AI stocks suffer worst plunge since DeepSeek shock; Goldman Sachs high-beta momentum portfolio gives back year-to-date gains

The AI momentum trade is sharply reversing. The high-beta momentum portfolio, which had once gained about 70% year-to-date at its June peak, has now given back all of its year-to-date gains; the Philadelphia Semiconductor Index (SOX) has posted its worst performance relative to the S&P 500 over the past 9 trading days in more than 20 years.

On October 23, doubts about AI investment returns intensified further. Goldman Sachs' trading desk called this round of concentrated selling in AI and semiconductors the "worst performance since DeepSeek shock day," and the debate over AI spending costs and returns triggered by Meta's earnings became the catalyst for the acceleration of this selloff.

The selling is spreading from stock prices to portfolios. Goldman Sachs Prime data show that hedge funds' fundamental long/short net leverage has fallen to the 1st percentile of the past year, and the global AI stock long/short ratio has also dropped to the 6th percentile; meanwhile, UBS data show that the short leverage of systematic strategies has risen to the 99th percentile of the past five years.

Market style has shifted accordingly. As the high-beta momentum trade fades, low-volatility and earnings-quality factors are strengthening again, and capital is beginning to shift from chasing high-beta assets to strategies that place more emphasis on defensiveness and earnings certainty. Whether the AI trade can regain support depends on the market's reassessment of the returns on massive capital expenditure.

Year-to-date gains wiped out entirely, momentum trade hit hard

Data show that the Goldman Sachs High Beta Momentum Pair had year-to-date gains of about 70% at its June peak, and by October had given back all of those gains. The Unconstrained Momentum strategy is also experiencing its worst drawdown since the global financial crisis, second only to the pandemic shock period.

Goldman Sachs' trading desk said that last Thursday, previously strong stocks once again faced concentrated selling, with AI and semiconductors performing especially weakly. Meta's earnings intensified market doubts about the return on AI capital expenditure, and investors began to reassess whether massive investment can translate into corresponding returns.

The semiconductor sector's decline is especially prominent. According to Peter Callahan's data, in the 9 trading days since October 12, SOX fell about 13% relative to the S&P 500, the largest relative decline in an equivalent time window in more than 20 years, at the 1st percentile of the historical distribution.

The sustained decline has also pushed SOX to near its 200-day moving average, the first time the index has touched this key technical level since June 2025. Whether this moving average can provide support will become an important indicator for observing whether the semiconductor sector's correction deepens further.

Hedge funds accelerate de-risking, short leverage rises to high levels

According to Goldman Sachs Prime data, fundamental long/short net leverage fell 3.1 percentage points this month to 49.4%, at the 1st percentile of the past year's distribution and the 4th percentile of the past three years, indicating that hedge funds are significantly compressing net market exposure.

The global AI stock long/short ratio has fallen to the 6th percentile of the past year, and momentum factor exposure has also dropped to its lowest level since late June 2025. However, based on the five-year historical distribution, momentum factor exposure is still at the 73rd percentile and has not yet fallen to a long-term low.

The short side continues to heat up. According to UBS data, the short leverage of systematic strategies is close to a five-year high, at the 99th percentile. But long leverage has also risen to the 94th percentile, indicating that leverage on both the long and short sides is at high levels; the market is not simply turning one-sidedly bearish, but is readjusting positions and risk exposure.

Low volatility and earnings quality strengthen, capital turns defensive

As the high-beta momentum trade fades, the low-volatility factor has performed strongly recently and has once again become a favored allocation direction for capital; the earnings-quality factor has also strengthened. According to Goldman Sachs Prime data, its exposure has risen to a five-year high.

Goldman Sachs' trading desk noted that after sector, industry, and factor-neutral adjustments, the Goldman Sachs Quality Index has risen about 20% over the past three months, equivalent to about 13 times volatility performance, mainly driven by the sharp underperformance of low-quality stocks.

This change indicates that capital is shifting from chasing high-growth, high-beta assets to targets that place more emphasis on earnings quality and performance certainty. As investors reassess the costs, returns, and realization timelines of AI capital expenditure, valuation repricing of related assets may continue. Risk disclosure and disclaimer

Markets carry risk, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Investment based on this is at the user's own risk.