Did the AI Bubble Just Pop? The 5 November 2026 Rout Explained

Key Takeaways

• On Thursday 5 November 2026, the Nasdaq Composite fell 5.52% to 24,315.00, its biggest daily drop in 19 months, while the S&P 500 fell 4.36% and the Dow fell 3.44%.
• The Philadelphia Semiconductor Index fell 12.25% to 9,011.14, its second-biggest fall on record and its worst since March 2020. SanDisk fell 22.26%, Micron 17.46% and AMD 17.26% in a single session.
• The reported causes were threefold: fear that the Democratic midterm sweep would slow the data-centre buildout and tighten AI regulation, Qualcomm’s weak smartphone-chip outlook, and doubts about returns on massive AI infrastructure spending.
• The value wiped out depends on the scope being measured: Reuters reported about $2 trillion off the Nasdaq’s market capitalisation, while other outlets put the figure at about $2.3 trillion for the Nasdaq 100 and about $2.5 trillion for the Nasdaq Composite.
• Search interest in the question surged: “did the ai bubble just pop” rose 1,400% over 7 days, and “is the ai bubble popping?” rose 850%, according to the trend data behind this story.
• On Friday 6 November morning, Nasdaq futures rose and chip stocks stabilised pre-market, with a Jefferies strategist calling the selloff “reactionary, not a fundamental reset”, according to Reuters. That is a stabilisation, not a verdict.
• One day does not prove a bubble has burst. It proves that investors repriced risk sharply, all at once, and that the AI trade is now sensitive to politics, chip demand and the return on infrastructure spending at the same time.

Thursday 5 November: the day the Nasdaq fell 5.52%

Thursday 5 November 2026 was the worst day for US technology shares in a long time, and the numbers leave little room for doubt about the scale. The Nasdaq Composite closed down 5.52% at 24,315.00, a fall of 1,421.52 points. That was the Nasdaq’s biggest daily drop in 19 months, since April 2025. The Nasdaq 100 fell 6.36%, its worst day since March 2020. The S&P 500 fell 4.36% to 7,353.73, and the Dow Jones Industrial Average fell 3.44% to 51,523.06, a drop of 1,833.06 points.

The damage was broad, not confined to one company or one corner of the market. All 11 S&P 500 sectors fell. Technology fell 7.3%. The VIX, the market’s widely watched volatility index, recorded its biggest jump since December 2024. Bitcoin fell 6% to $75,000. When every sector falls, volatility spikes and even assets outside equities drop on the same day, the market is not reacting to a single earnings miss. It is repricing a whole trade at once.

The Nasdaq is now more than 10% below its record close of 6 October, in correction territory, roughly 11.9 to 12% below that record. That matters because the language changes at that point. A fall of a few percent is a pullback. A fall of more than 10% from a recent high is, by the market’s own convention, a correction. It does not mean the market will keep falling, and it does not mean a bubble has burst. It means the gains of the autumn have been substantially given back, and quickly.

The chip stocks led the fall

If the Nasdaq’s fall was the headline, the semiconductor selloff was the story inside it. The Philadelphia Semiconductor Index fell 12.25% to 9,011.14 on 5 November, its second-biggest fall on record and its worst since March 2020. That is an extraordinary single-day move for an index of large, established chip companies.

The individual closes tell the same story. SanDisk fell 22.26%. Micron fell 17.46%. AMD fell 17.26% to $404.40. Marvell fell around 17%. Arm fell 13.08%. TSMC fell 10.78%. Broadcom fell 10.43%. Intel fell 10.33%. Nvidia fell 9.26% to $181.60, a fall that erased its 2026 gains. Qualcomm fell 9.5%. Tesla, which trades as part of the AI story as much as the car story, fell 10.77%.

The large technology companies fell too, but by less. Meta fell 3.96%. Microsoft fell 3.58%. Alphabet fell 2.20%. Amazon fell 2.19%. Apple fell 1.53%. The pattern is worth noting: the companies that make the chips, and the memory that goes into AI systems, fell roughly twice as far as the companies that buy the chips and run the services. Investors were not selling AI demand in general. They were selling the hardware at the centre of the AI buildout, and selling it hardest.

