Goldman Sachs says AI stocks have flipped to the short side of momentum
Goldman Sachs has reported that AI stocks have flipped to the short side of momentum trades, marking a significant shift in market leadership.
For roughly three years, betting on momentum and betting on artificial intelligence were effectively the same wager. If you owned one, you owned the other, whether you meant to or not.
Goldman Sachs now says that relationship has broken. AI stocks have moved to the short side of momentum trades, and the bank’s 12-month momentum basket posted a record gain over its shorter-term counterparts.
The call came from Guillaume Soria of Goldman’s thematic baskets team, in a note published Thursday. The bank repeated the theme a day later in its US Equities Weekly Rundown.
A 12-month momentum basket ranks stocks by their performance over the past year. Shorter-term versions, such as Goldman’s 3-month factor, react much faster to recent price action.
When the two windows disagree sharply, it usually means market leadership has changed recently. On the day Goldman flagged the divergence, the 3-month momentum factor surged 5% while the 12-month momentum factor plummeted 6.7%, marking the largest single-day spread between these metrics in five years.
The catalyst is the AI trade itself. Goldman’s AI theme basket, which trades under the ticker GSPUARTI, fell nearly 45% from its peak. That is its steepest drawdown since the launch of ChatGPT.
A decline of that size changes how a stock gets classified. Names that once dominated the long side of momentum screens have lagged long enough to slide toward the short side.
Chips out, software in
The shift has not simply drained money from tech. According to Goldman’s findings, capital has been rotating from semiconductor-related stocks toward software.
The relative performance of software versus semiconductors notched its second-best single-day gain on record during the decoupling. For a pair trade that has mostly favored chipmakers during the AI boom, that is a notable reversal of fortune.
Goldman has also flagged a broader rotation away from AI and momentum leadership. Value, quality and breadth factors have been gaining ground, and the bank tied that change to shifting hedge fund positioning.
Goldman did not just describe the shift. Soria’s note, circulated to top clients, recommended hedging against further losses through put options.
The suggested vehicles included baskets such as GSXUHMOM or GSTMTAIP, which track key AI beneficiaries and medium-term winners. A put option gains value when the underlying asset falls, so it works like an insurance policy on a position.
The biggest implication is for anyone who treated momentum and AI as interchangeable. Systematic funds that run momentum strategies were, for years, quietly running a large AI bet. Now their screens point the other way.
The software-over-semiconductors move deserves attention, too. If the rotation persists, it suggests investors are reassessing where AI value actually accrues: with the companies selling the picks and shovels, or with those building applications on top.
Goldman’s own read is that the shift reflects changing hedge fund positioning rather than a single news event. The bank’s recommendation to hedge with puts implies it sees meaningful downside risk remaining in AI beneficiaries, without claiming the AI theme is finished.


