Latest / Boardroom Conversations with Fexingo: CEOs, Strategy, and Corporate Leadership Explained / How Nvidia's Board Managed the AI Chip Boom
Transcript
- Lucas: So Nvidia reported its fiscal first-quarter results back in May, and the headline number was something like thirty-four billion dollars in data-center revenue alone. Think about that—that's more than the entire company made in a full year just six years ago. Luna: It is staggering. And it's easy to look at that and think, well, Jensen Huang just has the Midas touch. But a company this size doesn't get run by one person. Lucas: Exactly. So today I want to look at the people around the table—the Nvidia board. How did that board oversee a transformation from a gaming graphics-card maker to the infrastructure provider for the entire AI industry? Luna: And this is a particularly interesting governance case because Jensen is both founder and CEO, with a very strong vision. How do you provide oversight without getting in the way? Lucas: Right. So the Nvidia board currently has twelve directors. Jensen is the chairman. But the rest are independent—people like Mark Stevens, a former Sequoia partner, and Harvey Jones, who was an early software executive. They've been on the board for over two decades in some cases. Luna: So they've seen the full arc, from near-bankruptcy in the early 2000s to the crypto boom to AI. Lucas: Exactly. And one of the most consequential decisions the board made happened in 2006. They approved the development of CUDA—Compute Unified Device Architecture. At the time, it was a huge engineering investment with no clear commercial payoff. Nvidia was selling graphics cards to gamers. CUDA let programmers use those GPUs for general-purpose computing. Luna: I remember reading that some investors were skeptical. Why spend millions on a platform when your core business is consumer graphics? Lucas: And the board could have said no. They could have said, 'Stick to gaming, that's what we know.' But instead, they backed Jensen's bet. And they did it with a structure that allowed for long-term thinking—the board doesn't face quarterly pressure in the same way a company with activist shareholders might. Luna: So the board's patience was key. But patience alone isn't governance. How did they actually monitor progress? Lucas: They installed a technology committee. Mark Stevens and a couple of other directors with engineering backgrounds regularly reviewed CUDA adoption metrics. They weren't just nodding along—they dug into the developer ecosystem numbers, the number of PhD papers using CUDA, things like that. Luna: So they created a dashboard that wasn't just financial. Lucas: Exactly. And that discipline paid off when the AI wave hit around 2015. Suddenly, every deep-learning researcher wanted Nvidia GPUs. The board's earlier bet on CUDA created an unassailable moat. Luna: If these conversations have moved your work forward in some small way, we want to keep them ad-free. That's a deliberate choice on our part—no sponsors, no interruptions. It's the kind of show we'd want to listen to. If that resonates with you and you'd like to support that independence, you can at buy me a coffee dot com slash fexingo. No pressure, just a simple way to keep going. Lucas: Yeah, that's really appreciated. And it makes a difference. So back to Nvidia—the board also had to navigate the supply chain crisis of the early 2020s. When demand for AI chips exploded, Nvidia couldn't get enough manufacturing capacity from TSMC. Luna: That was a real test of board-level risk oversight. How do you allocate limited supply when every hyperscaler is desperate for your product? Lucas: The board worked closely with management to prioritize long-term relationships over short-term profit. They could have sold chips to anyone at inflated prices, but instead they allocated capacity to key cloud partners like Microsoft, Amazon, and Google. That strengthened those partnerships for the long haul. Luna: So the board was essentially saying, 'We'll take a hit on immediate revenue to secure our strategic position.' Lucas: Precisely. And that's the kind of decision independent directors can help with. Jensen might have been tempted to maximize quarterly numbers—he's a competitive guy. But the board provided a counterbalance, reminding everyone of the multi-decade vision. Luna: Let's talk about geopolitics. The US export restrictions on advanced chips to China have been a huge issue for Nvidia. How did the board handle that? Lucas: That's a fascinating board-level discussion. Nvidia had to design chips specifically for the Chinese market that met export compliance—like the A800 and H800—only to have those restricted later. The board had to approve significant R&D spending on custom designs for a market that could be shut off at any moment. Luna: So they were essentially building a product line with an expiration date. Lucas: Yeah, and the board had to weigh the risk: do we invest in China-compliance chips and possibly lose that investment, or do we cede the market to Chinese competitors like Huawei? They chose to stay in the game, but with strict oversight from the governance committee on compliance. Luna: That's a board that's deeply engaged with operational reality, not just rubber-stamping. Lucas: Now, one thing I find interesting is the board's composition. Nvidia has a relatively small board for a company of its size—twelve people—but they have deep tech roots. There's no one from the consumer goods world or a retired politician. It's very focused. Luna: Do you think that's a strength? Or does it create groupthink? Lucas: It's a double-edged sword. On one hand, they move fast because everyone understands the technology. But some governance experts argue they lack diversity of perspective—no one with a background in labor, or retail, or geopolitics itself. Luna: Has the board faced any notable proxy fights or shareholder activism? Lucas: Not really. Nvidia has been so successful that shareholders have been quiet. But that's changing as the valuation gets higher. There have been proposals about separating the chairman and CEO roles—Jensen holds both—but so far it hasn't gained traction. Luna: Let's talk about succession. Jensen is sixty-three now. The board must be thinking about life after Jensen. Lucas: They are. And this is a classic founder-led company challenge. Jensen is irreplaceable in a sense, but the board has a succession planning committee that works with him to identify internal candidates. They've been grooming people like Colette Kress, the CFO, and other executives. Luna: But if Jensen left tomorrow, the stock would drop significantly. Lucas: Undoubtedly. But the board's job is to ensure the company can survive that. They've institutionalized the culture around CUDA and the data-center strategy. It's not just Jensen's vision anymore—it's embedded in the organization. Luna: So what's the one lesson other boards can learn from Nvidia? Lucas: I think it's the importance of having a strategic metric that's not financial. The board's early focus on CUDA's ecosystem metrics—developer adoption, academic usage—gave them confidence to stay the course during lean years. That's hard to do when you're staring at a P&L. Luna: It requires patience and conviction. And a board that's willing to bet on a technology before it's proven. Lucas: Exactly. And that's the kind of board we don't always talk about. It's easy to focus on the CEO, but the people around the table made those bets possible. Luna: Yeah. Well, that's a good note to end on. The unsung heroes of the AI revolution.