Hedge funds rapidly unwound technology and AI-related positions in October: reports citing Goldman Sachs prime-brokerage data say five consecutive selling sessions erased their cumulative net buying in global AI stocks since early 2025. The selling was led by reductions in existing long positions, but the figures do not show that funds sold every AI share they owned—or that the AI investment outlook has changed decisively.
3
14
What the reported figures show
Goldman’s prime-brokerage data, as reported by several outlets, indicates that hedge funds net-sold global AI-related stocks for five consecutive sessions, reversing their cumulative net purchases since early 2025. That describes a change in net flows; it is not a measure of all shares held by hedge funds or a complete account of every investor’s activity.
3
14
The broader technology selloff was also large. Reports put hedge-fund technology-stock sales at about $20 billion over two weeks, close to the roughly $23 billion sold during July’s unwind.
49
59 On the heaviest selling day, existing holdings were sold about 3.5 times faster than new short positions were opened, suggesting that cutting long exposure played a larger role than establishing fresh bearish bets.
1
13
Technology positions nevertheless remained elevated: one report put technology at 23.5% of Goldman’s prime book, around the 90th percentile of its five-year range. The selloff, then, did not necessarily mean positioning had returned to normal.
53
Why hedge-fund favorites took a hit
Goldman’s basket of stocks favored by hedge funds fell 4.2% on October 22, its worst daily decline since 2011, and was down 10.4% for October, according to Forbes. The report also said fundamental long-short, systematic and multi-strategy funds each lost more than 1% that day—the first time all three had done so simultaneously since the pandemic began.
5
The pressure reached trend-following funds, too. Reuters, citing a Goldman note, reported that they had their worst week since 2024 as worries about the eventual payoff from heavy AI spending contributed to volatility across stocks and bonds.
49
Together, those reports describe a broad positioning shock, not just a few funds adding short bets. But they also point to a mix of explanations: investors were reducing risk, while concern about the returns on AI spending was part of the market backdrop.
5
49
What the unwind does—and doesn’t—mean
A reversal of net buying is not the same as selling all existing holdings. Nor does the 3.5-to-1 comparison mean every sale was a forced liquidation: it indicates that sales of existing positions outpaced new short openings in the reported activity.
1
13
The numbers support the conclusion that hedge funds sharply reduced exposure to AI and technology stocks. They do not, on their own, establish that underlying demand for AI products has collapsed, that the selling is finished, or that share prices have reached a bottom. Reuters’ account also notes concern about the eventual payoff from AI spending, so the evidence does not reduce the episode to positioning alone.
49
The reports available here substantiate the main figures primarily through Goldman Sachs. They do not provide enough detail to independently compare Morgan Stanley’s prime-brokerage data with Goldman’s or verify the additional leverage and stabilization claims sometimes attached to this story.