That pattern is also why this selloff felt different from the wobbles earlier in the autumn. In the earlier falls, a single company’s results or a single funding story moved the market. On 5 November, the selling hit the whole supply chain at once, from memory makers to chip designers to the foundry that manufactures many of their chips.

The week that set it up: AMD, the midterms and Qualcomm

The 5 November rout did not come out of nowhere. The three days before it each added a piece of the story, and the pieces fitted together badly.

Tuesday 3 November brought two very different kinds of news. AMD reported third-quarter revenue of $13.38 billion, up 45% year on year, with earnings per share of $2.09 and data-centre revenue of $8.55 billion, up 97% year on year. Its shares rose about 7.5% after hours. On its own, that was a strong result, and it showed that demand for AI data-centre chips was still growing fast. The same day, the US midterm elections delivered a Democratic sweep, with Democrats winning and taking control of Congress, the House and the Senate according to reports. That result would matter enormously two days later, for reasons that had little to do with AMD’s numbers.

Wednesday 4 November looked, on the surface, like a good day for the market. The Dow rose 3.0% to 53,356.12. The S&P 500 rose 1.6% to 7,688.62. The Nasdaq rose 1.4% to 25,736.51. But underneath the headline gains, the Philadelphia Semiconductor Index fell 3.1%, and AMD fell 4.6%, giving back much of its post-earnings rise. Qualcomm reported fiscal fourth-quarter revenue of $10.322 billion, down 8.4% year on year, with handset-chip revenue down 26% to $5.14 billion, missing estimates, and gave a weak outlook for smartphone chips. The same day, Representative Yassamin Ansari called for a pause on advanced AI development until adequate regulations are in place. The market’s broad gains on 4 November hid a split that would become the whole story on 5 November: the wider market was rising, the chip trade was already falling, and the political debate about AI was sharpening.

Why investors sold: three reported causes

Reporting from Reuters, CNN, Business Times and Le Monde pointed to three linked causes for the 5 November selloff, and it is worth keeping them separate, because they imply different things about what happens next.

The first cause was politics. The Democratic midterm sweep raised fears that Democrats would slow the data-centre buildout and tighten AI regulation. That fear was given specific numbers by the banks. UBS estimated that about 30% of expected 2028 US data-centre additions could face delays. Goldman Sachs said Democratic state and local wins could temporarily halt projects affecting about 45% of planned US data-centre capacity expansion. Whether those estimates prove right is not yet known, and they are estimates, not outcomes. But they explain why the selling hit data-centre and chip stocks hardest: if a large share of future data-centre capacity is delayed, the order books of the companies supplying those data centres are the first thing investors question.

The second cause was evidence that chip demand is cyclical after all. Qualcomm’s weak outlook and Intel’s warning on PCs and memory suggested that parts of the chip market still follow the old boom-and-bust cycle, even while AI data-centre demand is strong. That matters because the AI trade has been priced as though chip demand would grow almost without interruption. If memory and PC chips are cyclical, investors have to ask which other parts of the AI supply chain are cyclical too, and what happens to prices and margins when supply catches up with demand. The memory stocks’ falls on 5 November, SanDisk down 22.26% and Micron down 17.46%, were the sharpest expression of that doubt.

The third cause was the oldest question in the AI debate: whether the massive spending on AI infrastructure will generate sufficient returns. Companies have committed enormous sums to data centres, chips and power, and investors are asking, with growing impatience, when that spending turns into profit at a matching scale. This doubt was already present before November. What the midterm result and the Qualcomm numbers did was give it a trigger, a day on which all three doubts could be acted on at once.

Michael Burry’s AI-bubble warning also resurfaced in coverage of the selloff. Burry, the investor made famous by “The Big Short”, had taken bearish put positions involving Nvidia, Micron and Palantir, as reported in September. His warning did not cause the selloff, and this article does not claim that it did. But its return to the headlines shows how the mood had changed: arguments that had been dismissed as premature a few weeks earlier were suddenly being quoted as prescient.

