Nvidia's stock closed 4.27% lower on Monday, Nov. 9 after the board opened an investigation into Jensen Huang's relationship with a businesswoman whom intelligence agencies call a Chinese spy. Huang co-founded Nvidia in 1993, has run it ever since, and has no named successor. Bernstein analysts flagged "key man" risk by name.

What exactly counts as key-person risk?

Key-person risk is the exposure a business carries when its value or operations depend on one person or a few people. The usual test is substitutability: how long it would take, and how much it would cost, to replace what that person does.

The idea has a long paper trail. IRS Revenue Ruling 59-60 says losing the manager of a "one-man" business may depress the stock's value, especially when no trained successor exists. It sets no standard percentage for that discount. Public US companies disclose material risks under Item 105 of Regulation S-K. That rule doesn't specifically mention key personnel, but reliance on a founder can be disclosed as a material risk.

That explainer is mostly about small private firms. Nvidia shows the same logic at a far larger scale.

Why would a trillion-dollar company depend on one person?

According to Bernstein, Huang sets long-term strategy and has roughly 60 direct reports, and it's unclear who would step in if he were unwilling or unable to lead. Sixty people reporting straight to the top suggests a lot runs through one desk.

Analyst Patrick Moorhead described Huang's roles as product strategist, the closer with the biggest cloud customers, and the company's face in Washington and Beijing. Moorhead said this story puts the last of those roles at risk. Wedbush's Matt Bryson said removing Huang would be a significant negative because of his hands-on style.

Is a 4% drop normal for a CEO scandal?

Roughly, yes. Yahoo Finance cites a 2018 Journal of Financial Economics study that found stocks fell about 4% on average when the CEO was the one caught in misconduct. At Nvidia's size, Yahoo says, that comes to about $185 billion. That's more than McDonald's was worth when it fired Steve Easterbrook in 2019. (We're relying on Yahoo's summary here and haven't checked the study's sample or methods ourselves.)

What's unusual is the size of the company. Nvidia was worth about $4.5 trillion, nearly 7% of the S&P 500. Yahoo calculates that this is about four times the combined value of the nine largest companies ever hit by CEO personal-conduct scandals. The previous record was Intel, worth $249 billion when Brian Krzanich resigned in 2018. No earlier scandal-hit company was more than about 1% of the index. So a percentage drop that's ordinary for the stock means far more dollars, and more effect on index funds, than in any past case.

Do founder-CEOs really carry a premium?

The best evidence says founders matter more than other CEOs on average, but the effect varies a lot and earlier studies disagreed.

The clearest test comes from sudden deaths, which no market can see coming. Economists Dirk Jenter, Egor Matveyev and Lukas Roth identified 458 CEO deaths at public US firms from 1980 to 2012. Here's what happened to the stocks over the following three days, measured against what the market would otherwise have predicted:

  • All sudden deaths: -2.32% on average.
  • Founder-CEOs: -3.25%.
  • Founders in the youngest third by age: -8.82%.
  • Founders in the oldest third: +5.26%.

The authors conclude that founders matter more to shareholder value, likely because they have more control over their companies.

Losing a CEO doesn't always hurt shareholders. In the same study, sudden deaths of CEOs over 65 were followed by average gains of 3.59%. Slow deaths were followed by gains of roughly 2.6% to 3.6% over two months. The authors read this as a sign that some CEOs are poor matches for their firms or are overpaid.

The "founder premium" is also contested. Earlier studies used small samples and found average reactions close to zero. One 1985 study of 33 CEOs found stocks rose when founder-CEOs died. So the market reaction depends on the person. It isn't a fixed price.

Is an investigation the same as losing the CEO?

No. The death studies measure a CEO who is definitely gone. An investigation is about what might happen next. The CEO might be removed, might stay but spend his time defending himself, or might come out with his influence reduced.

Bernstein's note keeps these risks separate. It called the episode a "heavy distraction" at best and "significant ammunition" for lawmakers at worst. It said demand for Nvidia's products seems unlikely to be affected, but regulatory risk could rise. China, including Hong Kong, made up 8% of Nvidia's revenue in fiscal Q2 2027.

The facts also differ in how solid they are. Huang told staff in a Sunday email that he had an affair. Bernstein said the spy allegations look more tenuous for now because they rest on anonymous sources. The board said it had "no reason to believe" that national security or Nvidia secrets were compromised.

Washington is involved too. Three Republican House committee chairs asked Huang to attend a closed-door briefing on Nov. 18. Bernstein described that as a request, not a subpoena. Sen. Chris Coons said Congress would investigate.

How much of the drop is really about Huang?

Nobody knows yet. The fall could reflect fear of losing Huang, or it could reflect political and regulatory worries, such as congressional scrutiny and China export policy. No source separates the two.

Even the price path is murky. Investing.com's reports describe the $186.64 figure differently. One treats it as the latest close, down 7.6% on the week. The other calls it the Nov. 6 close, followed by after-hours trading at $178.70. Neither lines up neatly with CNBC's Monday close of -4.27%.

What happens next?

It's not known how long the board's investigation will take, what it will find, or whether Huang keeps his job. Analysts at Moor Insights and Wedbush think removal is unlikely unless the investigation finds wrongdoing. Nvidia has no public successor, and it's not known whether the board has a confidential emergency plan.

Bernstein argued that "the primary beneficiary" of Huang's removal would be China. It expects the cloud over the stock to persist, but it didn't change its rating. It still rates Nvidia outperform with a $400 price target.