sixty direct reports

Jensen Huang ran Nvidia for thirty-three years with about sixty people reporting to him and nobody in line behind him. On Thursday he resigned, and the board had to invent two jobs in a day. Max Weber described this exact moment a century ago.

Sixty direct reports. No chief operating officer. No board chair. No succession plan.

That was how Nvidia was run until Thursday, when Jensen Huang resigned after thirty-three years as its only chief executive. The board made Colette Kress, the CFO since 2013, interim CEO while she stays CFO. It made Harvey Jones, a director since 1993, the company’s first-ever board chair. Two offices that hadn’t existed on Wednesday existed by Thursday morning.

what happened

The week went fast. On Sunday Huang told staff he had been “unfaithful to my wife,” and Nvidia said it was investigating after the New York Times reported he had been the target of a foreign intelligence effort. On Wednesday the Times reported a second relationship, with an executive who reported to him. The board met that night. By Thursday morning he was out.

The board’s statement said it expects “operations, customer commitments and product roadmap to continue without interruption.” The market wasn’t sure. CNBC reported the stock opened down 13 percent and clawed back to close down 4.3, its sixth straight losing week. KeyBanc’s John Vinh called the departure “a material increase in strategic, organizational, and execution risk.” Wharton’s Michael Useem put it harder, as Fortune reported: “If the board had been doing its job, it would already have a slate of possible successors.”

he did pass it on

None of this was an accident. Huang built Nvidia flat on purpose, and he explained why many times.

He doesn’t do one-on-ones. “I don’t have one-on-ones with them because it’s impossible,” he told Lex Fridman in March. “We present a problem, and all of us attack it.” Everyone hears the same thing at the same time. Nobody gets power from knowing something the others don’t.

He didn’t believe in succession planning either. In a 2009 talk at Stanford he called a hand-picked list of successors a toxic process. On the same podcast this year he gave the alternative: “The most important thing you should do today, if you care about the future of your company, post you, is to pass on knowledge, information, insight, skills, experience as often and continuously as you can.”

And he did. For three decades Huang reasoned out loud in front of sixty senior people. Nobody at Nvidia forgot how to build a GPU on Thursday. The roadmap is written down, and so are the customer commitments. By his own standard, the transfer worked.

The stock still dropped, and it didn’t drop over what Nvidia knows. It dropped over who decides.

authority doesn’t sit in on meetings

Here’s what I think Huang got wrong. Knowledge transfers in a room. Authority doesn’t. You can watch someone make a thousand calls and come out knowing exactly how they think. That doesn’t make you the person whose call it is.

Max Weber worked this out in Economy and Society, published in 1922. He described three kinds of authority. One rests on tradition, one on rules and offices, and one on a person: the leader people follow because of who they are. He called the third charismatic. That last kind is the strongest and the least stable, because it can’t be inherited. It lives in one person. When that person goes, the group has two choices. It can turn the leader’s authority into something that outlasts them, which means offices, procedures and a line of succession. Or it can come apart. Weber called the first path routinization, and he said the pressure to do it peaks at succession, when the question can’t be avoided anymore.

Nvidia spent thirty-three years refusing to routinize. No layers, so there were no offices to hand down. No COO, so there was no obvious second. No chair, so the board had no independent head. No list of successors, because Huang thought lists were toxic. The flat structure worked because one person sat on top of it. Sixty people watched every decision get made, and none of them was ever the one making it.

So on Thursday the board routinized in a day, under pressure, the way Weber said it happens. It made a chair and an interim CEO, and it promised a search. My guess at the next step: whoever gets the job permanently won’t have sixty direct reports. Nvidia will grow a layer or two and look more like an ordinary company. That’s the price of having been run by one person for thirty-three years, and the bill came all at once.

the Jensen on your team

You probably have one, at a smaller scale. A lead who reasons in the open and teaches everyone by doing the work in front of them. People love working for them, and they mean it. They don’t name a second, because naming one feels like playing favorites.

Being in that room is a real education. You learn how a good engineer thinks. What you don’t get from it is standing. When the lead leaves, whether it’s burnout, a better offer or a bad week like Huang’s, the knowledge stays with you and the authority leaves with them. Then the team finds out that nobody was ever allowed to make the call, and the first few decisions after that are slow, political and bad.

If you’re the hub, handing over context isn’t enough. Hand over decisions. Pick something you’d have called differently, give it to someone else, and live with their call. It feels like a loss of quality, and in the short run it is. It’s also the only way anyone besides you learns to carry the weight. And making yourself replaceable is usually what gets you promoted, not what gets you cut.

If you’re one of the spokes, look at where your standing actually comes from. If it comes from being close to the person everyone listens to, it’s borrowed, and it goes when they go. Ask for something in writing: a named area you own, decisions that are yours. That way your authority belongs to a role and not to a relationship.

Huang said for years that he was passing it on, and he was. Sixty people watched him decide. What the market wanted to know on Thursday was which of them could.


Sources: TechCrunch: resignation, Kress interim CEO and CFO, Jones first board chair ↗ · CNBC live updates: timeline, board statement, stock moves, Vinh ↗ · Fortune: NYT reporting, Useem, sixty reports, no COO or succession plan ↗ · Al Jazeera: Kress background ↗ · Entrepreneur: Huang on one-on-ones, Lex Fridman Podcast ↗ · Wccftech: Huang on passing on knowledge, Lex Fridman Podcast ↗ · Stanford eCorner: Huang on succession planning (2009) ↗ · Stanford Encyclopedia of Philosophy: Max Weber, charismatic authority ↗