The 5 November AI Rout: The Nasdaq’s Worst Day Since Liberation Day

8 min read

On Thursday 5 November 2026 the Nasdaq Composite fell 5.52% to 24,315.00, its worst day since the “Liberation Day” tariff shock of April 2025, and closed in a technical correction. The following week it recovered part of the loss while the 10-year Treasury yield climbed back above 5%. This article looks at what the day showed in breadth and sector data, why the rebound week does not settle anything, and which parts of market history are worth comparing. It is history and data, not a forecast.

Index closes for 5 November are from Reuters via Investing.com unless stated. Sector, breadth and volatility figures are from CNBC’s market statistics for the day. Treasury yields are official US Treasury par yields. Weekly figures for 9–13 November are from Portfolio Terminal’s weekly review. Data runs to the close on Friday 13 November 2026.

What the tape did on 5 November

All three major US indexes had their steepest one-day fall since 4 April 2025.

Index 5 November 2026 Close 4 April 2025, for comparison
Nasdaq Composite −5.52% 24,315.00 −5.82%
S&P 500 −4.36% 7,353.73 −5.97%
Dow Jones Industrial Average −3.44% (about 1,833 points) 51,523.06 more than 2,200 points, or 5.5%

5 November closes: Reuters via Investing.com. 4 April 2025 comparisons: CNBC. Other wraps printed slightly different closes for the S&P 500 and the Dow, a tick or a few points apart; we use Reuters throughout.

The damage was concentrated in the AI trade, but it was not confined to it.

  • Semiconductors. The Philadelphia Semiconductor Index fell more than 12%, according to Reuters. Other reports put the fall at 12.3%, its worst day since March 2020.
  • Technology. The S&P 500 information technology sector fell 7.30%, its worst day since 16 March 2020, CNBC reported. The Technology Select Sector SPDR (XLK) fell 8.04% on more than three times its 30-day average volume.
  • The Magnificent Seven lost more than $1 trillion in combined market value, “with Nvidia leading the way at a loss of more than $440 billion”, according to CNBC. Nvidia fell more than 9%.
  • Everything else. All 11 S&P 500 sectors fell. Industrials dropped 4.12%, their worst day since April 2025. The equal-weighted S&P 500 lost 2.9%, TheStreet reported, a day after closing at a record.

Estimates of the value lost differ, and they cover different universes. Reuters said about $3 trillion of S&P 500 market value was erased. MarketWatch estimated $2.9 trillion. Separately, about $2.3 trillion was erased from the Nasdaq-100. These figures overlap and cannot be added together.

Breadth

Breadth measures how many stocks rose against how many fell. It tells you whether a move was carried by a few heavyweights or by the whole market. On 5 November it was the whole market.

  • NYSE: 2,344 decliners to 416 advancers, almost 6 to 1 (CNBC).
  • Nasdaq: 4,166 decliners to 855 advancers, almost 5 to 1 (CNBC).
  • S&P 500: 411 decliners to 75 advancers (investingLive).
  • Volume on US exchanges was 26.9 billion shares, against a 20-session average of 17.7 billion (Reuters).
  • The VIX rose 11.57 points, its biggest point gain since 4 April 2025 (CNBC), to close at 28.17 (TheStreet).

What set it off

Reuters tied the sell-off to the midterm elections. Democrats won control of Congress, and Tuesday’s results included states where opposition to new data centres had grown. Investors feared that tighter AI regulation could follow and slow the build-out of AI infrastructure.

Chip news after the close on 4 November added to it. Qualcomm fell 9.5% after missing Wall Street estimates for fourth-quarter smartphone-chip sales, Reuters reported. Data-centre names fell harder than the chipmakers: IREN and Applied Digital each lost more than 14%, and CoreWeave fell almost 10%.

One portfolio manager, Jed Ellerbroek of Argent Capital Management, told Reuters it felt “like the capitulation phase for the data center capex beneficiary stocks”. That is one view on the day, not a conclusion.

A correction, by the textbook definition

A correction is a fall of 10% or more from a recent high. It is a label, not a signal. It says how far the market has fallen, not what happens next.

