- On 12 November Jensen Huang left his roles as Nvidia CEO and board member, effective immediately.
- CFO Colette Kress is interim CEO, and Nvidia will still report quarterly results on 17 November as planned.
Nvidia’s board announced on 12 November that Jensen Huang, who founded the company in 1993 and ran it ever since, has resigned as CEO and as a board member, effective immediately. CFO Colette Kress takes over for now. This matters to anyone who pays for AI, because almost every large model is trained and served on Nvidia chips. The short version: nothing changes in your prices overnight, but there are clear things worth watching.
What Nvidia actually announced
The statement is short and factual. Colette Kress has been Nvidia’s CFO since September 2013 and before that spent almost 25 years in finance roles at companies including Cisco and Microsoft. As interim CEO she also keeps her finance job until the board picks a permanent successor from inside or outside the company. The second change is at board level: for the first time in its history Nvidia has a separate chair, Harvey C. Jones, a former CEO and executive chairman of Synopsys.
Jones says in the statement that operations, customer commitments and the product roadmap will continue without interruption. Nvidia also says it will report results for the third quarter of fiscal 2027 as planned on Tuesday, 17 November.
Why it happened so fast
Nvidia itself gives no reason. Two days earlier, the independent board members said they had hired Joon H. Kim, a former acting US Attorney for the Southern District of New York, and the law firm Wachtell, Lipton, Rosen & Katz to run an independent investigation. According to TechCrunch, the resignation came days after Huang admitted to an affair with a woman whom US authorities suspect of spying for China. That is media reporting, not part of the company’s statement.
TechCrunch also reports that Nvidia shares fell almost 12% in early trading on Thursday and as much as 19% since the previous Friday, wiping more than $855 billion off its market value. The Decoder, counting from the first reports on 29 October, puts the loss at about $1.4 trillion.
The maths for anyone paying for AI
Your bill for a language model does not depend on who runs Nvidia. It depends on what your provider pays for chips and electricity. The Decoder adds a revealing detail here: Nvidia runs a profit margin of around 75%, which is already pushing its biggest customers (Microsoft, Amazon, Alphabet) to build cheaper chips of their own. A change at the top could speed that shift up or slow it down, but not within a week.
What I would do now as a builder or business owner:
- check whether my product can switch to another provider’s model if prices or limits suddenly change;
- watch the 17 November results and what Nvidia says about supply and orders;
- avoid panic purchasing decisions: according to the company, contracts and deliveries continue as before.
I write more about how to budget for AI in the AI in business hub.
What could change for open models
Huang was the loudest champion of open models, the kind anyone can download and run on their own hardware. According to TechCrunch, Nvidia agreed to buy Hugging Face, the largest home for such models, for $12.9 billion, with the deal expected to close in the first half of next year if regulators approve. Huang also rallied more than 100 companies behind the Open Agent Safety Platform. Nobody knows yet whether the new leadership will push these projects as hard.
What I would keep an eye on
In my view, the key question is not who replaces Huang but whether Nvidia can hold together the web of deals he built personally. The Decoder writes about roughly $300 billion in guarantees for smaller customers and more than 60 people who reported directly to him. A board statement about continuity is a promise, not proof. The real test is the next set of results and the first decisions Colette Kress makes as interim CEO.