.png "Global Insight")
Headlines say artificial intelligence stocks have lost about $5 trillion since early October. That number is real — and it is only half the story. While the biggest AI names fell sharply, hundreds of other large companies actually gained value. What happened in October was less a market crash than a rotation: money moving out of a handful of crowded trades and into the rest of the market. This follows October's Nasdaq selloff, when AI spending fears first hit Wall Street.
How big was the fall?
According to Yahoo Finance calculations reported on 1 November, AI and semiconductor stocks have shed roughly $5 trillion in market value since 6 October. Nvidia and Alphabet each lost approximately $835 billion in market capitalisation, and Intel lost nearly half its value during October. The iShares Semiconductor ETF fell about 17% in October, dipping as low as $460 on Friday before closing near $472 — below its July low of $464 intraday, though it remains more than 50% higher for the year.
The headline indexes felt it: the S&P 500 fell 3.1% from its 6 October record of 7,818.93, and the Nasdaq declined 7.8% from its October peak. But the equal-weight S&P 500 — which treats every company the same instead of letting giants dominate — gained 4.9% since 30 September. And 625 other large US stocks collectively added approximately $1.3 trillion over the same period. In the latest week, nine of the S&P 500's 11 sectors rose at least 1%, an unusually broad advance in a declining week for the index that Yahoo Finance calculates has not occurred since at least 1999. The same figures appear in a second account of those calculations.
Why did AI stocks fall if earnings are still growing?
Part of the answer is valuation. After a long run, AI leaders were priced for perfection, so weaker-than-expected third-quarter results from Alphabet and Intel hit hard. Yet analysts still expect S&P 500 earnings per share to grow 16.1% in 2027, and projected fourth-quarter earnings growth remained at 30.2%, according to Yardeni Research. The S&P 500's forward price-to-earnings ratio fell to 18.2, its lowest since April 2025 — cheaper, not broken.
Part of the answer is macroeconomics. Bank of America's Michael Hartnett has stayed cautious, pointing to 42 central bank interest-rate increases worldwide over the past six months; the Federal Reserve held rates steady on Wednesday, but two officials voted for an increase. And part of it is seasonal: Jeff Hirsch, editor of the Stock Trader's Almanac, calls the October turbulence "Octoberphobia" and notes that his firm bought semiconductor stocks on Friday even as the sector struggled.
What history says — and what it cannot promise
Hirsch points to a striking seasonal pattern: since 1950, the S&P 500 has risen in all 19 periods stretching from the end of October in a midterm election year through the following June, with a median gain of approximately 15%. Semiconductor stocks rose in all eight comparable periods since 1994, with a median gain of roughly 22%. History, however, is not a guarantee — every one of those periods also survived wars, recessions and inflation shocks, and past patterns can break.
The week ahead will test the rotation. Tuesday brings the US midterm elections, followed by earnings and jobs data in a market already wary about AI spending. The question for investors is no longer whether AI stocks can fall — they just did, by trillions. It is whether the rest of the market, which quietly rose while they fell, can keep carrying the load. That is a very different question from the one the headlines ask, and a more useful one.