps the prior CEO but reassigns him to lead “new businesses” like solar and LEDs—areas the company hoped could complement core foundry growth. He explains why both initiatives ultimately failed: China’s solar subsidies crushed pricing, and LED patent control limited opportunity.

    • Chang’s view: IC foundry was always going to be great; expansions were additive, not replacement
    • Solar: China subsidies drive prices extremely low, undermining TSMC’s solar effort
    • LED: market smaller; key patents controlled by a few incumbents
    • Former CEO later departs; note on later leadership roles (MediaTek)
  6. 29:27 – 38:30

    Settling with NVIDIA: the pizza-and-salad negotiation and the $100M+ resolution

    After returning as CEO, Chang makes customer calls and prioritizes repairing the NVIDIA relationship. He personally investigates, computes what he believes is a fair figure, and delivers a take-it-or-arbitrate offer during a home dinner—preserving trust and enabling decades more partnership value.

    • Chang’s first weeks back: heavy focus on resolving the NVIDIA dispute
    • Framing: manufacturing fixes are essential for all customers, but NVIDIA bore disproportionate damage
    • Chang’s offer: >$100M, valid for 48 hours, no bargaining; arbitration as fallback
    • Outcome: Jensen accepts within two days; relationship strengthened for future billions in business
  7. 38:30 – 42:47

    Committing to 28nm: the 8% R&D rule and the ‘sweet spot’ thesis

    Chang explains how frustration at TI over R&D budgeting led him to set a fixed R&D percentage at TSMC: 8% of revenue regardless of recessions. With stable resources, R&D leaders argued 28nm would be a major performance-and-demand inflection (“sweet spot”), giving Chang conviction to push hard.

    • Lesson from TI: annual R&D bargaining is destructive; set a clear, durable rule
    • TSMC’s decision: lock R&D at 8% of revenue (even in downturns)
    • R&D argument: 28nm as “sweet spot” for the industry and performance
    • Strategic clarity: 40nm success is prerequisite to executing 28nm well
  8. 42:47 – 52:14

    Going all-in on 28nm: tripling capex, board skepticism, and the smartphone tailwind

    With market forecasting and business development inputs aligning, Chang authorizes a huge capex ramp (from ~$2–2.5B/year to nearly $6B) to capture the 28nm tide. He describes intense board scrutiny and how, in the end, leadership accountability—not perfect proof—won the argument.

    • 28nm-era capex surge: a major bet on capacity and leadership at the leading edge
    • Organizational enablers: market forecasting + strategic business development (marketing)
    • Board dynamics: independent directors push back hard; off-record pre-dinner Q&A tradition
    • Chang’s stance: ‘I’m responsible for operations—let me do this’
    • Result: 28nm aligns with the smartphone era and proves highly successful
  9. 52:14 – 1:05:50

    How CC Wei got the strategic ‘business development’ role—and why org design matters

    Before telling the Apple story, Chang explains a critical internal decision: consolidating advanced and mainstream groups and building a real strategic marketing/business development function. He details offering the role to Mark Liu first (who declines), then to CC Wei (who accepts)—a pivotal step in creating modern TSMC leadership capacity.

    • Foundry business fits functional org structure better than divisional (customer overlap, fab interchangeability)
    • Historical parallel: earlier attempt (Don Brooks) + McKinsey validation of functional structure
    • Two groups (advanced/mainstream) each had small BD teams; Chang consolidates them
    • Leadership development intent: broaden ops leaders with strategic/marketing experience
    • CC Wei accepts BD role—foreshadowing future CEO/chairman trajectory
  10. 1:05:50 – 1:21:00

    Apple arrives via Foxconn: the Jeff Williams dinner and the surprise 20nm request

    Chang recounts an unexpected evening: Terry Gou (Foxconn) brings an Apple executive to Chang’s home—who turns out to be COO Jeff Williams. Apple’s initial interest quickly becomes clear, but the technical ask shocks Chang: not 28nm (which TSMC is ramping), but a detour to 20nm.

    • Terry Gou’s visit setup and why it mattered (Foxconn as a key Apple supplier)
    • Jeff Williams’ style: direct, minimal small talk, immediate foundry request
    • Early margin talk: Jeff mentions “40% gross margin,” while TSMC is already ~45%
    • Big twist: Apple wants 20nm (a ‘half step’ detour) rather than 28→16 progression
    • Constraint: at the time, TSMC R&D can’t pursue two nodes in parallel
  11. 1:21:00 – 1:36:24

    Financing the Apple bet: Goldman Sachs, bonds, and taking only half the demand

    To fund Apple’s capacity needs on top of an already aggressive capex plan, Chang leans on a long-cultivated Goldman Sachs relationship. He walks through the decision set—cut dividends, issue stock, borrow, or partially fulfill—and explains why TSMC chose corporate bonds and only committed to half of Apple’s requested volume.

