Goldman Sachs: In July, the crowded trades were smashed, and while the U.S. stock market bull run hasn't ended, it's become harder to navigate.

In July, the U.S. stock market didn't experience a crash at the index level; rather, it felt more like a liquidation at the positioning level. The S&P 500 held its ground this week, with a total fluctuation range of only 3.5% throughout July, remaining less than 2% away from its peak. Paradoxically, the equal-weighted S&P, low-volatility S&P, and S&P 500 excluding AI all hit record highs this week. Tony Pasquariello, head of Goldman Sachs' hedge fund business, wrote in the latest market observation: "After experiencing a truly exponential rise in high-speed trading, a heavy hammer has smashed through the consensus positions over the past month; I tend to believe this frenzy has cooled off." The emphasis is not on the disappearance of risk, but rather that the most crowded, easiest-to-leverage trades are being forced to cool down. A surface calm coexists with underlying volatility. The S&P 500's average daily fluctuation was less than 1% this week, yet Goldman Sachs' flagship momentum basket had an average daily fluctuation of nearly 10%. On June 22, Goldman Sachs' TMT momentum basket had a year-to-date gain of as much as 145%, only to experience the most significant recorded pullback, followed by a single-day rebound of 17%. Asian fundamental long-short funds had record performances in the first half of the year but then faced the largest single-month pullback in the past decade, while the South Korean KOSPI surged 18% overnight. This framework ultimately leads to an uncomfortable conclusion: the outlook for the U.S. stock market still leans favorable, but the risk-reward ratio is no longer cheap, and the upside elasticity of global stocks is weaker than before. The bull market has not been declared out, but the next phase is not one of "buy and hold for easy gains." The index didn't collapse; instead, the crowded trades did. The most easily misjudged aspect in July was focusing solely on the S&P 500. The index did not send panic signals. With the S&P 500 less than 2% away from its peak and a fluctuation range of only 3.5% in July, it seems like a normal oscillation. However, active managers at a deeper level have experienced a different market: popular momentum, the AI chain, South Korean stocks, and Asian long-short strategies have all been squeezed out of leverage. The issue is not about how much it falls on a given day, but rather that the previously most profitable trades suddenly lost liquidity. Betting on the S&P 500 itself reveals stability; betting on high-momentum tech stocks, however, shows near-uncontrolled volatility. The critical split in July lies here: little turbulence at the index level, but a boat has already capsized at the positioning level. De-leveraging is not just a minor adjustment; it is a genuine cleansing. Several data points indicate that this round of de-leveraging has extended beyond normal portfolio adjustments. Global technology exposure has seen the largest sell-off in over five years. The asset management scale of Korean stock leveraged ETFs peaked at $53 billion in June and has now dropped to $15 billion. The total exposure seen by Goldman Sachs' prime brokerage has been reduced the most since the end of 2022. More detailed positioning changes also point in the same direction: fundamental long-short clients' exposure to momentum factors has fallen to the 28th percentile for the past year. Crowded trades have shifted from "everyone is on board" to a significant portion having already exited, even being forced to exit. This doesnt mean that painful trades won't return. It just signifies that compared to early July, the impulse to chase rallies in the market has evidently diminished, while liquidity and discipline have noticeably increased. The contradiction of AI trading has shifted from narrative to return on investment. In the latter half of July, AI trading faced not just simple profit-taking, but a more fundamental question: Can the colossal capital expenditure from ultra-large cloud vendors on AI yield sufficiently clear and sustainable returns? Market skepticism around this question intensified last week. This week, responses varied but were better than the most pessimistic scenarios. Meta has not demonstrated significant AI returns are already realized; Microsoft provided a clearer signal that capital expenditures are converting into revenue and AI products, with scalability in conversion; Amazon followed suit with results showing AWS growth accelerating and cloud business margins expanding. The credit spreads on bonds of ultra-large cloud vendors have also narrowed correspondingly. These changes are significant. If AI trading is left with "huge investments and distant returns," valuations will be pressured; however, if some companies can prove that investments are starting to revenue, the market will not treat the entire AI chain with a broad brush. However, differentiation has already emerged. The phase where merely attaching an AI label could boost valuations is now much more challenging after this round of cleansing. The Federal Reserve's communication has become less clear, and long-term interest rates are back to being a concern for the stock market. After the FOMC meeting, stock market traders did not feel significantly more at ease. The volatility on the long end of the U.S. treasury curve briefly spilled over into the stock market. Whats more troublesome is the change in communication style. The market was accustomed to higher transparency, but now it seems to have entered a more restrained, less explicit phase. Traders have to decipher policy direction with fewer clues, which inherently creates friction. What truly needs to be monitored is the policy direction, not every single word. However, for stocks, changes in long-term interest rates cannot be overlooked, especially for long-duration stocks. The valuations of AI, technology, and growth stocks are more sensitive to remote discount rates; once the long end of the global bond market continues to exert pressure, "a stable foundation" does not mean it will be comfortable every day. U.S. stocks still lean favorably, but upside elasticity has thinned. From a broader framework, the U.S. stock market has not lost its support. Economic performance is solid, earnings growth is strong, and capital flows are expected to turn more aggressive, with nearly $1 trillion in AI capital expenditures still circulating in the system. This explains why the S&P 500 can maintain its position while there is significant underlying de-leveraging. The index does not lack risk, but it has a sufficient number of support factors beneath it. However, this is not a signal to be aggressively bullish. The direction for U.S. stocks remains favorable, but the risk-reward ratio is at a mid-range, and the upward elasticity of global stocks is not as robust as in the previous phase. Short-term volatility is likely to persist. Summer liquidity is not conducive to risk transfer; once a certain type of position becomes crowded, illiquid, and structurally complex, volatility will be amplified. At the portfolio level, it is more suitable to enhance liquidity and reduce complexity than to continue pursuing the steepest trades. The NASDAQ's response: The bull market is still there, but the path will be tough. The NASDAQ 100 index has currently retraced 8% from its June peak but is still up 12% for the year. Over the past nine months, it has experienced six months of decline, yet the point-to-point increase remains at 9%. The price-to-earnings ratio has fallen back to the lower end of the range observed over the past few years. These figures clearly articulate the market's state: the trend is not deteriorating, but the process is difficult. For traders, the endpoint and the path are not the same thing. The main bull market in the NASDAQ is still ongoing, but if the future continues to follow a rhythm of "climb for a while, dump positions, then recover," making a profit will be more difficult than simply being right about the direction. July has already given a warning: the market does not reward overcrowding and does not forgive leverage. This article is reproduced from Wall Street Insight, author Pan Lingfei; GMTEight editor: Wenwen.