Will Jensen Huang leave as Nvidia CEO?
GitHub · suggest a change · how to contribute
Base case
Before this month, the chance of Huang leaving in a given year was low single digits. He co-founded Nvidia, has run it for more than three decades, and said in January he will stay "for as long as I deserve it".5 The model adjusts from there.
The affair, and the allegations
On Nov 8, 2026 Huang told Nvidia staff he had been unfaithful to his wife.2 The woman he had the affair with is reported to be an executive of Megaspeed, an Nvidia customer, whom U.S. intelligence officials reportedly consider an asset of China's Ministry of State Security.3 Nvidia's board opened an independent investigation.1
The affair needs to be separated from the allegations that she is a Chinese spy and an executive of a customer. The affair, admitted at once, is unlikely to cost him the job on its own. The allegations could.
The critical question
If the dealings were above board, he has a very high probability of staying. If there was a conflict of interest, a preference given, or information passed, keeping his job would be a long shot. So the critical question is the chance that something improper happened and comes to light. If nothing is found, nothing comes to light.
The forecast
Exits of this kind come fast or not at all. Most of the risk falls in the first months. If he is still CEO after that, the chance falls back toward the base case.
Drag across the chart to pick a date. The open circles are the market's dates.
Inputs
Each input is my estimate. Change one and everything above updates.
Low and high set the shaded range on the chart. Low means a lower chance that he has left, so for the two timing inputs low is the slower case.
Why these numbers
- Chance that something improper happened and comes to light
- My gut feeling is that an affair between a chief executive and an executive of one of his company's customers more likely than not leads to something improper coming to light. It has to be found and made public, not only to have happened, which is why this is not higher.
- Chance he leaves, if something improper comes to light
- Keeping his job after that would be a long shot. Given the stakes of the AI race with China it should be higher than the base rate for chief executives with findings against them, but 95% felt too high. This one is uncertain, and the low and high values are placeholders.
- Chance he leaves over the scandal, if nothing more comes to light
- If the dealings were above board, he has a very high probability of staying. I put this below the base rate, given his commitment, founder status, long tenure and relative youth.
- Median days until the decisive finding or report
- 85 days is a starting point: the median in my count of 11 past board investigations of a sitting chief executive. The list is not yet published. It feels long to me: the board would want to move rapidly, and the press will. Against that, there is no known successor, which is a reason to take time. The low and high values are placeholders.
- Days from the decisive finding or report until he has actually left
- Placeholders, not yet estimated.
- Chance per year he leaves for a reason unrelated to the scandal
- Low single digits. He is a co-founder with more than three decades in the job who says he will stay "for as long as I deserve it".5 This covers mortality and the tail risk of extremely poor performance.
Assumptions
- An exit caused by the scandal follows the same clock whether or not something improper comes to light.
- Half of the decisive findings come by the median, three quarters by twice the median, and so on.
- Unrelated exits happen at a constant yearly rate, independent of the scandal.
Weak points
- The critical question is a gut estimate. No count of past cases supports it yet.
- There is almost no precedent. A chief executive tied to an alleged agent of an adversary state has few earlier cases to compare with.
- The board has already said it "has no reason to believe that either national security or NVIDIA secrets were compromised".1
- The President's backing. He has said Huang should stay.3 If the tide turns against Huang, I expect him to go with the tide.
- Only an actual departure counts. An announced future exit does not, and the contract terms do not address a leave of absence or an interim chief executive.4
What would change the forecast
- The board committee naming its members, its counsel or a timeline, or reporting a finding.
- Nvidia starting to put a succession plan in place.
- Huang's answer to the House chairs' request for a closed briefing on Nov 18.6
- Public hearings, a subpoena, a review of Nvidia's export licences, or charges in the Megaspeed case.
- New reporting on what he knew, or on what passed.
Open questions
- How long did the relationship last, and what access did it give her?
- Did the board know of it before Nov 8?
- Did Megaspeed receive anything other customers did not?
- What is the FBI examining?
- How long from a decisive finding to him actually being out? The values in the model are placeholders.
- Would a leave of absence count as leaving?
Something missing? Suggest an improvement
Suggest a change on GitHub, or copy the repository and send your version (how to contribute). The same calculation as a short script: forecast_curve.py.
Sources
- NVIDIA. "Statement from NVIDIA Board of Directors." 2026-11-08.
- CNBC. "Read Huang's email to his staff and statement from Nvidia's board of directors." 2026-11-08.
- NBC News. "Nvidia opens investigation after CEO Jensen Huang's affair with alleged Chinese spy." 2026-11-09.
- Kalshi. CEOCHANGE contract terms.
- Quartz. Report of Huang's remarks to reporters at CES. 2026-01-07.
- House Select Committee on the CCP. "Moolenaar, Mast and Crawford seek briefing from Jensen Huang." 2026-11-05.
Last updated 2026-11-09 · history
Nothing here is financial advice or a recommendation to trade.