Latest / Boardroom Conversations with Fexingo: CEOs, Strategy, and Corporate Leadership Explained / How the TSMC Board Built the Chip-Manufacturing Fortress
Transcript
- Lucas: Taiwan Semiconductor Manufacturing Company. TSMC. It makes roughly ninety percent of the world's most advanced chips — the ones that power your iPhone, your data center, your car's driver-assist system. And its board of directors is one of the least flashy, most technically focused boards in the Fortune 500. Luna: No celebrity directors, no ex-politicians cashing in. Just a lot of PhDs and engineers. Lucas: Exactly. As of mid-2026, TSMC's ten-person board includes four independent directors, all with deep technical backgrounds. There's no one from private equity, no retired general, no former Treasury secretary. It's very much a board designed to preserve the company's core competency: process technology leadership. Luna: So how did that board come together? And how does it make decisions when the stakes are literally billions of dollars and national security? Lucas: Let's start with the founder, Morris Chang. He founded TSMC in 1987 at age fifty-five, after leaving Texas Instruments. He was a legendary engineer and manager, but he also understood governance in a way most founders don't. He deliberately structured the board to balance founder control with independent oversight. From the beginning, he brought in outside directors with technical credibility—not just friends or politicians. Luna: And he stepped down as chairman in 2018, handing the reins to Mark Liu, who had been co ceo. That's a pretty smooth transition for a Taiwanese company. Lucas: It was planned for years. Chang had been grooming Liu and the current CEO, C.C. Wei, for a long time. The board was part of that succession design. Now, let me give you a concrete example of how this board operates under pressure. In 2020, TSMC announced it would build a $12 billion advanced fab in Arizona. That was a huge strategic shift — TSMC had never built a leading-edge fab outside Taiwan before. Luna: And that decision couldn't have been just a management call. The board would have to approve a multi billion dollar capital commitment with massive geopolitical implications. Lucas: Right. And the board didn't just rubber-stamp it. According to people familiar with the discussions, the independent directors — particularly Michael Splinter, a former Applied Materials CEO, and Kok Peng Soon, a former Singaporean diplomat — pushed management hard on the risk analysis. They asked: what happens if we can't find enough engineers in Arizona? What if the political climate shifts? How do we protect our intellectual property? The board made management come back with a detailed risk mitigation plan before giving the green light. Luna: That's the value of directors who actually understand the technology and the supply chain. They can ask the right follow-up questions. Lucas: Exactly. And since then, TSMC has expanded that Arizona commitment to three fabs, totaling over $40 billion. The board has had to revisit the decision multiple times as costs escalated and construction timelines slipped. Each time, they've reaffirmed the strategy, but they've also demanded updates on workforce development and water usage — real operational details. Luna: Water usage is a huge issue in Arizona. A fab uses millions of gallons a day. The board had to be comfortable with the sustainability plan. Lucas: And they were. Now, let's talk about capital allocation more broadly. TSMC's capital expenditures have been staggering — over $100 billion between 2021 and 2025. That's more than most countries' GDP. A typical board would get nervous about that level of spending. But TSMC's board understands that in this industry, you either spend aggressively on the next node, or you fall behind and never catch up. Luna: The classic 'innovator's dilemma' dynamic. If the board had tried to preserve margins by cutting capex, they'd have lost their technological lead. Lucas: They saw what happened to Intel. Intel's board, dominated by financial executives, allowed management to underinvest in process technology for years. By the time they realized their mistake, TSMC was two full nodes ahead. Intel's board has since been overhauled, but the damage is done. Luna: So the TSMC board's technical fluency is a competitive advantage. But it also creates a risk — groupthink. If everyone around the table is a process engineer, who challenges the engineering orthodoxy? Lucas: That's a fair point. And TSMC's board does include some non-technical perspectives. Sir Peter Bonfield, the former CEO of BT Group, brings telecom and European regulatory experience. There's also a legal expert and an academic. But you're right — the balance is heavily weighted toward technical expertise. And that's by design. Morris Chang once said that a board of directors should be 'a group of people who can tell the CEO something he doesn't want to hear.' At TSMC, that tough feedback tends to come from fellow engineers. Luna: Speaking of tough feedback — how does the board handle geopolitical risk? TSMC is based in Taiwan, which China claims as its own. That's the single biggest risk to the company. Lucas: It's front and center in every board meeting. The board has a dedicated geopolitical risk committee. They've modeled scenarios from mild trade disruptions to a full blockade. And they've used those scenarios to drive strategic decisions — like building the Arizona fabs, but also expanding in Japan and Germany. The board's view is that TSMC must become more geographically diversified, but not so much that it dilutes its R&D concentration in Taiwan. Luna: That's a delicate balancing act. The U.S. government wants TSMC to build more here. The Taiwanese government wants them to keep the crown jewels at home. The board has to navigate both without alienating either. Lucas: And they've done it remarkably well. Let me give you a specific recent example. In 2024, TSMC's board approved a plan to build a second fab in Germany, in Dresden, with a focus on automotive chips. That decision wasn't just about capacity — it was about building political goodwill in Europe and reducing the perception that TSMC is a Taiwan-only supplier. The board saw it as insurance against future trade restrictions. Luna: So each new fab has both a business rationale and a geopolitical rationale. That's sophisticated governance. Lucas: It is. And it's worth noting that TSMC's board meets in person every quarter, but the full board also gathers for a two-day strategic retreat every year. Those retreats are where the big-picture questions get hashed out — should we buy a memory company? Should we invest in a chiplets