Jensen Huang has left Nvidia. What happened, and what it means for AI
Jensen Huang resigned as Nvidia's CEO on 12 November 2026. What the filing says, the record quarter that followed, and the new US rule on AI chip access.
Jensen Huang resigned as Nvidia's president and chief executive, and from its board, on 12 November 2026, with immediate effect. Nvidia's filing names its chief financial officer, Colette Kress, as interim chief executive. Harvey C. Jones, a former head of Synopsys, became the company's first board chair. The filing gives no reason. Five days later Nvidia reported record revenue of $115.8 billion for the quarter. In the same week the US government issued a rule limiting remote access to advanced chips from China. For teams that build with AI, nothing changes in the next month. The longer questions are about who supplies computing power, and under what rules. This post is not investment advice.
Why did Jensen Huang resign?
Nvidia has not said. The filing records the resignation and the interim arrangements, and nothing more.
It followed two weeks of press reports. On 29 October the New York Times reported that US intelligence had been monitoring an executive at Megaspeed, a Singapore company that buys Nvidia chips, over ties to China. According to the San Francisco Standard, Nvidia's board opened an independent investigation on 8 November. Megaspeed denies wrongdoing, and nobody has been charged.
Huang co-founded Nvidia in 1993 and had led it ever since. On the 17 November earnings call, the new chair said the board did not intend to reinstate him, CNBC reported.
Who runs Nvidia now?
Colette Kress, as interim chief executive. She remains chief financial officer. The board says it is considering internal and external candidates for the permanent job.
On 17 November the board also set up a national security and export control oversight committee, chaired by the director Suzanne Nora Johnson. Its job is to oversee how Nvidia complies with export rules.
How did Nvidia's shares react?
They fell hard, then steadied. Nvidia closed at $173.73 on 11 November and opened at $151.00 the next morning. By TechCrunch's count, the shares were down as much as 19% from 7 November at one point, more than $855 billion of market value. They rose more than 3% after the earnings report and then fell 2.89% the next day.
What did Nvidia report on 17 November?
A quarter larger than any it had reported before, and a bigger forecast. For the three months to 25 October, Nvidia reported:
- Revenue of $115.8 billion, up 103% on a year earlier.
- Data centre revenue of $107.6 billion, up 110%.
- A gross margin of 74.8%.
- Net income of $65.3 billion.
Analysts surveyed by LSEG had expected about $109.1 billion. Nvidia forecast $130 billion for the current quarter, which has an extra week in it. The forecast assumes no data centre computing revenue from China at all.
What is the new US rule on remote access to AI chips?
It closes the route by which Chinese companies rent American chips instead of buying them. The Commerce Department's Bureau of Industry and Security announced the rule on 16 November. It is due in the Federal Register on 19 November, takes effect on publication, and is then open for comment for 60 days.
According to the rule's text:
- Overseas data centres need a licence to provide large amounts of computing power, above a set threshold, to companies headquartered in China, Macau or other countries in the same export group. Providers already doing so get 90 days.
- Some uses are exempt, including the recommendation systems that online shops and social media run.
- Advanced chips need a licence to go to Indonesia, Malaysia, Singapore, Thailand or Brazil, unless the buyer is a US-headquartered company that files reports.
- It formally withdraws the AI diffusion rule from the Biden administration.
Chip stocks fell on the news, the Philadelphia semiconductor index by as much as 5.5%, according to The Next Web.
What is Congress doing?
Asking questions in public. House committee chairs have asked Huang for a closed-door interview on 1 December. The chair of the Senate Banking Committee, Tim Scott, has asked him to testify on 3 December, ABC News reported. Commerce Secretary Howard Lutnick is expected to appear voluntarily. The House already passed a Remote Access Security Act in January 2026, and the Senate's version is still pending.
How much are the big cloud companies spending on AI?
More than ever, and they say next year will be higher. These are each company's own guidance for 2026 capital spending, given with their results in October. They define spending differently, so do not add them up as if they matched.
| Company | 2026 guidance | Source |
|---|---|---|
| Amazon | about $220 billion in cash | Results release |
| Alphabet | $195 billion to $205 billion | CNBC |
| Meta | $135 billion to $145 billion, including leases | Results release |
Microsoft spent $45.2 billion in its July to September quarter alone and expects more than $50 billion in the next. CNBC reports that the industry's spending is expected to pass $1 trillion in 2027. Most of that money buys Nvidia's products.
What does this mean for teams that build with AI?
In the short run, very little. Model prices are set by the AI labs, and the November releases from Anthropic and OpenAI cut prices at the top by about a fifth. We compare them in which AI model is best for coding agents.
Over a longer period, two things are worth watching:
- Supply and policy now move together. In June a single export directive took two of Anthropic's models offline for every customer for three weeks. The remote access rule shows how quickly the rules on computing power can change.
- Cost is per task, and it is yours to watch. List prices can fall while a bill rises, if agents spend more tokens to finish the same work. A spend cap on each agent run keeps one bad run from becoming one bad month.
Cognibl's cloud agents draw prepaid credit at a published rate for each model, and you choose the model on each agent. Every turn of a run records the model its tokens were spent on, so you can see what each piece of work cost. See how cloud agents are priced and run.
Keep reading
- What is OpenAI worth, and when will it go public?OpenAI raised money at a $1.25 trillion valuation in November 2026 and says it will list in 2027. Its owners, revenue, and how it compares with Anthropic.
- Should you run coding agents on open-weight models?Open-weight models cost less per token than Claude or GPT, but not always per task. What their licences allow, and what tests found on agent hijacking.
- The best open-weight AI models in November 2026Kimi K3.1 Pro, DeepSeek V4.1 and GLM-5.3 lead open-weight AI in November 2026. How far they trail Claude, GPT and Gemini on independent tests.