AI Momentum Trade Hits a Wall
TREE NEWS reports: Goldman Sachs is warning clients that the artificial intelligence trade is facing its most severe structural test since the current bull market began, as momentum strategies split apart in ways not seen in five years. In a client note dated September 15, strategist Guillaume Soria highlighted an extraordinary divergence: the three-month momentum index (GSPRHMO3) rose 5% on the day, while the twelve-month momentum index (GSPRHIMO) fell 6.7% — the widest single-day performance gap in five years. At the same time, the bank’s index tracking AI-linked stocks (GSPUARTI) has fallen nearly 45% from its peak, the deepest drawdown since ChatGPT was released.
Inside the Momentum Factor’s Fracture
The internal structure of momentum trading is shifting rapidly. Capital is clearly rotating away from the “old momentum leaders” that had dominated the factor. Overall momentum exposure has declined: on a one-year basis, positioning sits at the 41st percentile, a notable retreat, but on a five-year view it remains at the 88th percentile — meaning systemic deleveraging has not yet fully played out.
Momentum volatility is also running well above the broad market. On the same day the S&P 500 fell less than 30 basis points, yet the momentum factor saw violent internal divergence. This low-correlation, high-volatility state makes hedging short-term momentum more cost-effective.
Money Rotates From Semiconductors to Software
The sector composition of momentum is rotating too. Short-term momentum flows are moving out of semiconductors and into software — a sector that was among the market’s most unloved at the start of the year. The relative performance index tracking software versus semiconductors (GSPUSOSE) posted its second-largest single-day gain on record. If the trend persists, flows could gradually feed into six- and twelve-month momentum baskets, forcing future rebalancing that buys software — through long additions and short covering — while selling semiconductors.
Notably, software had long held the largest weight on the short side of the twelve-month momentum index. Its rapid rebound signals not just a sector style shift but also a further widening of the gap between short- and long-term momentum.
The AI-Momentum Link Faces Unraveling
Although AI stocks and the momentum factor remain highly correlated — with one-month to one-year correlations between 90% and 96% — Goldman warns this long-standing bond is at risk of loosening. Sustained AI weakness would have profound implications for the momentum factor’s composition. As software gradually replaces semiconductors as a key driver of short-term momentum, the previously aligned paths of momentum and AI could decouple.
The highest-concentration momentum themes in the U.S. market remain in AI-related areas. But if the software-versus-semiconductor rotation continues, the sector makeup of momentum baskets will adjust further.
Key Takeaways for Investors
- Hedging matters: For portfolios with AI exposure, Goldman explicitly recommends buying puts on the mid-term winners basket (GSXUHMOM) or the AI beneficiaries basket (GSTMTAIP).
- Reassess the AI-as-momentum proxy: Strategies that treated AI as a proxy for momentum exposure need to be re-evaluated as the two decouple.
- Watch the rotation: Software strength and semiconductor weakness could drive the next round of momentum rebalancing.
- Deleveraging is incomplete: Five-year positioning remains elevated at the 88th percentile, leaving room for further unwinds.
Momentum trading is entering a new risk-and-reward phase, one where the crowded AI trade no longer guarantees the momentum factor’s direction.
