On 12 November 2026 Nvidia’s board announced that Jensen Huang had resigned with immediate effect as president, chief executive and director, ending a 33-year tenure at the company he co-founded. The board named chief financial officer Colette Kress interim CEO and said it would search for a permanent replacement. That is the whole of what is officially established.
The price action was violent and then restrained. NVDA fell as much as 14.35% shortly after the open — briefly erasing roughly $600 billion of market value — then pared the loss to close down 4.34%, about $182 billion. The Nasdaq Composite closed roughly 0.4% lower and the S&P 500 about 0.35% lower at 7,588.9.
Most coverage treated the gap between those two numbers as the story. The more useful number is 27%.
The part that was already priced
Through Wednesday 11 November, NVDA was already 27% below its 6 October record. The resignation did not create that drawdown. It explained it.
This is what key-person risk looks like in practice. It is not a disclosure item and it is not a line in a model. It sits inside the discount rate as an unquantified premium that rises quietly as a founder ages, as succession stays unresolved, and as the company’s valuation comes to depend more on one person’s judgement than on its balance sheet. Analysts rarely name it because there is no input field for it.
So the correct reading of 12 November is not “a $600 billion shock out of nowhere”. It is a five-week repricing that finally got a label.
The index math hides the damage
Nvidia is the largest single weight in the Nasdaq-100 and among the largest in the S&P 500. When a stock carrying the heaviest weight falls 4.3% and the index falls 0.4%, the arithmetic only works if the other names went up. They did.
AMD and Broadcom each rose about 6%. So the index was not resilient — it was rotated. That distinction matters to anyone holding exposure through a fund, a pension or a CFD: you own the winners and the losers at index weight, you captured the rotation on neither side, and the only thing you actually received was the volatility.
Two names in the same complex went the other way and are the more informative signal. CoreWeave fell 9% to 11%, and Nebius dropped 17% after a soft 2026 revenue guide. That is not a vote on Nvidia’s management. It is a vote on the second derivative of AI capital spending — and guidance, not governance, is what repriced the neoclouds.
Three scenarios, three levels
The market will resolve this on succession, not on sentiment.
A permanent external CEO is the highest-variance outcome: it introduces strategy risk into a company whose entire moat is an architectural roadmap executed on a fixed cadence. An internal promotion, or Kress confirmed outright, is the continuity outcome and the closest thing to a non-event. A strategic reset — reshoring decisions, a capex pause, a change in the platform model — is the tail that would matter most and is the least likely to be announced early.
On US100, that maps to three round numbers. A close back above 25,500 says the market has decided the succession is manageable. Holding the 5 November low says it is waiting. A break of 25,000 would turn a governance shock into a trend, because at that point the index has lost the level that defined the November bounce.
The honest leak in the AI bull case
The macro backdrop did not cause this, but it decides how far it travels. The 2-year Treasury yield is 4.73%, the dollar index sits at a 19-month high, and the December FOMC — 8-9 December, with the funds rate already at 3.75%-4.00% — is priced around 87% for a 25bp hike.
AI equities are the longest-duration assets listed anywhere. Long duration is precisely what breaks when the discount rate rises. A governance shock lands hardest on the asset class that had already been relying on multiple expansion rather than cash flow, which is the uncomfortable symmetry between this story and the correction that began on 5 November.
Expressing it rather than predicting it
UZFX lists seven index CFDs — US100, US500, US30, DE40, JP225, HK50 and AU200 — so the transmission can be traded without a single-stock position, with a 0.01-lot minimum and a $10 minimum deposit on the Web Terminal, H5 and the iOS, Android, Windows and Mac apps. There is no MT4 or MT5. Negative balance protection applies.
The cleanest structural trade for a concentrated book is a breadth trade: short US100 against long US500. US500 carries lower single-name concentration, so the ratio isolates the key-person effect from the market effect. Be clear about what it is not — both indices share the same top holdings, so this reduces exposure, it does not hedge it.
Leverage is the part that decides outcomes, and it is set by the entity on your client agreement, not by the platform. ASIC-regulated retail accounts cap major indices at 20:1; offshore entities advertise up to 1:500. The same position therefore needs 25 times more margin on the ASIC account. Overnight financing on index CFDs also rises with the hiking cycle, so a multi-week short is not a free position.
The sizing arithmetic that actually matters
Take a US100 position with 3% of notional posted as margin. That is roughly 20:1. A 3% adverse gap — entirely normal for an index that just printed a 14% single-stock intraday move — consumes 60% of the margin posted.
Now run the same gap at 1:500, where margin is 0.2% of notional. A 3% gap is 15 times the margin posted. That is not a drawdown, it is a liquidation, and the platform closes you at the worst price of the session rather than the one you would have chosen.
At 0.01 lots the numbers are small in dollars and identical in percentages. That is the whole argument for starting there: the percentage is what teaches you, and the small dollar amount is what lets you survive the lesson.
Calendar
US PPI prints on Friday 13 November — the first macro test of whether the rate story is still tightening. The December FOMC is 8-9 December. Nvidia’s next earnings date has not been confirmed since the board change, and until it is, the succession announcement is the only scheduled catalyst.
FAQ
Q: Did Jensen Huang resign or was he removed?
The board announced on 12 November that he resigned with immediate effect as president, CEO and director after 33 years, and named CFO Colette Kress interim CEO. No reason has been confirmed on the record. Anything beyond the board statement is unconfirmed reporting and should be treated that way rather than repeated as fact.
Q: Is this the top for AI stocks?
Nobody knows. What is documented is narrower and more useful: NVDA had already fallen 27% from its 6 October record before the announcement, so the repricing began five weeks earlier without anyone naming key-person risk as the cause.
Q: What is key-person risk?
The share of a company’s valuation that depends on one individual staying in place. It is rarely disclosed and usually invisible, because it hides inside the discount rate as an unquantified premium that rises as succession stays unresolved.
Q: Why did the Nasdaq fall only 0.4% when Nvidia fell 4.3%?
Because the money leaving NVDA went into AMD and Broadcom, which each rose about 6%. The index was held up by rotation, not resilience. A passive holder or an index CFD holder owns both sides at index weight and captures neither move.
Q: Does more leverage protect a small account?
No. At 20:1 a 3% adverse gap costs 60% of the margin posted. At 1:500 the same gap is 15 times the margin posted, which means liquidation at the worst price of the session. Leverage changes survival odds far more than it changes expected profit.
Q: Can this be traded on a 10-dollar account?
Yes, at 0.01 lots on US100 or US500. The purpose is not the size of the profit. It is that a 3% gap does not end the account, which is the only property that keeps a small account in the game long enough to be right once.