All the speculative themes are being dumped! Goldman trading desk: Thursday’s momentum trades in U.S. stocks saw the biggest decline since the DeepSeek shock.

```

Market concerns over the massive financing demands of the AI cycle are brewing, leading to a sharp sell-off in US technology stocks. Momentum trading strategies have suffered heavy losses, and speculative sectors such as AI-related thematic baskets and Bitcoin-sensitive stocks have also faced fierce selling pressure.

The Nasdaq 100 index plunged more than 2% on Thursday, closing at the day's low, and has declined on five of the past six trading days. This sell-off has concentrated its impact on momentum trading strategies and AI-related assets. Although the Nasdaq 100 is only about 5% from its historical high and is attempting to hold the key support of its 50-day moving average, market sentiment has clearly shifted to defensive.

Goldman Sachs data shows that its high-beta momentum paired trade (GSPRHIMO) plunged 7% on Thursday, marking its second worst performance this year and the largest single-day drop since the DeepSeek event. AI-related thematic baskets, Bitcoin-sensitive stocks, and speculative areas like quantum computing have all seen severe sell-offs.

Peter Callahan, a Goldman Sachs expert in the technology, media, and telecom (TMT) sector, said the trading desk has received many inquiries about "why the decline" and "what the catalyst is." The Goldman trading desk attributes the volatility to multiple resonant factors: investors taking profits ahead of Nvidia's earnings, a hawkish shift in Fed statements, and intensifying doubts about returns on AI infrastructure investments.

With only 30 full trading days left in the year, investors are waiting for market narratives and price trends to stabilize. Goldman Sachs notes that momentum exposure in prime broker portfolios remains high (76th percentile over one year, 88th percentile over five years), while seasonal pressures from year-end deleveraging and tax-loss harvesting are starting to appear. For now, investors are waiting for clearer signals—whether Nvidia’s earnings will beat expectations, or whether the Fed’s monetary policy path will become clearer—to judge when this wave of selling will stabilize.

Five Major Pressure Factors Lead to Sell-Off, Including the AI Bubble

Goldman's trading team listed five major triggers for this wave of declines.

Firstly, investors are choosing to cash in and reduce holdings of AI leaders ahead of Nvidia's earnings next week. Secondly, a Financial Times article questioning the "inflated" power demand from data centers has sparked concerns about an AI infrastructure investment bubble.

Hawkish statements from Federal Reserve officials have further dampened market sentiment. Comments by Boston Fed President Collins and Minneapolis Fed President Kashkari have weakened expectations for a rate cut in December.

Additionally, corporate cost-cutting news such as Verizon’s layoff announcement, and the volatility that could be triggered by an intensive schedule of economic data releases before the Fed decision in the next three weeks, have increased market uncertainty.

Goldman Sachs TMT expert Callahan noted that the macro backdrop has continuously worsened over recent months—from the third-quarter earnings season for internet companies, to the nonlinear growth in AI, to fatigue evident in the price trends of leading sectors (AI and large tech stocks). Multiple factors have left investors in a state of "confusion."

Momentum Trading Strategies Closely Tied to AI Narrative

Goldman Sachs warned last Friday that sell-offs in high beta (GSP1BETA) and idiosyncratic volatility (GSP1RSVL) could escalate into a momentum unwind by year-end.

Thursday's plunge confirmed this prediction. Data shows that momentum factors currently have a significantly higher correlation with high short interest, high idiosyncratic volatility, and high beta, while correlation with the high-quality factor is far below normal levels.

More noteworthy is that the long side that has outperformed both the short side and the broader market in high-beta momentum paired trades is now the worst hit area. This is directly tied to the rising correlation between momentum strategies and the AI narrative.

Goldman Sachs data shows that AI beneficiary stocks, relative to the S&P 500 excluding the "Magnificent 7," have already declined by 9%, while similar previous corrections have typically reached around 20%.

Goldman's unconstrained momentum strategy (GSPRHIMO) is currently clearly tilted toward high beta, high idiosyncratic volatility, and cyclical sectors—long information technology and industrials, while shorting healthcare. This structure makes the strategy particularly vulnerable in the current sell-off.

Rising AI Skepticism Weighs on Speculative Themes

AI-related sectors and the most speculative areas of the market saw broad declines on Thursday. Bitcoin-sensitive stocks and quantum computing, which are highly correlated with the momentum factor, also fell sharply.

Goldman Sachs notes that skepticism over AI is rising—Oracle’s CDS spreads have widened and leverage has increased, SoftBank’s sale of Nvidia shares, among other factors, are impacting its AI themed basket.

Previously, the performance of momentum strategies was mainly driven by the long side, while the short side performed in line with the broader market. But as the correlation between the AI narrative and the momentum factor nears historic highs, this structural exposure is now becoming a source of risk.

The Goldman trading team notes that Nvidia’s earnings next week will be a key catalyst. Until then, hedging of momentum exposure remains cautious. The firm’s momentum-ex-AI basket (GSPUMOXX) offers investors a tool for factor hedging, while avoiding a direct bearish bias toward AI itself.

Risk Warning and DisclaimerThe market involves risk, and investment should be undertaken with caution. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their particular circumstances. Investing based on this information is at your own risk. ```