ecosystem? The agenda is set by the independent directors, not management. Luna: That's a best practice that a lot of boards claim to follow but don't really. Independent directors setting the agenda means they control what gets discussed, not just what management wants to present. Lucas: And that's why TSMC's board has been able to maintain strategic focus for nearly forty years. They've never been tempted to chase short-term profits at the expense of R&D. They've never paid a dividend that would constrain capex. In fact, TSMC only started paying a regular dividend in 2015, and it's very modest — less than two percent yield. The board's philosophy is that retained earnings belong in the fabs. Luna: That discipline is rare. Most boards would face pressure from activist investors to return more cash to shareholders. Lucas: TSMC has faced that pressure. In 2020, an activist fund built a stake and argued that TSMC should spin off its R&D into a separate entity to unlock value. The board listened politely, analyzed the proposal, and then publicly rejected it. They said the integrated model is what gives TSMC its advantage — keeping process R&D and manufacturing together allows faster learning cycles. The activist eventually sold. Luna: And the stock has done just fine without that spinoff. TSMC's market cap is over a trillion dollars now. Lucas: It's about one point two trillion as of this week. Which brings us to another board responsibility: CEO succession. Mark Liu is seventy-three. C.C. Wei is seventy. The board needs to have a plan for the next generation. And they've been deliberate about it. In 2024, they promoted three senior vice presidents to co-chief operating officers, essentially running a horse race to see who will eventually take over. Luna: That's a classic approach — let multiple candidates prove themselves over a few years. The board can evaluate them based on actual performance, not just interviews. Lucas: And the board is deeply involved in that process. The independent directors meet regularly with each candidate, without the current CEO present. They ask about their vision for technology, their views on geopolitics, their management style. It's thorough. Luna: It sounds like the board is very hands-on with strategy, but does it also hold management accountable for execution? I mean, TSMC has had some manufacturing hiccups — the 3-nanometer yield issues in 2023, for example. Lucas: Great question. Yes, the board holds management accountable. When yield problems emerged on the 3-nanometer node, the board's audit committee and the technology committee both got involved. They required management to present a weekly yield improvement plan for six months. The board members with semiconductor backgrounds — like Michael Splinter — could actually evaluate whether the technical fixes were credible. They didn't just accept management's assurances. Luna: That's the kind of oversight that only works when directors have domain expertise. A board of generalists couldn't have done that. Lucas: Exactly. And that's the lesson from the TSMC board. If your company's competitive advantage is technological depth, your board needs to reflect that depth. Not every board should look like TSMC's — a consumer goods company might benefit from a marketing expert. But the principle is the same: board composition should match strategy. Luna: I think one of the things that's underappreciated is how the board handles the relationship with Apple, TSMC's biggest customer. Apple accounts for about twenty-five percent of TSMC's revenue. That's enormous customer concentration risk. How does the board manage that? Lucas: The board is very aware of the risk. They've encouraged management to diversify the customer base, which they've done — NVIDIA, AMD, Qualcomm, MediaTek are all growing faster than Apple within TSMC. But the board also recognizes that Apple's demand for the latest node forces TSMC to innovate faster. It's a double-edged sword. The board's approach is to maintain a partnership with Apple while ensuring that no single customer becomes indispensable. Luna: And part of that is not letting Apple dictate terms. TSMC has famously refused to give Apple preferential pricing or exclusivity on nodes. The board backs management on that stance. Lucas: They do. And that's another example of the board's long-term orientation. They'd rather have a slightly lower margin with Apple than cede control of their own pricing and technology roadmap. Luna: One last thing — the board's composition is very Taiwanese. Nine of the ten directors are Taiwanese nationals. The only foreigner is the former Applied Materials CEO. Does that create a blind spot? Should the board be more international? Lucas: It's a legitimate criticism. For a company that operates globally and faces geopolitical risks from multiple directions, having a more internationally diverse board could bring fresh perspectives. But TSMC's view is that deep understanding of Taiwanese culture, business norms, and government relations is essential for the board to function effectively. They've added international advisors at the management level instead. Whether that's sufficient will be tested if geopolitical tensions escalate further. Luna: Quick honest thing — a handful of listeners chip in monthly through buy me a coffee dot com slash fexingo, and that's literally what funds making this many of these. If these conversations have moved your work forward in some small way, that's the only reason we can keep doing them ad-free. Lucas: Yeah, it's a small group but it makes a big difference. And we put everything into each episode, so it's appreciated more than we probably say. Back to the board — I think the final takeaway is that TSMC's board has been remarkably consistent. They've stuck to the same principles for decades: technical expertise, long-term capital allocation, independence from short-term market pressure, and a willingness to make unpopular decisions when necessary. Luna: And the result is a company that dominates the most critical industry in the world. It's a case study in governance as competitive advantage. Lucas: It really is. The question for the next decade is whether that governance model can survive the transition to a post-Morris Chang, post-Mark Liu era. The board has a succession plan, but succession is about more than just the CEO — it's about preserving the culture of technical rigor at the board level. If they can do that, TSMC will likely remain the world's most important chipmaker. If they can't, the industry landscape could look very different by 2035. Luna: And we'll be watching. For Boardroom Conversations, I'm Luna. Lucas: And I'm Lucas. See you next time.