Nvidia after Jensen Huang: what developers should watch

Jensen Huang, who co-founded Nvidia and ran it for 33 years, is no longer its chief executive. The board accepted his resignation as president, CEO and director on Wednesday 11 November, effective immediately, and announced it the next morning. Colette Kress, the chief financial officer, is interim CEO. For developers, nothing about CUDA, drivers, chip supply or pricing has changed, and nothing has been announced that would change them. What has changed is who decides, and how closely Washington is now looking at where Nvidia's chips end up.

This post separates what Nvidia has confirmed from what is only reported, then covers the parts that matter if your work runs on Nvidia hardware.

What Nvidia has confirmed

These points come from Nvidia's Form 8-K and the press release filed with it.

Resignation	Accepted by the board on 11 November 2026, "effective immediately". Huang left as president, CEO and board member
Interim CEO	Colette Kress. She stays executive vice president and CFO, a job she has held since September 2013
Board chair	Harvey C. Jones, a director since 1993 and a former CEO of Synopsys. He is the first chair Nvidia's board has had
Successor	The board "will consider internal and external candidates". No timeline, and no search firm named
Reason given	None
Separation terms	None disclosed
Earnings	Tuesday 17 November, "as previously scheduled"

The release contains no quote from Huang. It thanks him for "immeasurable contributions to the company and the industry" and says nothing about any continuing role.

Two sentences in it are aimed at people who depend on the company. Jones said the board expects "the company's operations, customer commitments and product roadmap to continue without interruption." Kress said: "Our focus remains on serving our customers, partners, developers and shareholders."

How it got here

The sequence below uses only Nvidia's own statements.

• Sunday 8 November. The board announced an independent investigation "in light of recent reporting that founder and CEO Jensen Huang has been the target of a foreign intelligence effort." It added that it had "no reason to believe that either national security or NVIDIA secrets were compromised" and that it had confidence in Huang.
• The same day. Nvidia published an email Huang sent to employees. "I was unfaithful to my wife," he wrote. "There is no excuse for what I did." He said he supported the investigation and would cooperate fully.
• Tuesday 10 November. The independent directors said they had retained 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 it.
• Wednesday 11 November. The board accepted his resignation.

Press coverage adds the detail Nvidia's statements leave out. CBS News, Fortune and others report that the relationship was with a woman US officials suspect of spying for China, and that she is linked to a company under investigation over the export of restricted chips. Those claims come from officials and investigators the outlets do not name. CNBC and CBS, each citing people familiar with the matter, report that the board had been preparing to remove Huang before he resigned. Nvidia has not said so. Later reports made further claims about his personal conduct. They have no bearing on anyone's build pipeline and are not repeated here.

The investigation has not reported any findings.

What does not change this week

• The roadmap, according to the board. The statement above is the only official word. Several analysts quoted since then say much the same. "The risk is not that Nvidia suddenly loses its technology roadmap," Michael Feldman of ClearBridge Investments told Reuters. "The risk is execution." Jeremiah Buckley of Janus Henderson told CNBC: "We think the product roadmap is solidified through 2028." These are opinions, not disclosures.
• Supply. No report describes a change in what Nvidia can ship. A Wedbush analyst quoted by Yahoo Finance said "component and material access, not end demand, is defining shipments" through at least next year. That was true before this week and is still the constraint.
• The earnings date. Nvidia reports on 17 November. Its outlook from August was revenue of $108.0 billion, plus or minus 2%, with the note that it was "not assuming any Data Center compute revenue from China." As CFO, Kress already writes the commentary that goes out with the results. This time she presents as chief executive too.

The stock is a different matter, and it is not a measure of the product. CBS News reported that Nvidia closed down 4.3% at $166.19 on Thursday 12 November, after a far steeper fall at the open.

What could change for developers

Four things are worth watching. None of them is a reason to change code today.

1. Export controls

This is the most likely source of real-world effect, and it is where Congress has already acted. According to CNBC's reporting on Thursday:

• Senator Tim Scott, who chairs the Senate Banking Committee, sent letters asking Huang and Commerce Secretary Howard Lutnick to testify publicly on 3 December and 10 December, and raised the possibility of subpoenas.
• Senator Ron Wyden and Representative Richard Neal wrote to Lutnick to announce a probe of the administration's approval of Nvidia chip exports to companies linked to the Chinese government.
• Representative Brian Mast, who chairs the House Foreign Affairs Committee, said Huang had agreed to appear before his committee.

An HSBC analyst quoted by Yahoo Finance said "one possible outcome could be enhanced export restrictions." That is a forecast and no rule has been proposed. If your product ships to, or runs on hardware in, a country covered by US chip export rules, this is the thread to follow. For everyone else, the practical effect of a tighter rule is usually indirect: who can buy which chip, and where capacity gets built.