    • Why Goldman: relationship seeded early (ADR work; Chang served on Goldman’s board)
    • Financing options evaluated: dividends, new stock, debt, or limiting Apple allocation
    • Decision: borrow via corporate bonds; don’t cut dividends (dividend-sensitive shareholder base)
    • Customer negotiation reality: Apple’s incentives vs. TSMC’s capital risk → TSMC takes half demand
    • Foundry capacity discipline: deposits/commitments as protection against over-ordering
  12. 1:36:24 – 1:46:46

    Intel enters the Apple conversation: ‘Intel doesn’t know how to be a foundry’

    Jeff Williams pauses discussions after Intel’s CEO approaches Tim Cook, raising the possibility of Intel as an iPhone chip supplier. Chang isn’t overly worried and later visits Apple, where Tim Cook delivers a concise verdict: Intel lacks foundry operating culture—especially the customer-centric responsiveness required.

    • Jeff’s call: pause for two months due to Intel outreach at the highest level
    • Chang’s competitive assessment (2011): tech near parity, TSMC manufacturing stronger, higher customer trust
    • Tim Cook’s message: Intel doesn’t know how to be a foundry
    • Interpretation: not just conflict-of-interest—foundry is a service business with demanding customers
    • Implication: why TSMC’s pure-play model is inherently advantaged for Apple
  13. 1:46:46 – 1:55:18

    20nm and beyond: pricing battles, R&D trade-offs, and the Samsung 16nm scare

    Chang explains the practical friction in making Apple work: pricing must clear both cost realities and Apple’s component budgets. Strategically, the 20nm detour delays 16nm, allowing Samsung to get ahead—leading to a moment of alarm when Apple places initial 16nm orders with Samsung before ultimately returning most volume to TSMC.

    • Pricing fundamentals: cost first, then customer acceptance; negotiation can be tense
    • Real trade-off: 20nm work delays 16nm because R&D couldn’t do both simultaneously then
    • Competitive consequence: Samsung ‘skips’ 20 and advances on 16nm earlier
    • Crisis moment: Apple’s first 16nm orders go to Samsung; Chang emails Jeff immediately
    • Resolution: Apple commits to shift to TSMC once 16nm is ready; most 16nm volume returns
  14. 1:55:18 – 2:02:18

    Reading the industry from the foundry seat: IBM–Qualcomm and refusing ‘co-development’

    Chang shares how foundry vantage reveals upstream trouble early: as Qualcomm shifts business away from IBM, Chang infers IBM Semiconductor is in distress. When IBM proposes co-developing 130nm, TSMC refuses—avoiding dependency and cultural friction—while UMC accepts and later regrets it.

    • Qualcomm’s early reluctance; IBM as its primary foundry supplier in the 1990s
    • Signal: Qualcomm’s pivot to TSMC implies IBM’s foundry business is failing
    • IBM asks TSMC to co-develop 130nm; TSMC declines immediately
    • Why refusal mattered: avoid becoming junior partner and losing process autonomy
    • IBM turns to UMC; Chang notes strategic regret for UMC later
  15. 2:02:18 – 2:21:08

    The learning curve and the ‘flywheel of returns’: scale as destiny in semiconductors

    Chang traces his relationship with BCG’s Bruce Henderson and Bill Bain at TI, where he helped refine the learning/experience curve into an actionable management tool for semiconductors. The discussion connects volume, declining unit cost, aggressive positioning, and reinvestment—forming the strategic flywheel that can concentrate industry leadership.

    • Learning curve basics: more output → lower unit cost; but real value is strategic application
    • Origin story: Bruce Henderson brings experience curve to TI; Bill Bain embeds with Chang
    • How the tool shapes behavior: race through low-volume phase; scale creates advantage
    • Flywheel logic: lower cost → better pricing → more volume → more reinvestment capacity
    • Chang’s takeaway: a practical thinking framework used throughout his career
  16. 2:21:08 – 2:46:53

    Post-interview reflections: pure-play foundry, Arm’s role, Hsinchu’s ecosystem, and the Taiwan ‘rifle shot’

    Ben and David debrief the interview and connect dots across semiconductor industry structure. They emphasize the non-competition principle, the enabling role of Arm and other modular value-chain companies, and what visiting Hsinchu Science Park made tangible: an ecosystem density that’s extremely hard to replicate elsewhere.

    • Why ‘don’t compete with customers’ became a decisive long-term advantage
    • Industry modularization: architecture (Arm), EDA, equipment, design, and manufacturing separating into specialists
    • Hsinchu Science Park as physical proof of tight collaboration (Arm/Synopsys/Cadence/customers/universities)
    • Why recreating TSMC abroad is so hard: ecosystem, talent pipelines, supplier proximity, iteration speed
    • Government initiative success framing: a highly targeted industrial policy win
  17. 2:46:53 – 2:54:55

    Carve-outs and closing credits: recommendations and episode thank-yous

    The episode ends with carve-outs—AAA, Defunctland, and Everything Everywhere All at Once—before a long list of industry experts who helped the hosts prepare. Ben and David point listeners to related semiconductor episodes (TSMC remaster, NVIDIA, Qualcomm) and invite community discussion.

    • Carve-outs: AAA membership; Defunctland YouTube channel; Everything Everywhere All at Once
    • Prep thank-yous: semiconductor leaders, analysts, and prior Acquired guests
    • Pointers to the broader Acquired semiconductor catalog (TSMC, NVIDIA, Qualcomm, Arm/Synopsys on ACQ2)
    • Where to engage: Slack, email list, show notes and linked research

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.