One paragraph of background: the late-October wobble

The 5 November selloff was not the first AI wobble of the autumn, but it was a much bigger one than the events that preceded it. In late October, Alphabet fell about 12 to 13% in a day, losing roughly $500 billion in value after its third-quarter free cash flow came in at minus $9.07 billion on AI capital expenditure of $51.36 billion; OpenAI cut its funding-round target from $30 billion to about $20 billion and its target valuation from $1.4 trillion to $1.25 trillion; Intel fell about 28 to 29% in a day, its worst since the 1970s, after weak fourth-quarter guidance; and the Nasdaq 100 shed over $2.5 trillion in market value across October. Our earlier coverage, AI Bubble Fears Surge as Markets Reprice the AI Boom, covered those events in detail. They matter here only as context: the market was already nervous about AI spending and returns before November began, and the midterm result gave that nervousness a new, political shape.

How much value was wiped out?

The value wiped out on 5 November was reported differently by different outlets, and the differences are about scope, not accuracy. Reuters reported about $2 trillion off the Nasdaq’s market capitalisation. CNBC TV18 put the figure at $2.3 trillion for the Nasdaq 100. FirstPost, the Straits Times and CNN put it at about $2.5 trillion for the Nasdaq Composite. These figures are not contradictory. They measure different baskets of stocks. The honest way to report them is the way this article does: name the scope alongside the number, and do not present any single figure as “the” loss.

Whichever scope is used, the loss was enormous, and it came in a single session. That is what made 5 November feel like a rupture rather than a bad day. A $2 trillion-plus fall in one day is the kind of event that forces every investor in the AI trade to re-examine their assumptions, even if they believe the underlying technology is sound. It also forces a distinction that matters for readers: a fall in market value is not the same as a fall in the value of the technology. Share prices can fall because expectations were too high, while the products themselves keep improving. Both things can be true at the same time, and on 5 November, both probably were.

What the search data shows

The search data behind this story shows that the public question changed on 5 November. Over 7 days, search interest in “did the ai bubble just pop” rose 1,400%, “is the ai bubble popping?” rose 850%, “ai bubble crashing” rose 400%, “sandisk stock” rose 850%, “intel stock” rose 650%, “amd stock” rose 450% and “nasdaq composite” rose 450%. These are observed rises in search interest over 7 days, not measures of how many people searched, and they should be read that way.

What they show is that people were not only watching the market fall, they were asking whether the fall meant something structural. The question “did the ai bubble just pop?” is a different question from “why is the Nasdaq down?”. The first asks for a verdict on an era. The second asks for an explanation of a day. The surge in the first kind of question tells us that the selloff was being interpreted, in real time, as a possible turning point rather than a routine dip. Search interest does not settle whether that interpretation is right. It settles that the interpretation is widespread, and that the bubble question is now the frame through which many people are watching the market.

Friday morning: stabilisation, not a verdict

The morning after a rout is always informative, and Friday 6 November brought a measure of calm. Nasdaq futures rose, and chip stocks stabilised pre-market, with AMD, Intel, Broadcom and Micron rising 5 to 8.3%. Nasdaq 100 e-minis were up 1.48% at 04:14 ET. A Jefferies strategist called Thursday’s selloff “reactionary, not a fundamental reset”, according to Reuters. That is a meaningful view from a professional strategist, but it is a view, not a fact, and it should be read as one.

Asia told a different part of the story on Friday. South Korea’s Kospi plunged almost 9% at the open, triggering a 20-minute trading halt. Tokyo’s Nikkei slumped more than 6%. The Asian falls were a delayed reaction to the US selloff, and they show that the shock travelled across markets rather than staying in New York. A pre-market bounce in US futures and a sharp fall in Asian markets can both be true on the same morning. Together they describe a market that is unsettled rather than settled, and that is the honest description of 6 November so far.

Did the AI bubble just pop?

This is the question the article is named for, and it deserves a careful answer rather than a dramatic one. A bubble, in the strict sense, is a period in which prices rise far beyond what the underlying assets can justify, and then fall sharply when that gap becomes impossible to ignore. By that definition, 5 November looks like evidence for the bubble case: prices fell sharply, across the whole AI supply chain, on doubts about regulation, demand and returns, and the Nasdaq entered correction territory.