The Nasdaq’s record close was 27,599.79 on 6 October 2026. Its 5 November close of 24,315.00 was 11.9% below that. Reuters said the index ended more than 10% below its record, confirming a correction; AP put it at 11.9%, which matches the arithmetic.

The S&P 500 did not enter a correction that day. TheStreet noted it closed at its lowest level since 29 July.

Then the rebound week, and a 10-year at 5.15%

The week of 9–13 November reversed part of the damage.

Index Close, 13 November Week
S&P 500 7,561.67 +2.51%
Nasdaq Composite 25,136.27 +2.27%
Dow Jones Industrial Average 52,347.55 +2.07%

Source: Portfolio Terminal weekly review, 13 November 2026. It was the S&P 500’s best week since 7 August 2026.

Two details matter more than the headline. First, the Nasdaq’s Friday close was still about 8.9% below its 6 October record, so the correction had narrowed, not closed. Second, the leadership changed. Energy rose 5.42% on the week and industrials 3.76%, while Nvidia, the stock that led the 5 November losses, fell 12.75% after its chief executive resigned. We cover that separately in Nvidia without Jensen Huang.

The 10-year above 5%

Treasury yields moved the opposite way from stocks on the rout day and then rose through the rebound. On 5 November, as stocks fell, Treasuries rallied and yields fell. The 10-year rose above 5% only in the following week.

Date 10-year 2-year
4 November 4.90% 4.58%
5 November 4.84% 4.46%
6 November 4.86% 4.51%
9 November 4.88% 4.56%
10 November 4.98% 4.73%
12 November 5.10% 4.86%
13 November 5.15% 4.88%

Source: US Treasury daily par yield curve. There is no 11 November print because the bond market was closed for Veterans Day; the stock market was open. Market quotes for the 10-year closed the week at 5.154%; the official par yield is 5.15%.

The 2-year rose 37 basis points in five sessions, including a jump from 4.56% to 4.73% on 10 November, the day of the consumer price index release. Stocks rallied and yields climbed through the same week, which suggests the bond market was looking past the 5 November scare. That is not unusual. It also means the equity rebound came with a higher cost of money attached.

For context, the 10-year was not at a new high. The official par yield was 5.31% on 5 October and 5.19% on 14 October. For how yields and equities feed into each other, see our intermarket analysis guide and US Treasury yields are breaking out.

What a crowded-trade unwind looks like

A crowded trade is a position so many investors hold, for the same reason, that there is nobody left to buy and plenty of people who could sell. While the reason holds, a crowded trade can rise for a long time. When something challenges the reason, everyone heads for the same exit.

5 November had the features of that kind of exit:

  • The crowded names fell most. Semiconductors fell more than 12%, tech 7.3%, the S&P 500 4.36% and the equal-weighted S&P 500 2.9%. The closer a stock sat to the AI theme, the harder it fell.
  • Volume surged. US exchange volume was about 50% above its 20-session average, and XLK traded more than three times its usual volume. Forced and urgent selling looks like that.
  • Breadth collapsed with it. Almost six NYSE stocks fell for every one that rose. When a crowded trade unwinds, the selling spills into things that are only loosely linked, because the same funds own them.

The week that followed shows why one rebound does not close the book. Money rotated into energy and industrials rather than back into the AI leaders. The Nasdaq remained in correction territory, and the 10-year moved above 5%. A bounce after a crowded unwind tells you that sellers ran out for a while. It does not tell you whether the reason the trade was crowded still holds.

The historical rhymes, used carefully

History can show what similar episodes looked like. It cannot tell you which one this is.

The dot-com bust. The Nasdaq peaked in March 2000 and fell about 78% over the following two and a half years, with sharp rallies on the way down. The early sharp falls looked like healthy corrections, and some of the rallies looked like the end of the selling. Our full account is in Pets.com and a Trillion Dollars of Hope, and the data comparison with today is in AI Bubble vs Dotcom Bubble: What the Data Actually Says. That piece also sets out where the comparison breaks down.