2. The GPU clouds Nvidia finances

Nvidia does not only sell to cloud providers. It invests in them. TechCrunch's report lists $2 billion each in CoreWeave and Nebius this year and $36 billion in cloud service commitments. Both companies' shares fell sharply on Thursday. TechCrunch put the drops at about 9% and 16%.

A falling share price is not an outage. No report describes a change in capacity, pricing or contracts at any GPU cloud. But if a large share of your training or inference runs on one specialist provider, know what your contract says about reserved capacity, and know where you would run the workload if you had to move it.

3. Nvidia's stakes in the AI labs

TechCrunch also lists $30 billion in OpenAI and $10 billion in Anthropic, and says Huang's exit "raises questions about the future" of commitments like these. That is the publication's judgement. Nvidia has said nothing about changing any investment, and the board's statement about customer commitments points the other way. It is on this list because these were deals Huang made personally, and the next chief executive inherits them.

4. Who gets the job

Reuters reports that investors and analysts see two leading internal candidates: Kress, and Ian Buck, who runs Nvidia's hyperscale and high-performance computing business and is the engineer credited with creating CUDA. A chief executive from the software side and one from finance would probably set different priorities, but that is speculation until the board names someone. Kress's memo to staff, quoted by Reuters, says the company will hold a meeting on 18 November.

Huang remains a large shareholder. Reuters puts his stake at about 3.6%.

Is the CUDA moat weaker?

Some commentary this week says so. Fortune quotes Bret Greenstein of the consultancy West Monroe, who wrote that Anthropic "could move its inference workloads from Nvidia's CUDA software stack to AMD in just three days" and that the "software moat may no longer be as strong as it once was." Daniel Newman of Futurum told CNBC that "AMD is clearly the immediate beneficiary."

Treat the three-day figure as one consultant's claim. Fortune reports no confirmation from Anthropic or AMD, and it is a statement about inference at one company with a very large engineering team. It says little about a codebase with custom kernels, or a training pipeline tuned for one generation of hardware.

The useful question is about your own stack, and you can answer it without waiting for the news:

• Where is the Nvidia dependency? Code written against a framework is one kind of dependency. Code that calls vendor libraries directly, or that was tuned for one card, is a much harder one to move.
• Has it ever run anywhere else? A second backend that has never been tested is not a fallback. Yesterday's post has a table of what a fallback really costs for model providers, and its point carries over: the cheap options only work if you set them up in advance.
• Is portability already there in the tools you use? Some projects are built for many backends from the start. The notes for llama.cpp 1.0, released the day after Nvidia's announcement, list work across CUDA, Metal, Vulkan, SYCL, OpenCL and WebGPU in a single release.

None of this is new advice, and a change of chief executive does not make it urgent. It makes it a good week to find out the answer.

Dates to watch

Date	What
Tuesday 17 November	Nvidia's third-quarter results at about 1:20 p.m. Pacific, call at 2 p.m.
Wednesday 18 November	Company meeting, according to Kress's memo to staff as quoted by Reuters
Thursday 3 December	Date the Senate Banking chair asked Huang to testify
Thursday 10 December	Date the same committee asked Commerce Secretary Lutnick to testify
No date	The board's investigation, and the search for a permanent CEO

The short version

• Confirmed: Huang resigned as president, CEO and director on 11 November. Colette Kress is interim CEO and stays CFO. Harvey Jones is the board's first chair. The filing gives no reason.
• Also confirmed: three days earlier Huang told employees he had been unfaithful to his wife, and the board opened an independent investigation into reports that he was the target of a foreign intelligence effort. It has reported no findings.
• For developers today: no change to CUDA, drivers, supply or pricing has been announced. The board says the roadmap and customer commitments continue.
• Watch: export controls, where Congress has already sent letters, and then the GPU clouds and AI labs that Nvidia finances, and the choice of successor.
• Worth doing anyway: find out how hard your Nvidia dependency really is, and whether your fallback has ever been run.

Frequently asked questions

Who is Nvidia's CEO now?

Colette Kress, Nvidia's chief financial officer since September 2013, is interim CEO and keeps the CFO role. The board accepted Jensen Huang's resignation on 11 November 2026 and says it will consider internal and external candidates for a permanent successor. It has given no timeline.

Why did Jensen Huang resign from Nvidia?

Nvidia's filing gives no reason. Three days earlier Huang told employees he had been unfaithful to his wife, and the board opened an independent investigation into reports that he had been the target of a foreign intelligence effort. The board had said it had no reason to believe secrets were compromised.

Does the CEO change affect CUDA or Nvidia's product roadmap?

Nothing has been announced. The new board chair said the company expects its operations, customer commitments and product roadmap to continue without interruption. No Nvidia document or report read for this post describes a change to CUDA, drivers, chip supply or pricing.

When does Nvidia next report earnings?

Tuesday 17 November 2026, as previously scheduled. Results are due at about 1:20 p.m. Pacific time and the call starts at 2 p.m. Nvidia's outlook from August was revenue of $108.0 billion, plus or minus 2%, with no data centre compute revenue from China assumed.