Goldman Sachs Trading Desk: Momentum Trade Bottoming May Take Weeks; AI Capex Narrative Shaken

AI models achieving frontier performance with computing power inputs far below market expectations have once again challenged the market narrative that "only continuous expansion of capital expenditure can win the AI race." Goldman Sachs' One-Delta trading desk believes that the adjustment in momentum strategies has not yet ended, but there is currently no systemic risk in U.S. stocks, and the market structure remains resilient.

Rich Privorotsky, head of Goldman Sachs' One-Delta trading desk, stated that his momentum model shows it will likely take several more weeks for this round of momentum correction to truly bottom out, with the final decline probably close to the historical median level. However, given that the previous upward slope was far steeper than the historical average, there is also a possibility that this correction could exceed the historical mean.

He also pointed out that the recent emergence of new-generation efficient AI models is reigniting market reflection on the investment logic for AI infrastructure. As model training efficiency continues to improve, the core narrative of "continuously investing huge amounts of capital to build larger-scale computing clusters" is facing increasing skepticism, yet AI capital expenditure remains the most important pricing theme for global tech stocks today.

Momentum Trade Not Yet Fully Cleared; Market Internal Rotation Persists

Privorotsky stated that the momentum indicators he tracks show relative volatility remains at elevated levels, with no signals yet sufficient to lift the alarm.

However, significant divergence has begun to appear within the market. On one hand, some AI hardware-related stocks have entered oversold territory; on the other hand, some previously laggard sectors are rebounding without significant fundamental improvements, showing clear signs of fund rotation.

From an index perspective, U.S. stocks continue to demonstrate strong resilience. Correlations between sectors remain low, with funds switching more between industries rather than evolving into a broad-based sell-off. Even though implied volatility rose last Friday, this market structure has not fundamentally changed.

AI Efficiency Gains Again Challenge Capex Logic

Privorotsky stated that after practical testing of the Kimi K3 model, he was impressed by its engineering capabilities. The model has 2.8 trillion parameters and self-hosting still requires enterprise-grade GPU clusters; it cannot be run on ordinary local devices.

He believes that what is truly worth attention is not the inference phase, but training efficiency.

Rather than relying solely on larger computing scale, new-generation models rely more on algorithm optimization, innovations in model architecture, and more efficient Mixture of Experts (MoE) routing mechanisms to enhance training efficiency. Taking Kimi K3 as an example, it has 896 expert modules, but only 16 are activated during each inference, significantly reducing computational resource consumption.

This has prompted the market to rethink: if frontier models can significantly improve training efficiency through algorithmic innovation, does the AI industry still need to continuously build ever-expanding, capital-intensive data centers and training clusters?

However, Privorotsky believes this primarily challenges the investment logic on the training side, while demand for computing power on the inference side remains strongly supported, meaning the long-term demand for AI infrastructure has not undergone a fundamental reversal.

Earnings Season Will Determine Whether the AI Theme Can Continue

As the Federal Reserve enters its quiet period ahead of the interest rate decision, the market's short-term focus will be on macro events such as the European Central Bank's interest rate decision, the UK CPI, and the preliminary PMI readings for major global economies.

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

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

Risk Warning and Disclaimer

The market carries risks; investment requires caution. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial status, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Responsibility for investments made based on this content lies solely with the investor.