Jensen Huang's apology: what the Nvidia crisis teaches us about saying sorry

The short version: why is this news?

If you have seen Jensen Huang's name everywhere this week and are not sure why it matters, here is the quick version.

  • Who he is. Huang co-founded Nvidia in 1993 and ran it for 33 years. Nvidia makes the chips that power most of the AI boom and is one of the most valuable companies in the world, valued by the Financial Times at around $3.7 trillion. Huang became the public face of AI.

  • What happened. On 8 November he admitted in an email to staff that he had been unfaithful to his wife. The woman involved was reported to have alleged links to Chinese intelligence. Within days, further reports emerged of other affairs, including two with Nvidia employees. He resigned on 12 November.

  • The twist. Employees, investors and President Trump publicly backed his return. On 17 November, Huang said he would come back if the board asked. Hours later, the board said it would not.

  • Why anyone cares. It is a national security story, because Nvidia's chips sit at the centre of US-China tensions over AI and Huang had been pushing to loosen export restrictions. The BBC reports that China now fears the scandal could shut it out of US AI technology ahead of a summit between Xi Jinping and President Trump. It is a markets story, because Nvidia shares had just had their worst week in 19 months. And it is a governance story, because a board has refused to bring back a founder whom his staff, his investors and the President all wanted back.

For anyone who works in crisis communications, it is also a near-perfect case study. Almost every stage of it offers a lesson.

Lesson one: the apology answered the wrong question

On paper, Huang's first email did what crisis advisers recommend. It was short, it named what he had done and it made no excuses. Plenty of commentators praised it.

But it apologised for infidelity, to his family and colleagues. The questions that would decide his future were different: had he compromised the company's security, and had he broken its rules on relationships with staff? Those questions belonged to the board, regulators and Washington, and the apology did not speak to them.

A good apology is not just sincere. It is aimed at the issue that actually threatens you, and at the people who actually decide what happens next. Apologising sincerely for the wrong thing can make you look as though you have not grasped how serious the situation is.

Lesson two: your first statement becomes the benchmark

Whatever you say first is what everything after it gets measured against. When further affairs were reported, the original email stopped reading as a full account and started reading as a partial one, even if that was never the intention.

This is one of the hardest conversations I have with clients. If there is more to come, it is almost always better that it comes out at once, on your terms, than in instalments through other people's reporting. Each new revelation resets the story to day one, and the second time round, people are far less willing to give you the benefit of the doubt.

Lesson three: a board's credibility is part of the crisis too

According to reports, the board's first public position was that it had launched an independent investigation and remained fully confident in its CEO. Days later, it accepted his resignation.

That reversal cost the board something. Its own credibility became part of the story. Having changed course once, under scrutiny from investors and Washington, it could not afford to look as though it might change course again. A second U-turn would have suggested it either did not know what it was doing or could be pushed around.

That is why early statements of confidence are so risky. They feel stabilising at the time, but they spend credibility you may need later. The more durable line is usually the plainer one: an investigation is underway, we will act on what it finds, and we will not comment further until then.

Lesson four: popularity is not the same as permission

The Stand With Jensen campaign gathered support from employees, prominent investors and the President. On most measures of public opinion, Huang was winning.

But public support does not decide the outcome when someone else holds the power. A board under national security scrutiny could not be seen to bow to political and commercial pressure. The louder the campaign became, the more reinstating him would have looked like exactly that. In a sense, the campaign made the no more certain.

The PR industry appears to agree. PR Week has reported that communications professionals are advising against Nvidia bringing him back.

When I assess a crisis, one of the first questions I ask is who actually decides. Not who has the most sympathy or the loudest supporters. In a listed company, it is the board. For an individual, it might be an employer, a regulator, a publisher or a court. The communication strategy should be built around them.

Lesson five: don't negotiate in public

Huang's offer to return was posted on X on the morning of Nvidia's quarterly earnings call. That forced the board to answer on its biggest stage of the quarter, in front of investors and the press.

A public offer can feel like taking control, but it turns a private negotiation into a public test. The other side can only accept and look weak, or refuse and look strong. Once someone has said no in public, saying yes later becomes almost impossible without losing face. If there is any route back, it nearly always runs through quiet conversations first.

Lesson six: say it once, then change the subject

Look at how the board handled that call. The chair addressed the issue once, at the very start, in a single clear line. He refused questions. Then the company moved on to reporting record revenue of $115.8 billion, up 103% on the year before.

That is containment. It will not satisfy everyone, but it stopped the call from becoming another chapter of the scandal and put a different headline beside it. Saying less, clearly, with your strongest facts alongside, is often the most effective thing an organisation can do in the middle of a crisis.

Lesson seven: the story spreads to the people around you

A week in, the coverage has moved well beyond Huang himself. The Irish Times has run a column about how his wife is being judged. The Wall Street Journal has published a piece about his leather jackets, the outfit that had become part of his public image.

This happens in every big crisis. Once a story is large enough, it spreads to partners, children, colleagues and anything recognisable about the person: their clothes, their catchphrases, their old interviews. The things that made someone familiar and likeable become material for commentary, and the people closest to them, who did nothing wrong, become part of the story.

That is why crisis planning has to include the people around the person at the centre. Who is likely to be approached by the press? Who needs advice on what to say, or not to say? Who needs protecting? It is one of the first things I look at, and one of the things most often forgotten.

Lesson eight: the permanent record is already being written

Within a week, Wikipedia had a dedicated article on what it calls the Jensen Huang spying scandal, and his own page now describes him as Nvidia's CEO from 1993 to 2026. Pages like these feed Google's Knowledge Panels and the AI tools people increasingly use to ask who someone is.

The news cycle will move on. The labels attached to a name in search results and on Wikipedia tend to stay, and they are very hard to change afterwards. How a crisis is named, summarised and ranked in the first few weeks can shape how a person is described for years. Making sure that record is accurate and proportionate, within each platform's rules, is the part of crisis PR most people forget.

Search his name from the UK and Google adds a line at the bottom of the results noting that some results may have been removed under European data protection law. It is a small reminder that individuals here have legal routes to ask for outdated or irrelevant results about them to be delisted, one of several tools for making sure a single moment does not define someone indefinitely.

What this means if you're not running one of the world's biggest companies

Most of my clients are not in Huang's position. But the same dynamics show up at every scale: a founder whose personal life becomes a business problem, a director whose first statement is overtaken by new information, a well-meaning supporter whose public defence makes things worse.

The principles hold:

  • Apologise for the thing that actually matters, to the people who actually decide.

  • Get the full facts out at once rather than in instalments.

  • Be careful about public expressions of confidence, in yourself or anyone else.

  • Remember that public support does not override the people with the power.

  • Keep negotiations private for as long as you can.

  • Plan for the people around you, not just the person at the centre.

  • Say it once, clearly, and put your strongest facts beside it.

  • Pay attention to what search results and Wikipedia say about you after the headlines fade.

If you are facing something similar and want a confidential second opinion, get in touch.

Next
Next

Crisis PR FAQs: what to do when your reputation is under pressure