I'm LongbridgeAI, I can summarize articles.The UBS trading desk believes that last week's US stock market movement was essentially "deleveraging rather than de-risking"—with targeted selling in AI and momentum sectors, memory stocks dropping 16% in a single week, and capital rotating into defensive and value sectors, yet overall market exposure was not systematically abandoned. Hedge funds' total exposure remains at a high level in the 90th percentile, indicating lingering risks of further deleveraging. The previous "AI guru liquidation" was similarly driven by forced liquidation of crowded positions, rather than fundamental-driven systematic selling
Last week's experience in the US stock market was not a broad-based sell-off, but a targeted position reset. Torsten Sippel, a trader at the UBS trading desk, pointed out in his latest weekly report that last week's market movements bore a strong resemblance to the deleveraging episode triggered months ago by "Situational Awareness" (a well-known AI quantitative fund)—with AI and momentum sectors under continuous pressure, capital rotating into defensive and less crowded sectors, yet overall risk exposure was not systematically abandoned.
AI and momentum sectors became the epicenter of this correction. Semiconductors, memory, and AI infrastructure faced sustained selling pressure, with almost no capital stepping in to buy the dip. The memory sector fell 16% in a single week, speculative growth stocks dropped 12%, and momentum long/short portfolios declined 8%. Meanwhile, consumer staples rose 5%, and value long/short portfolios gained 6%, making the shift in market style from growth and crowded momentum toward quality and value quite evident.
Hedge funds' net exposure has dropped to the 30th percentile of its five-year history, but gross exposure remains above the 90th percentile. This structure implies that if profit-and-loss pressure persists, the market still faces the risk of further deleveraging. Sippel explicitly stated in the report that the recent market landscape is "fragile rather than thoroughly bearish," with risks being reallocated rather than abandoned.
Deleveraging Rather Than De-Risking: The Essence of This Correction
Sippel characterized last week's market behavior as "De-Grossing rather than De-Risking."
From the perspective of capital flows, high-touch trading volume in the US market increased significantly. The buy/sell ratio for long-only funds was 36/64, while the buy/sell/short ratio for hedge funds was 42/50/8. Long positions continued to be reduced, but new single-stock short positions were relatively limited, which does not align with the characteristics of a comprehensive risk clearance.
This pattern mirrors the deleveraging episode months ago involving Situational Awareness, the hedge fund founded by "AI Guru" and former OpenAI researcher Leopold Aschenbrenner—at that time, it was also a case of forced liquidation of crowded positions, rather than fundamental-driven systematic selling. In this round of market moves, software, defensive sectors, and low-crowding exposures absorbed some of the rotating capital mid-week, while consumption, housing, healthcare, and certain industrial sectors also saw capital inflows, with small-cap stocks and equal-weight indices outperforming the broader market.
AI Financing Logic Questioned: CoreWeave Becomes a Key Bellwether
CoreWeave has become an important benchmark for observing AI trades. Sippel pointed out that as concerns over the company's creditworthiness continue to expand, market discussions around AI have shifted from the demand side to more fundamental questions: Can the economics and financing structure of AI infrastructure support the next phase of capital expenditure expansion?
Notably, both ASML and Taiwan Semiconductor released strong fundamental data, but this failed to quell the aforementioned concerns. This divergence is crucial—it means that even though hardware demand data remains impressive, the market has developed an independent pricing logic regarding the sustainability of AI capital expenditure financing.
In this context, the upcoming earnings reports of Hyperscalers will face a higher threshold for validation—investors need to see that cloud revenue, order backlogs, and cash generation capabilities can genuinely support the scale and financing needs of AI capital expenditures.
Macro Factors Amplify Rather Than Trigger Deleveraging
Sippel emphasized that macro factors played the role of an amplifier rather than a trigger in this correction.
Inflation and interest rate shocks accelerated systematic selling, while the escalation of tensions in the Middle East increasingly transmitted volatility to the markets. The VIX spiked on Friday, with the 1-month/3-month VIX term structure inverting for the first time in six months, indicating that hedging demand has spread from AI-specific risks to broader market risks.
The market weakness on Friday afternoon was particularly noteworthy—the sell-off in the tech sector showed distinct indiscriminate characteristics. Sippel believes this aligns more with the behavioral patterns of ETF redemptions, option expirations, and hedge rollovers, rather than incremental selling based on individual stock fundamentals.
Earnings Season Has Not Yet Become the Dominant Variable
Currently, the overall performance of earnings season is decent, but its impact on market trends is limited. 10.4% of the S&P 500's market capitalization has been disclosed, with expected Q3 revenue growth of 12.1% and earnings per share growth of 26.5%. Actual earnings beat expectations by 4.1%, with 75% of companies exceeding expectations.
However, the strong performance of the banking sector and positive AI-related data failed to effectively digest the existing selling pressure. Earnings reports currently remain subordinate to position logic.
In the next five trading days, 80 companies will disclose their results, accounting for a combined 13.8% of the S&P 500's market capitalization, including Netflix, General Electric, Meta Platforms, Tesla, Procter & Gamble, and Western Digital. Sippel specifically mentioned that Meta is reporting earnings a week earlier than usual, implying with a question:
"Is it trying to get ahead before the (secondary offering) window closes?"
Overall, Sippel's judgment on the recent market is "fragile rather than thoroughly bearish."
Risks are being reallocated rather than abandoned, but given the lack of genuine bottom-fishing capital in the AI sector, gross exposure remaining at high levels, and recent pressure on the volatility curve, there is still room for further deleveraging. Additionally, the recent rise in volatility has sparked a new round of doubts about the sustainability of capital market issuance windows.