The telecom bust of 2000–2002. The closer parallel to the data-centre build may be the fibre build of the late 1990s, financed by debt on the assumption that demand would grow into the capacity. We covered that in When the Bond Market Spoke First: CoreWeave, AI Debt and the Telecom Bust. CoreWeave’s near-10% fall on 5 November is a reminder of why that piece focused on the debt.

The 1994 bond massacre. In 1994 the Federal Reserve doubled its policy rate from 3% to 6% in a year, and long-term yields rose sharply with it. Stocks did not crash, but rising yields squeezed anything priced on distant earnings. It is the relevant history for a rebound that came with a 10-year above 5%. We told that story in The Great Bond Massacre of 1994.

None of these is a template. 1987, 1998 and April 2025 all produced violent one-day falls that were followed by recoveries. The comparisons are useful for one thing: knowing what the possibilities look like, so that a single good week is not mistaken for an answer.

Mind, Method, Money. A crowded-trade unwind tests all three. Mind: the urge to call the bottom after one green week. Method: whether your setup still works when breadth and leadership change. Money: whether a 12% day in your sector would breach your daily loss limit. If it would, the position was too large before 5 November.

Frequently asked questions

How much did the Nasdaq, S&P 500 and Dow fall on 5 November 2026?

The Nasdaq Composite fell 5.52% to 24,315.00, the S&P 500 4.36% to 7,353.73 and the Dow 3.44%, about 1,833 points, to 51,523.06, according to Reuters. All three had their worst day since 4 April 2025.

Did the Nasdaq enter a correction?

Yes. The 5 November close was 11.9% below the 6 October record close of 27,599.79, past the 10% threshold that defines a correction. After the rebound week it was still about 8.9% below the record.

What is a crowded-trade unwind, in practical terms?

Many investors holding the same position for the same reason, and selling at the same time when that reason is questioned. The signs are heavy volume, the most popular names falling most, and selling that spreads to unrelated stocks held by the same funds.

Why did the 10-year yield fall on the rout day and then rise above 5% the next week?

On 5 November investors bought Treasuries as a safe asset, pushing the 10-year down to 4.84%. The following week stocks recovered, inflation data came out on 10 November, and yields rose each session to 5.15% by 13 November.

Does a strong rebound week mean the sell-off is over?

No one knows. Sharp falls have been followed by full recoveries, as in April 2025, and by long declines with strong rallies along the way, as in 2000. One week of gains does not tell you which, and the Nasdaq was still in correction territory on 13 November.

How does 5 November compare with Liberation Day in April 2025?

Liberation Day was bigger for the broad market: the S&P 500 fell 5.97% on 4 April 2025 against 4.36% on 5 November. The 2026 fall was more concentrated in technology and semiconductors, which had their worst days since March 2020.

This article is education, not financial advice. It describes market data to 13 November 2026 and makes no forecast about indexes, individual stocks or yields.

Market Mayhem

The history of crashes, bubbles and panics

From the tulip mania to 2008, told as it happened, with what traders can take from each one.

See the book →

Sources: Reuters via Investing.com (5 November 2026); CNBC market statistics (5 November 2026); TheStreet (5 November 2026); MarketWatch via Dow Jones (6 November 2026); investingLive (5 November 2026); AP (5 November 2026); Dow Jones Data Talk via Morningstar (6 October 2026); US Treasury daily par yield curve; Portfolio Terminal, stock market week in review (13 November 2026).

Louw van Riet
Written by
Louw van Riet
Author · Trader · Coach

Louw is the author of The Complete Trader's Edge — a 70-chapter trading framework covering psychology, technical analysis, ICT concepts, and professional risk management. He has spent years studying institutional price action across forex, indices, and crypto, and built this platform to provide the complete, honest trading education he wished existed when he started.

The Complete Trader's Edge compass logo
Mind · Method · Money
Free Trading Plan Template

Get Your Complete Trading Plan

Subscribe and get the 8-page Trading Plan Template free — includes pre-session checklist, trade journal, risk rules, and weekly review system. Plus weekly insights on psychology, strategy, and risk management.

No spam. Unsubscribe anytime. Free forever.