Latest / Boardroom Conversations with Fexingo: CEOs, Strategy, and Corporate Leadership Explained / How the TSMC Board Keeps the World's Chip Supply Flowing
Transcript
- Lucas: Luna, here's a number that stopped me this morning: TSMC makes over 90% of the world's most advanced chips — the ones in every iPhone, every NVIDIA data center GPU, every AMD processor. Luna: Ninety percent? That's essentially a monopoly on leading-edge logic. Lucas: Exactly. And the question I kept coming back to is: How does a board govern a company with that kind of market power — and that kind of geopolitical exposure? TSMC's headquarters sits ninety miles from the Chinese coast. The company is building factories in Arizona and Japan, and it's spending $30 billion a year on R&D. Luna: That's a board that can't just think about next quarter's margin. They have to think about national security, supply chain resilience, and technology roadmaps that are ten years out. Lucas: Right. And the structure they've built reflects that. TSMC's board has ten members — only one is an executive, the CEO C.C. Wei. The rest are independent directors. That's an unusually high proportion, even by US standards. Luna: Morris Chang, the founder, set that up deliberately, didn't he? He wanted the board to be a strategic check on management, not a rubber stamp. Lucas: He did. Chang stepped down as chairman in 2018, but he engineered a governance model where the board includes former CEOs of major industrial companies, academics, and even a former UK minister. They meet every two months, and each meeting includes deep dives on technology roadmaps, not just financials. Luna: So the board is almost like a technology advisory council. That's unusual for a manufacturing company. Lucas: Completely. Most boards spend the majority of their time on audit, compensation, and compliance. TSMC's board spends at least half its time on process technology and capacity planning. They have to understand the difference between a three-nanometer node and a five-nanometer node, and they have to sign off on which one gets the next $20 billion fab. Luna: That requires a level of technical fluency you don't see on, say, a consumer goods board. Lucas: Exactly. And the board's composition reflects that. Take Mark Liu, the current chairman — he's a PhD in electrical engineering from Berkeley. He spent his entire career at TSMC before moving to the board. The independent directors include Sir Peter Bonfield, former CEO of BT, who sits on the audit committee, but also L.K. Lim, a former executive at Texas Instruments who understands chip design. Luna: So they're not just financial overseers. They're domain experts who can push back on the CEO's technical assumptions. Lucas: That's exactly right. And the board has had to exercise that muscle in some very tense moments. During the chip shortage of 2021, TSMC was flooded with demands from automakers, Apple, and the US government. The board had to decide how to allocate capacity — who gets priority, and at what cost. Luna: I remember reading that they actually prioritized automotive chips even though those are lower margin, partly because of pressure from governments. Lucas: Right. And that decision went through the board, not just management. The board also had to approve the massive capital expenditure increases — TSMC's capex went from $17 billion in 2020 to $36 billion in 2022. That's a board-level bet on demand that could have gone wrong if the global economy slowed. Luna: And it didn't slow. But the risk was real. What about the geographic expansion? The Arizona fab alone is a $40 billion commitment. Lucas: That's another board-level decision that goes beyond normal corporate strategy. When TSMC announced its Arizona fab in 2020, it was a direct response to US government pressure and national security concerns. The board had to weigh the cost — building in the US is at least 50% more expensive than in Taiwan — against the strategic necessity of diversifying production. Luna: And they had to do that while maintaining the company's culture of secrecy and operational excellence. TSMC is famously insular. Lucas: Right. The board's role in preserving that culture is underappreciated. Morris Chang used to say that the board's most important job is picking the next CEO. In 2024, the board oversaw the transition from Mark Liu to a new chairman, and they did it without any drama. That's a sign of a mature governance system. Luna: So the board's real value is not just in approving big spends — it's in maintaining the institutional knowledge and stability that lets TSMC execute decade-long roadmaps. Lucas: Exactly. And that brings us to the biggest challenge ahead. TSMC's next generation of chips — the two-nanometer node — will cost something like $28 billion to develop. The board has to approve that investment, but they also have to think about what happens if geopolitical tensions escalate and Taiwan becomes inaccessible. Luna: That's a risk no amount of governance can fully hedge. But the board can at least make sure the company has options. Lucas: And they are. TSMC is building a second Arizona fab, a fab in Japan, and has even explored sites in Germany. Each of those decisions went through the board. The directors are effectively running a multinational logistics operation disguised as a chip company. Luna: It's striking how different this is from the typical boardroom. Most boards are reactive — they respond to activist investors or regulatory changes. TSMC's board is proactively managing the world's most critical supply chain. Lucas: And they do it with a level of discretion that's rare. TSMC's board meetings are not leaked. There are no proxy fights, no activist campaigns. The company has a single class of stock, so no dual-class structure to protect management. The independence is real. Luna: I guess when you have a monopoly on advanced chips, you can afford to focus on the long term. But I wonder if that governance model is transferable to other industries. Lucas: That's the million-dollar question. The TSMC model works because the board members are deeply technical and the company's mission is clear. For a consumer goods company, you might need different expertise. But the principle — board as strategic partner, not just overseer — is universal. Luna: Speaking of universal principles, one thing that keeps our conversations ad-free is listener support. If these episodes have helped you think differently about corporate strategy, you can support the show at buy me a coffee dot com slash fexingo. Lucas: It's a small way to keep this kind of deep-dive conversation going without any ads. And it really does make a difference. Luna: So back to TSMC — what's the one governance lesson you'd take away from this episode? Lucas: I'd say: build a board that can challenge management on the things that actually matter. For TSMC, that's technology and geopolitics. For your company, it might be something else. But the structure has to match the strategic reality. Luna: And make sure your directors have the expertise to ask the right questions. Otherwise, they're just a rubber stamp. Lucas: Exactly. Next time, we'll look at how the board of a very different company — one in the middle of a turnaround — handled a crisis. Stay tuned.