Nasdaq Worst Day Since 2025: Why Chip Stocks Crashed

The AI trade had its worst session in more than a year and a half on Thursday, November 5, 2026. The Nasdaq Composite fell about 5.5%, its steepest one-day drop since April 2025, and closed in correction territory. The PHLX Semiconductor Index lost more than 12%, and the Dow Jones Industrial Average shed more than 1,800 points, erasing all of Wednesday’s post-election rally.

Several forces hit the market at once. Investors weighed the results of Tuesday’s midterm elections, in which Democrats swept key races and voters rejected data center projects in local ballot measures across the country. Earnings guidance from Qualcomm and Arm also fell short of expectations. CNBC attributed the selloff to concerns on Wall Street that the artificial intelligence buildout is poised to slow down after the midterms.

The scale of the selloff

The numbers were large across every major index:

• Nasdaq Composite: down about 5.5%. Fox Business put the decline at 5.52%.
• Nasdaq 100: down 6.36%, its worst day since 2020, according to investingLive.
• S&P 500: down roughly 4.3% to 4.4%, its worst day since April 2025.
• Dow Jones Industrial Average: down 1,833.06 points, or 3.44%, to 51,523.06. A day earlier, the Dow had rallied 1,569 points after the election.

CNN reported that the selloff wiped about $2.5 trillion from the stock market. CNBC calculated that the “Magnificent Seven” megacaps alone lost more than $1 trillion in combined market value.

Volatility spiked. The Cboe Volatility Index (VIX) jumped 11.57 points, its biggest point gain since April 4, 2025, according to CNBC. CNN said the move was a gain of roughly 70%, the largest one-day percentage jump in almost two years, and left the gauge at its highest level in seven months. Risk assets outside equities also fell, with Bitcoin dropping more than 5%, according to CoinDesk.

Chip stocks take the hardest hit

Semiconductors led the market lower. Benzinga reported that the PHLX Semiconductor Index fell 12.3%, its worst day since March 16, 2020, and said 28 of the index’s 30 components are now in bear-market territory. Yahoo Finance data showed the index closing at 9,011.14, down 1,257.52 points.

The biggest declines came from memory makers, chip designers and equipment suppliers:

• Micron Technology: down 17.46%
• AMD: down 17.26%
• Lam Research: down 16.4%
• Applied Materials: down 13.6%
• Broadcom: down 10.4%, closing at $291.53, according to StockTitan
• Qualcomm: down more than 9%
• Nvidia: down 9.3%

CNBC reported that Nvidia lost more than $440 billion in market value during the session, the largest single-company decline of the day, and that the drop erased the stock’s gains for the year. Elsewhere in the megacap group, Tesla fell about 11% and SpaceX about 9%.

AI infrastructure names also fell. CNBC reported that CoreWeave lost almost 10% and Oracle about 7%. The debt-funded neocloud operators had already been under pressure in October, when CoreWeave’s bonds sold off during a global chip rout (see our coverage of the Kospi crash and CoreWeave bond selloff).

Qualcomm’s guidance disappoints

Qualcomm reported fiscal fourth-quarter results after the close on Wednesday, November 4. The quarter itself came in at the high end of the company’s guidance, according to the earnings call transcript: non-GAAP revenue of about $10.4 billion and earnings of $2.17 per share. Some coverage cited total revenue of $10.32 billion.

The outlook was weaker. MarketBeat reported that Qualcomm’s guidance midpoint of $2.25 per share for the current quarter came in just below the consensus of about $2.27, though revenue guidance was slightly above expectations. Tech Insider reported that Qualcomm expects an operating margin of 20% to 22% for its chip business, down from 24% in the fourth quarter, and tied the decline to rising supply costs.

The handset business, still Qualcomm’s largest, was a sore spot. Tech Insider reported that handset revenue fell 26% from a year earlier to $5.14 billion. Analysts cited weak handset demand, Apple’s shift to in-house modems, and the cost of ramping Qualcomm’s data center business, which the company laid out at its June investor day (see Qualcomm’s Dragonfly data center bet). Automotive revenue grew about 90% year over year, according to coverage of the call.

The shares fell between 5% and 8% in extended trading Wednesday, depending on the snapshot, and extended the decline to more than 9% during Thursday’s selloff.

