Goldman Sachs Trading Desk: Momentum trading may still take several weeks to bottom out, and the AI capital expenditure narrative is starting to waver.

```

AI models have achieved cutting-edge performance with computing power far below market expectations, once again challenging the narrative that "only continuous expansion of capital expenditures can win the AI competition." Goldman Sachs One-Delta trading desk believes momentum strategy adjustments are not yet finished, but there is no systemic risk in the current US stock market, and market structure remains resilient.

Rich Privorotsky, head of Goldman Sachs One-Delta trading desk, said that their momentum model indicates the current momentum trading (Momentum) correction is still several weeks away from truly bottoming out, with the final decline likely to be close to the historical median level. However, given that the previous upward slope far exceeded the historical average, there is also the possibility that this adjustment may surpass the historical mean.

He also pointed out that the emergence of a new generation of efficient AI models is prompting the market to rethink the logic of investing in AI infrastructure. As model training efficiency continues to improve, the core narrative of "continuously investing huge capital to build larger-scale computing clusters" is facing more and more skepticism. AI capital expenditures remain the most important pricing theme for global technology stocks at present.

Momentum trading has not yet been fully cleared; market rotation persists within sectors

Privorotsky said that his tracked momentum indicators show relative volatility is still at a high level, and no signals strong enough to lift alarms have yet appeared.

However, clear divergence has started to emerge within the market. On one hand, some AI hardware-related stocks have entered oversold territory; on the other hand, some previously lagging sectors are seeing rebounds despite no obvious improvement in fundamentals, with pronounced capital rotation characteristics.

From an index perspective, US stocks overall still demonstrate strong resilience. Correlations between sectors remain low, and capital is switching more between industries rather than turning into broad-based selling. Even though implied volatility rose last Friday, this market structure has not fundamentally changed.

AI efficiency improvements challenge capital expenditure logic once again

Privorotsky stated he was impressed by the engineering capabilities of the Kimi K3 model after actual testing. This model has 2.8 trillion parameters; self-hosting still requires enterprise-grade GPU clusters, not ordinary local devices.

He believes that the real point of focus is not the inference phase, but training efficiency.

Compared to simply relying on larger computing power, the new generation of models mainly improves training efficiency through algorithm optimization, innovation in model architecture, and more efficient Mixture of Experts (MoE) routing mechanisms. For example, Kimi K3 has 896 expert modules, but only 16 are activated during every inference, significantly reducing computational resource consumption.

This has prompted the market to rethink: if cutting-edge models can significantly improve training efficiency through algorithmic innovation, does the AI industry still need to keep building ever-larger, capital-intensive data centers and training clusters?

However, Privorotsky believes this mainly challenges the investment logic for the training side, while the demand for inference-side computing power still has strong support, and long-term demand for AI infrastructure has not fundamentally reversed because of this.

Earnings season will determine whether the AI theme continues

With the Federal Reserve entering its quiet period before the rate-setting meeting, the market's short-term focus will turn to macro events such as the ECB meeting, UK CPI, and preliminary PMI readings from major global economies.

However, Privorotsky believes that the upcoming earnings season will truly determine the direction of the market.

Besides Alphabet, the performance of tech companies such as Tesla, Texas Instruments, Intel, and AMD's upcoming "Advancing AI" event will serve as important windows for observing the AI investment cycle, further testing whether the trillion-dollar AI capital expenditure logic can continue to gain market recognition.

Risk Warning and DisclaimerThe market carries risks; investment requires caution. This article does not constitute individual investment advice and does not take into account any specific user's investment objectives, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their particular circumstances. Investment based on this content is at your own risk. ```