• Net leverage fell 4.2 percentage points to 73.4% in October, the sharpest monthly drop since March, according to Goldman Sachs (NYSE:GS)' prime brokerage data.
  • Fundamental long-short funds lost 3.5% and systematic strategies 3.6%, marking the first month since 2016 that both cohorts declined more than 3% together.
  • Hedge funds sold approximately $20 billion in technology stocks over two weeks, with five straight sessions of AI-stock selling erasing cumulative net purchases since early 2025.

A Swift De-Risking

Hedge funds aggressively reduced risk in October, cutting net leverage by 4.2 percentage points to 73.4%, the fastest monthly reduction since March, according to Goldman Sachs' prime brokerage desk. The move came as fundamental long-short funds lost 3.5% and systematic strategies declined 3.6%—the first time since Goldman began tracking in 2016 that both groups fell more than 3% in the same month.

The selling was concentrated in technology, industrials and materials, while healthcare and real estate attracted buying. Over two weeks, funds dumped roughly $20 billion in technology stocks, close to July's $23 billion unwind. Five consecutive sessions of net AI-stock selling from October 8–14 wiped out cumulative net purchases recorded since early 2025.

Risk Reduction, Not Bearish Bets

Despite the heavy selling, Goldman's data suggests managers were trimming exposure rather than building short positions. Selling of existing long positions exceeded short selling by about four to one across October. Funds also continued to buy macro and index instruments, and their overweight to Europe reached a one-year high of 3.5%.

"What we're seeing is a broad-based de-risking, not a wholesale abandonment of equities," said a person familiar with the bank's prime brokerage flows, who spoke on condition of anonymity. "Managers are rotating, not retreating."

The turbulence was linked to growing doubts over the returns from massive AI investments. Meta Platforms (NASDAQ:META) fell about 10% in the week before the selloff after indicating it would need to spend more on AI infrastructure. The Nasdaq 100 dropped 5.5% that week, its worst weekly performance since April 2025. Asia-focused managers were hit hardest, losing 9.4% in October.

The Financial Stability Board warned in August that borrowing, high valuations and cross-investment among AI companies could amplify a correction. Those concerns appear to have materialized, at least partially.

A Strong Quarter for Goldman

The bank reporting the client activity is not the one suffering the strategy losses. Goldman Sachs reported third-quarter net revenue of $18.68 billion, up 23% year over year, and diluted earnings per share of $17.01, up 39%. Equities revenue jumped 72% to $6.44 billion, while asset and wealth management revenue rose 14% to $5.05 billion.

Goldman's prime brokerage unit, which tracks hedge fund client positions, generated higher financing revenue even as those clients posted October losses. A spokesperson for Goldman declined to comment on specific client flows.

For now, the data points to a cautious hedge fund industry navigating a treacherous market for AI-related trades. Whether the de-risking continues into November remains an open question. "The pain has been real, but positioning is cleaner," said the person familiar with the flows. "That could set up a more stable market—or simply precede another leg lower."

Correction: An earlier version of this article misstated the year of Goldman's tracking start. It began in 2016, not 2015.