Hedge Funds Erase Two Years of AI Bets in Five Days
TREE NEWS reports: In a dramatic reversal that has rattled global markets, hedge funds have liquidated their entire cumulative net long position in global AI stocks built up since the start of 2025 — all within five consecutive trading sessions from October 8 to 14. Led by Vincent Lin, the net selling in AI names has completely unwound every net purchase made over nearly two years, marking one of the most violent position unwinds in recent memory.
The selling pressure was not confined to AI. Goldman’s Prime book shows global equities were “heavily net sold,” with single-day net selling reaching -2.1 standard deviations on a one-year lookback. North America and emerging Asia led the decline, with South Korea and Taiwan the most concentrated markets for net selling. The immediate fallout has been severe: South Korea’s KOSPI tumbled into bear market territory, triggering a circuit breaker in Seoul, while Taiwan saw a record single-day foreign net outflow of approximately $9.3 billion. Several Asian central banks were forced to intervene in currency markets.
Market Implications: A Broad Defensive Rotation
The sell-off reflects a sweeping defensive pivot rather than an isolated AI correction. Goldman’s data shows eight of 11 US sectors were net sold, led by information technology and industrials. Within tech, every sub-sector — software, semiconductors and semi equipment, and IT services — saw net selling. Meanwhile, hedge funds rotated into healthcare and consumer staples, which Goldman characterized as a “marginal increase in defensive posture.” On Thursday morning, healthcare and consumer staples each rose more than 2%, supporting the Dow, while the Nasdaq slid to a near one-month low. CoreWeave fell another 8.5%, and the Philadelphia Semiconductor Index dropped for a seventh straight session, erasing roughly $2 trillion in market value.
Quant funds emerged as the only winners. Global systematic long-short funds gained 0.5% on the day, up 1.9% for October and leading all strategies with a 21.7% year-to-date return. Discretionary TMT long-short funds lost 2.7% in a single day — a -2.6 standard deviation move — bringing October losses to 4.5%.
Leverage Structure Remains Fragile
The scale of the unwind is explained by leverage. Goldman’s Prime book shows gross leverage edged down just 0.4 percentage points to 310.3%, still in the 72nd, 91st, and 94th percentiles on one-, three-, and five-year horizons — near five-year highs. Net leverage fell 1.5 percentage points to 73.3%, down 4.3 percentage points in October, but at the 6th percentile for the past year. The logic is stark: a fund with 310% gross and 73% net leverage can sell far more than its net exposure suggests. When crowded longs (AI) and crowded shorts move against each other, deleveraging becomes the only exit — and the door is narrow.
Goldman’s Delta-One desk captured the dilemma: “In hindsight, the pain trade was long AI, short bonds. Now we see both legs unwinding… The market is discovering that compute demand can grow exponentially, but the capital to finance it cannot.”
Key Takeaways for Investors
- Positioning remains stretched: Despite the sell-off, global AI net allocation is still 26.1% of global single-stock net market value — 83rd percentile on a three-year basis. The long-short ratio for AI portfolios sits at 3.27, down from a June peak of 4.34 but still 64th percentile over three years.
- Deleveraging may not be over: JPMorgan’s TPM positioning gauge is at the 46th percentile — neutral but far above July’s 19th percentile low — suggesting room for further de-risking before key tech catalysts.
- Watch Asia flows: Record foreign outflows from Taiwan and Korea, plus central bank intervention, signal persistent pressure on Asian tech-heavy markets.
- Defensive rotation is underway: Healthcare and consumer staples are the new destinations for hedge fund capital, while semiconductors face continued headwinds.
The core debate now: is this a buying opportunity or the start of a larger deleveraging cycle? JPMorgan leans bullish, but Goldman is more cautious. With AI net exposure still at the 83rd percentile over three years and gross leverage at the 94th percentile over five years, the evidence suggests this is closer to the beginning than the end.