Arm’s royalty miss

Arm Holdings also reported after Wednesday’s close. The company posted record fiscal second-quarter revenue of $1.44 billion, up 27% year over year. Licensing and other revenue rose 50% to a record $773 million, beating the LSEG consensus estimate of $625.1 million. Chief financial officer Jason Child said licensing came in about $100 million above Arm’s own expectations as customers accelerated long-term commitments.

Royalties, which track actual chip shipments, were weaker. Royalty revenue rose 8% to $670 million, below the LSEG estimate of $720.8 million, although data center royalties doubled from a year earlier.

For the fiscal third quarter, Arm guided revenue to $1.54 billion, plus or minus $50 million. The company said royalty revenue would grow by low single digits from a year earlier, as strength in cloud and physical AI is offset by weaker smartphone demand and a tough comparison with last year. Arm also said it is more confident it can reach up to $2 billion in revenue from its AGI CPU across fiscal 2027 and 2028, with first commercial deliveries still targeted for the fiscal fourth quarter.

Arm shares fell between 2% and 5% after hours, depending on the source, before the broader selloff hit Thursday. Investing.com reported that Asian chip stocks slid on Friday as the two companies’ results weighed on the sector.

The midterm factor

The market rallied on Wednesday after the election results, then reversed on Thursday as investors focused on what the results mean for AI infrastructure.

Democrats won control of Congress and several governor’s races. In Texas and Georgia, Democrats who campaigned partly on pausing data center development won offices the party had not held in decades. Goldman Sachs estimated that Democratic statehouse wins in Texas, Michigan and Georgia could put roughly 45% of planned U.S. data center capacity at risk. Goldman also said a Democratic Congress “might limit data center tax incentives, but a national moratorium looks unlikely in any election outcome.”

At the local level, voters in 10 states backed restrictions or rejected pro-development proposals in 40 of 41 data center ballot measures, according to tallies cited by CNBC and Ballotpedia.

Analysts said they doubt any new law could meaningfully stop data center construction. They expect more congressional hearings and investigations into the AI buildout instead. Evercore ISI warned of more negative sentiment toward AI infrastructure stocks in a more restrictive development climate. One strategist quoted by Investing.com said the market narrative had shifted to the view that data centers would no longer be built in the U.S. That view reflects market sentiment, not any announced policy.

Valuations were already stretched going into the vote. The Nasdaq 100 had entered a correction in late October as investors grew concerned that data center spending was growing more slowly than expected.

What it means

Thursday’s drop reflects a change in what the market is worried about. For most of 2026, the main concerns about the AI trade were financial: whether hyperscaler capital spending would ever pay off, how much debt the neoclouds were carrying, and whether interest rates would rise. The midterms added a physical constraint. Data centers need land, power, water and local permits, and voters have now shown they will use the ballot to block them.

Who is most exposed. The stocks that fell hardest are the ones most dependent on new data center construction. Equipment makers such as Lam Research and Applied Materials, memory suppliers such as Micron, and GPU vendors such as AMD and Nvidia all earn money as new capacity gets built. If construction slows, their order books are at risk first. The debt-funded neoclouds are exposed for a related reason: they need new sites to come online on schedule to service their borrowing.

Who could hold up. Some hyperscalers held up better than the chip sector as investors bet that slower construction could rein in Big Tech’s AI spending and make existing facilities more valuable. Companies that already own large amounts of powered capacity have something the market now values more because it is harder to replace.

The Qualcomm and Arm signal. Both companies point to the same split. Data center and automotive revenue are growing, but smartphones, still the largest source of royalty and chip revenue for both, are soft. Neither company’s data center business is yet large enough to offset weakness in phones. That kind of guidance would have drawn little attention in a calm market, but it hit harder on a day investors were already selling.

What to watch next:

• Nvidia’s fiscal third-quarter report later this month, the most direct read on whether hyperscaler orders are holding up.
• Hyperscaler commentary on 2027 capital spending, particularly any change to U.S. site plans in Texas, Georgia and Michigan.
• State-level action by the new Democratic governors, including any moratoriums on new large-load grid connections similar to New York’s hyperscale pause.
• The VIX and credit spreads. A sustained rise in neocloud bond yields would show that the stock selloff is spreading into the debt that finances the buildout.

A one-day move of this size does not by itself mean the AI cycle has peaked. It does show that the market now has to price political risk into AI infrastructure stocks alongside the usual financial risks.