1A Jensen Huang exit went from tail risk to the favourite outcome in one evening
On Wednesday evening, US time, a wave of reports said Nvidia's board is weighing whether to remove Jensen Huang as chief executive. They followed a report of a second extramarital relationship, on top of the one he confirmed to staff on 8 November. Several outlets attributed the board story to the Wall Street Journal. None of the reports we saw said a decision had been made.
The betting moved faster than the reporting. Polymarket's contract on Huang's resignation or removal being announced by 31 December traded at 76.5% overnight. That is up 61 points in 24 hours, on $3.7m of volume. A day earlier the same contract was 15.5%. The contract for an announcement by Thursday 12 November alone traded at 58.5%, on $1.9m. Polymarket also prices Colette Kress, Nvidia's finance chief, as the most likely successor at 22.4%, up 17 points, though on only $14k of volume.
The stock had not finished reacting. closed Wednesday at $173.85, up 3.0%. Alpaca's last after-hours print at 16:59 ET was $168.27, down 3.2% from the close, and several outlets reported later after-hours losses of up to 8%. A Prof G Markets segment put the cost of the affair story near $1 trillion, because Nvidia fell while other chip stocks rose about 10%. The same show had cited Kalshi odds of his departure at 33% earlier in the day, a level the overnight price has since overtaken.
The sources disagree on what an exit would mean. Michael Batnick and Ben Carlson, on The Compound and Friends, both named Huang one of the largest key-person risks in the market, second only to Elon Musk. Carlson expected months of investigations now that Democrats control both chambers. Batnick, recording on Tuesday before the second report, said flatly that Huang was not going anywhere. A Prof G Markets segment argued the opposite: few chief executives anywhere would have survived the first disclosure. Ross Givens re-posted his case that chief-executive affairs have preceded rallies at Amazon, Microsoft and Google. That call dates from 10 November, three anecdotes, and none involved a board vote.
So what: the question has moved from whether China sales survive to who runs the company. A board decision would settle the second question quickly. It would not settle the first.
2The chip rebound ran on CoreWeave's numbers and left Nvidia behind
CoreWeave's results drove Wednesday's session. rose 24.7% to $60.80, after a pre-market briefing cited 168% revenue growth. A separate report said it plans to raise billions of dollars in debt secured on its GPUs. The read-through ran to other suppliers of data-centre chips and capacity. rose 7.3%, the semiconductor fund rose 4.8%, and rose 15.6%. The Nasdaq-100 fund gained 2.3% and gained 1.3%.
rose only 3.0% on the same news. CoreWeave's fleet runs on Nvidia GPUs, so the lag is idiosyncratic rather than a demand signal. The company-specific discount from claim 1 absorbed part of a sector-wide bid.
The rebound follows a violent week. Batnick and Carlson recapped last Thursday, 5 November. The Nasdaq-100 fell about 6%, its worst day since March 2020, and semiconductors fell 11%. The memory fund was down 54% from its high at the lows. Carlson cited at three times earnings. Batnick's view is that the market is right to price memory at a cyclical peak. Carlson's view is that low multiples assume earnings fall, and if they hold, the stocks are mispriced. That disagreement stayed open.
Two explanations for the swing compete. Givens tied the selloff to the midterm results, in particular a new Texas governor who campaigned on a data-centre moratorium. He said 20 to 28% of US data-centre construction is in Texas. Carlson's read of the ballot measures was milder: local opposition raises the cost of building, which shrinks returns but favours the largest spenders, who can pay. He also cited JPMorgan data showing the biggest hedge-fund degrossing since 2016, which would make the swing about forced selling more than about politics.
So what: one strong earnings report was enough to restart the chip trade. Nvidia now trades as a governance story inside a sector that is trading on demand.
3Money leaving AI stocks is staying in the market and moving into older industries
Jim Paulsen, on Excess Returns, made the case that the economy has split in two. Business spending on information processing and software is about 14% of real GDP and grew about 15% a year. The other 86% grew 0.7% over the past year, which he called recession-like. Payrolls are up only about 0.5% from a year ago, slower than 86% of non-recession periods since the war, by his count. His conclusion is that the newer sectors could fall 20 to 30% while the rest of the market holds up, leaving the S&P 500 with only a mild correction. He said he would lean further toward the older sectors.
