Why did chip stocks have their worst day since 2020 on November 5, 2026?

On Thursday, November 5, 2026, the stocks most tied to the AI buildout had their worst day in more than six years. The Philadelphia Semiconductor Index, which tracks the largest chip companies listed in the United States, fell about 12%, its steepest one-day drop since the market crash of March 2020. The broader market fell with it: the Nasdaq Composite lost 5.52% and closed more than 10% below its October record, which traders call a correction.

The day before had looked very different. On Wednesday, the first trading day after the November 3 midterm elections, the Dow gained over 1,500 points. By Thursday's close it had given all of that back. This article explains what happened and why; it does not recommend buying or selling anything.

What fell, and by how much

These are closing moves for Thursday, November 5, as reported after the market closed.

Where two outlets give different figures for one company, both are shown. NBC News, whose report was updated late Thursday night, gives 17% for AMD and Micron; a CNBC live update gives 15% for both.

All three major U.S. indexes had their worst day since April 2025, CNBC reports. NBC News estimated that the S&P 500 lost about $2.8 trillion in market value in one day. The damage reached beyond chip designers: memory makers, data-storage companies and "neoclouds", the companies that rent out AI computing, such as CoreWeave and Nebius, also fell sharply.

Why: the election and data centres

The main explanation in every report is political. Democrats won both chambers of Congress and several key governor races on November 3. During the campaign, the cost of electricity and the spread of power-hungry data centres became a major issue, and some winning candidates promised to slow or restrict new projects. CNBC notes that Senator Bernie Sanders, who caucuses with the Democrats, has called for a national moratorium on data centres.

Investors read this as a risk to the plan that props up chip prices: years of very large spending on AI data centres by a handful of companies. If building gets slower, more expensive or more regulated, fewer chips are needed, or they are needed later. "Markets are pricing in the midterm results," eToro analyst Jakub Rochlitz told Fox Business.

The guide's article on why communities oppose AI data centres explains where that resistance comes from, and on election night, voters in many local ballot measures sided against new data centres.

Why chips fell first: Qualcomm and Arm

Politics explains the mood; earnings explain why chip stocks led. On Wednesday evening, Qualcomm and Arm both reported quarterly results that investors saw as mixed, NBC News reports.

• Qualcomm reported revenue of $10.44 billion and adjusted earnings of $2.17 a share, close to forecasts, but its smartphone chip sales fell 26% to $5.14 billion and its outlook for the current quarter came in below expectations, SiliconANGLE reports. The reason: memory has become so expensive, because memory makers are selling as much as they can to AI data centres, that phone makers raised prices and sold fewer phones. Chief executive Cristiano Amon called it a supply problem, not a demand problem.
• Arm beat its headline numbers, but its royalties, the fee it earns on each chip shipped with its designs, came in below what analysts wanted.

This shows a side of the AI buildout that often gets missed. The same demand that lifts Nvidia and memory makers squeezes companies that depend on cheap memory, such as phone chip makers. The guide's article on the memory shortage explains why.

How this connects to the week before

This was not a sudden turn. A week earlier, on October 30, AI stocks fell when Alphabet and Amazon showed spending on data centres outrunning the cash their businesses produce; the guide covered that in Why did AI stocks fall this week when cloud sales were booming? Doubts about whether all this spending will pay off were already in the market. The election gave those doubts a political reason.

Not everyone reads it as the end of the boom. "I don't think this is an AI bubble bursting," Ed Yardeni of Yardeni Research told NBC News, describing it instead as a bubble "losing some air very, very rapidly." The guide's AI bubble debate sets out both sides.

Why do reports of the same day show such different numbers?

On a day this fast, the same stock can be down 5% at one moment and 17% a few hours later. Reports written in the morning describe moves during trading, which can be much larger or smaller than the close; a widely shared morning report on November 5 had AMD down 13%, and an earlier summary posted before U.S. markets opened showed far smaller drops. Even after the close, live blogs updated during the afternoon can show a figure that changed by the final bell, which is likely why CNBC and NBC News differ on AMD and Micron.

When you read market news, check three things: the time the report was written or updated, whether the figure is a close or a move during the day, and which index is meant. The Philadelphia Semiconductor Index and the SOXX fund track similar but not identical lists of companies, so one can fall 12.3% and the other "more than 12%" on the same day. For a reliable number, use closing figures published after 4 p.m. Eastern, and prefer the exchange's own data or a market wrap updated that evening.