But one day, however dramatic, does not prove a bubble has burst. Three cautions apply. First, the causes reported for the selloff are partly political and partly cyclical, and both can change. A delayed data centre is not a cancelled data centre, and a weak quarter for smartphone chips does not prove that AI data-centre demand has peaked. Second, the companies at the centre of the trade are still reporting strong growth in AI-related revenue, as AMD’s third-quarter results showed just two days before the rout, with data-centre revenue up 97% year on year. A bubble bursting would normally be accompanied by evidence that the underlying business has stopped growing. That evidence has not arrived. Third, markets have produced sharp one-day falls before without ending a technology cycle. The Nasdaq’s worst day since March 2020 is a serious event, but March 2020 itself was followed by a recovery, not by the end of the technology era.

The stronger reading of 5 November is that investors repriced risk, sharply and all at once, because three doubts arrived on the same day. The political doubt is about whether the data-centre buildout will proceed at the pace investors assumed. The cyclical doubt is about whether chip demand grows smoothly or in booms and busts. The returns doubt is about whether the spending turns into profit. None of those doubts is new. What is new is that they are now being priced at the same time, by the same market, on the same day. For context on how the AI buildout is being financed, and how much debt is involved, see Oracle and Alphabet Borrowed Billions for AI on the Same Day. And for the funding pressures that were already visible before November, see OpenAI Cuts Its Funding Target by $10 Billion as AI Stocks Slide.

What to watch next, without guessing the outcome

If you want to follow this story honestly, watch a short list of verifiable developments rather than daily commentary. First, whether the pre-market stabilisation on 6 November holds through the trading day, or whether selling resumes. Second, what Democrats actually say and do about data-centre permitting and AI regulation in the coming weeks, rather than what investors fear they might do. The gap between feared policy and actual policy is where much of this selloff will be judged. Third, the next round of results from the chip companies, and specifically whether memory prices and data-centre orders hold up or soften. Fourth, whether AI-related revenue at the large technology companies keeps growing fast enough to justify the capital spending behind it. Fifth, whether the value wiped out on 5 November is treated by investors as a buying opportunity or as the start of a longer de-rating. That last question will be answered by flows and prices over weeks, not by any single analyst’s call.

For businesses that buy AI tools rather than invest in AI shares, the practical question is different. It is not whether a bubble exists in the abstract, but whether your vendor will still be there, at a price you can afford, with a product that does useful work. That question is answered by contracts, service levels and measured results, not by headlines. A correction, if that is what this turns out to be, may actually help buyers, by forcing vendors to compete harder on price and proof.

The bottom line

On 5 November 2026, the Nasdaq fell 5.52%, the Philadelphia Semiconductor Index fell 12.25%, and roughly $2 trillion to $2.5 trillion in market value was wiped out, depending on the scope measured. The reported causes were a Democratic midterm sweep that raised fears of slower data-centre building and tighter AI regulation, weak chip demand signals from Qualcomm and Intel, and persistent doubts about the returns on AI infrastructure spending. The next morning brought stabilisation in US futures and sharp falls in Asian markets, and a Jefferies strategist’s view that the selloff was reactionary rather than fundamental. None of that, on its own, proves that the AI bubble has burst. It proves that the AI trade is now exposed to politics, cyclicality and the returns question all at once, and that investors are no longer willing to price it as though none of those risks exist. The bubble question will be settled by policy, orders, revenue and profits over the next several quarters, not by a single day’s headlines, and not by a surge in search interest. Until then, separate what happened from what is feared, attribute every figure to the outlet that reported it, and be sceptical of anyone who claims to know the ending already.

Sources: Reuters, Global Markets Trading Day graphic, 5 November 2026; CNN, Tech-heavy Nasdaq plunges, 5 November 2026; Business Times, US stocks: Nasdaq posts biggest drop in 19 months; Le Monde, On Wall Street, Democrats’ midterm victory bursts the AI bubble; Reuters, Nasdaq futures rise after steep AI selloff, 6 November 2026.