Hedge funds spent nearly two years building up AI positions, only to see them completely erased within five trading sessions. The intensity and speed of this sell-off is fueling fears of a much larger deleveraging wave.
According to a latest report from Vincent Lin, head of Goldman Sachs Prime Brokerage, hedge funds net sold global AI stocks for five consecutive trading days from October 8 to 14, completely reversing all cumulative net buying since the start of 2025 鈥?in other words, every single net purchase over the past nearly two years was wiped out in a single week.
Meanwhile, Goldman Sachs Prime Book data shows global equities were hit by "massive net selling," with single-day net selling reaching -2.1 standard deviations on a one-year lookback basis. North America and Emerging Markets Asia led the declines, with South Korea and Taiwan becoming the most concentrated markets for net selling.
The direct impact of this sell-off has already manifested across multiple markets. South Korea's KOSPI index fell into bear market territory, triggering a circuit breaker in Seoul. Taiwan recorded a historic single-day foreign net outflow of approximately $9.3 billion, and central banks across multiple Asian countries were forced to intervene in currency markets. TMT directional long-short funds lost 2.7% in a single day, an extreme move of -2.6 standard deviations, with October cumulative losses widening to 4.5%.
Two Years of Net Buying Erased in Five Days
The scale of this AI stock sell-off ranks second only to the July crash in Goldman Sachs' historical data. According to Goldman Sachs Prime data, the single-day net selling of global AI stocks (GSXGGLAI constituents) was the second-largest in the past year, with the ratio of long selling to short selling at approximately 3.5 to 1 鈥?meaning this sell-off was primarily driven by long positions actively reducing exposure rather than short positions adding to their bearish bets.
Looking at the flow chart, hedge funds' net flow on the AI list (black line) has fallen back to the starting point of early 2025. Long flow (blue line) has continued to decline from its peak around February to March of this year, while short flow (red line) has remained near a low of approximately -13% throughout the entire period. Goldman Sachs' chart clearly illustrates this "round trip": the entire 2025-2026 net buying cycle was completely reversed within a week, while shorts continued to add positions during the decline.
Just last week, Goldman Sachs noted that hedge funds' net exposure to Mag7 as a proportion of total US exposure in its Prime Book was close to 22%, the highest level on record. Five days later, the same group of funds rushed for the exits en masse.
Positions Still Heavy, Deleveraging Far From Over
Despite the extreme selling pressure this week, from a longer-term perspective, AI positions remain elevated, and the room for deleveraging cannot be ignored.
Goldman Sachs data shows that global AI stocks' net allocation as a proportion of global single-stock net market value currently stands at 26.1%, down from a peak of 31.7% in early June. This places it at the 49th percentile over the past year 鈥?but still at the 83rd percentile on a three-year basis. The long-short ratio of AI portfolios has fallen from a June peak of 4.34 to 3.27, sitting at the 7th percentile over the past year, but still at the 64th percentile on a three-year basis.
In short, one-year data suggests positions have been relatively cleared out, but three-year data indicates that hedge funds as a whole remain significantly overweight. Rachel Hu from Goldman Sachs' Asia team noted that trading feedback from the Taiwan market suggests the "capitulation" level of this sell-off was less than in July, implying there may be more selling pressure to come.
JPMorgan's market intelligence TPM positioning indicator currently sits in "neutral" territory at approximately the 46th percentile, far above the 19th percentile low seen at the end of July. This means that room for "further deleveraging" still exists before key tech catalysts arrive.
Defensive Shift: Selling Tech, Buying Toothpaste
This sell-off was not confined to the AI sector but reflected a comprehensive defensive rotation. Goldman Sachs Prime Book data shows that 8 of 11 US sectors were net sold, with information technology and industrials leading the declines. Within information technology, all sub-sectors 鈥?software, semiconductors and semiconductor equipment, and IT services 鈥?were net sold.
Meanwhile, hedge fund buying was concentrated in healthcare and consumer staples, which Goldman Sachs characterized as a "marginal strengthening of a defensive posture." On Thursday morning, healthcare and consumer staples each rose more than 2%, supporting the Dow, while the Nasdaq slid toward a near one-month low. CoreWeave fell another 8.5%, and the Philadelphia Semiconductor Index declined for a seventh consecutive day, with approximately $2 trillion in related market value evaporating.
Amid this collective bloodletting by human fund managers, quantitative funds emerged as the sole winners. Global systematic long-short funds gained 0.5% on the day, bringing October cumulative gains to 1.9%, with a year-to-date return of 21.7% leading all strategies.
Fragile Leverage Structure, Violent Deleveraging Risk Persists
Leverage data reveals why this sell-off was so severe. Goldman Sachs Prime Book's overall gross leverage declined only marginally by 0.4 percentage points to 310.3%, sitting at the 72nd, 91st, and 94th percentiles on one-year, three-year, and five-year bases respectively 鈥?near five-year highs. Net leverage fell 1.5 percentage points to 73.3%, down 4.3 percentage points cumulatively in October, sitting at the 6th percentile over the past year.
Goldman Sachs' logic is this: a fund with 310% gross leverage and 73% net leverage has actual sellable capacity far exceeding what net exposure suggests. When crowded longs (AI) and crowded shorts move in opposite directions simultaneously, deleveraging becomes the only way out 鈥?and the exit is extremely narrow, with everyone rushing toward the same door at once. This is precisely a replay of the late-July crash.
Rich Privorotsky from Goldman Sachs' Delta-One business summed up the current predicament in an overnight report: "In hindsight, the pain trade was long AI, short bonds. Now we're seeing both ends unwinding simultaneously... The market is discovering that compute demand can grow exponentially, but the capital that can finance it cannot."
Goldman Sachs: This Looks More Like the Beginning Than the End
The core debate in the current market is whether this is a buying opportunity or the prelude to a larger deleveraging wave.
JPMorgan leans toward the former, believing it is an opportunity to buy the dip. Goldman Sachs' flow trading desk had spent the previous two days observing positive year-end technical signals. But Goldman Sachs' overall assessment is more cautious: AI net exposure remains at the 83rd percentile on a three-year basis, gross leverage sits at the 94th percentile on a five-year basis, and Asian foreign capital is withdrawing at record scale 鈥?these signals collectively point to this deleveraging round being closer to the starting point than the endpoint.
Hedge funds have already sold off all the AI positions they bought since 2025. But the positions accumulated in 2023 and 2024 remain fully on the books.
Comments