The data from last week fits his frame. Batnick cited Fidelity's Jurrien Timmer: during the drop, AI stocks were 30% off their highs and the S&P 500 6% off, but the equal-weight index was down only 1.3% and the S&P 500 excluding AI was up 0.5%. Batnick's summary was that sellers wanted out of certain stocks, not out of stocks. Givens saw the same rotation from the chart side. Medical devices, pharmaceuticals and biotechnology now lead his 40-sector strength ranking, while a neocloud fund he follows is down 58% since August.
Paulsen added a warning about rates. For most of the year, falling Treasury yields helped stocks, because the worry was inflation. That correlation has flipped. Falling yields now arrive with falling stocks, which he reads as growth fears entering prices. The ten-year yield printed 4.88% on 9 November, from 4.86%, and the gap between ten-year and two-year yields narrowed to 0.25 points on 10 November, from 0.32.
Felix Prehn, on his own channel, offered the opposite worry. He argued money supply growth of 5.7% this year means asset prices broadly keep rising and savers lose ground. His framing is inflationary, where Paulsen's is a slowdown. Both cannot be the main risk in the next quarter, and neither showed data that settles it.
So what: the sell-off so far has rotated money rather than removed it. Paulsen's thesis needs that to continue through a rebound in tech, which has not been tested yet.
4A December hike is priced as settled, and the half-point tail held after Waller
Kalshi's December Fed market prices a quarter-point hike at 86%, up from 84%, with a hold at 5%. A half-point hike is 11%, from 12%, on $111k of 24-hour volume. Fed Governor Christopher Waller was reported to lean toward a quarter-point move while staying open to half a point. The market took that as confirmation of the base case, not as a push toward the larger move. January is a coin flip: a second hike is 51%, unchanged, and a hold is 46%, down from 50%, on thin volume of about $10k.
The two forces still pull apart. October consumer prices rose 0.35% in a month, the print that started this repricing. Fed Governor Michelle Bowman was reported to attribute about a quarter of excess inflation to AI-related demand. Paulsen calls hiking into a supply-driven inflation a "double mistake", because higher rates neither cool the AI build-out nor reopen shipping lanes. The market is siding with the Fed for December. Paulsen's view would need weaker growth data to win, and the next jobs report is weeks away.
Oil eased the pressure at the margin. Reuters reported crude down more than $1 overnight, on easing geopolitical fears and oversupply. It also reported OPEC+ output in October near its pre-war rate, led by Saudi Arabia. Against that, Saudi Arabia intercepted Houthi missiles near Mecca, and Reuters reported Iranian attacks on shipping navigation systems off Oman.
So what: nothing in the overnight data argues for a December pause. The live question is January.
What the sources recommended
Only two single-name buys came out of five episodes, both from one trader, and one triggers only on a breakout that had not happened by Wednesday's close. Ross Givens said he bought 1,000 shares of on a pullback to its 50-day average and a retest of its $29 breakout, with a stop at $28.40, or 2.9% of risk. He framed Pfizer as the "boring" half of a bet on healthcare leadership. His second name, , is conditional: he set an alert for a break above a "high tight flag" near $54.50, with a stop at the flag's low near $51. closed Wednesday at $53.88, below that trigger. Both calls came in a video that also sold a 99-cent membership, and neither came with a fundamental case.
Givens also re-posted his 10 November purchase, 200 shares worth about $35,000. That is a re-cut of an earlier call, not a new one. No source called a sell.
The next test is Nvidia's board, not its chips
Nvidia's board is the catalyst with the most money riding on it. Polymarket gives an announcement by the end of Thursday 12 November a 58.5% chance, which makes Thursday itself a test of the overnight repricing. An announcement would settle claim 1; silence through Friday would leave the 76.5% year-end price looking early.
Earnings will test claim 2. reports before Thursday's open, and and report after the close. Cisco and Applied Materials both depend on the data-centre orders that CoreWeave's results just validated. A weak outlook from either would undercut Wednesday's rebound. A strong one would sharpen the gap between demand across the sector and Nvidia's governance discount.














