After a historic selloff in momentum stocks during July, institutional investors across Asia have significantly cooled their enthusiasm for AI-linked trades, yet Goldman Sachs believes the market's prevailing caution has now gone too far.
John Flood, a partner in Goldman Sachs' FICC and equities division, completed a week-long client survey across Asia and released a report on September 10 stating that Asian investors currently harbor widespread skepticism toward AI trades, a sentiment deeply tied to the "psychological trauma" left by the sharp July drawdown.
According to Goldman Sachs' prime brokerage data, Asian hedge funds have still posted roughly 21% positive returns year-to-date despite the July pullback. Flood argues that multiple positioning and sentiment indicators show investor wariness has exceeded what fundamentals can justify.
Flood also raised a key observation: American investors may be underestimating the actual adoption pace of China's open-source AI models, while Asian investors might be overestimating it, with the final reality likely landing somewhere in between. He further noted that the upcoming IPO pipeline could act as a catalyst to reignite active participation from both institutional and retail investors.
Asian hedge funds suffered heavy losses in July, and caution has spread across the entire market
In his report, Flood described being "stunned" by the widespread caution and skepticism toward AI trades among local investors during the Asia survey. The historic July momentum selloff severely eroded year-to-date gains for many Asian hedge funds, with some funds losing over 20 percentage points of their annual returns in a single month.
This experience has left obvious "psychological scars" on market participants. Flood noted that following July's historic momentum reversal and August's lackluster performance, he observed a general reluctance among investors to take proactive risk. In every conversation with clients, concerns about geopolitics, interest rates, and inflation were repeatedly raised, while positioning and sentiment indicators remained subdued.
Multiple data points suggest underweight positioning and elevated bearish sentiment
Goldman Sachs listed a series of positioning and sentiment indicators to support its view that the market is excessively cautious.
On the sentiment front, retail investor sentiment has consistently remained in bearish territory, with Goldman Sachs' sentiment indicator retreating into negative territory, its lowest level since March. Short positions in Nasdaq 100 index futures have risen 35% since mid-June, and average short interest across S&P 500 components sits at elevated levels.
On institutional positioning, Goldman Sachs' prime brokerage data shows that US fundamental long-short funds carry gross and net leverage of 207% and 49.8%, respectively, placing them at the 20th and 6th percentiles over the past year, and the 59th and 10th percentiles over three years. The US fundamental long-short ratio has pulled back to near its "Liberation Day" low.
In the tech and AI sectors, total and net allocations to US information technology stocks have fallen sharply from five-year highs reached in June, now sitting at the 43rd and 71st percentiles over the past year. The long-short ratio for Goldman Sachs' US broad AI stock basket (GSTMTAIP) has dropped to 3.54, down from its one-year high of 4.27, placing it at the 38th percentile for the past year.
For mutual funds, while cash balances as a percentage of total assets remain low, absolute nominal amounts still exceed historical averages. AI-related stocks have seen increased weightings in mutual fund portfolios but still lag index weights significantly, leaving funds generally underweight AI-exposed stocks. During the second quarter, most mutual funds trimmed positions in "Magnificent 7" constituents, with the most pronounced reductions seen in Alphabet (GOOGL) and Meta (META).
Strong Q2 earnings reveal a divergence between fundamentals and sentiment
Flood emphasized that current caution stands in clear contrast to corporate fundamentals.
In the second quarter of 2026, S&P 500 earnings per share grew roughly 30% year-over-year (excluding certain "other income" tied to private equity investments). Hyper-scale cloud providers and AI infrastructure companies benefiting from their capital expenditures saw earnings surge 54% year-over-year, contributing approximately 50% of the S&P 500's quarterly EPS growth.
Notably, earnings growth in the rest of the market remains robust and accelerating. Excluding the energy sector, which was boosted by rising oil prices, the remaining S&P 500 constituents posted 14% year-over-year EPS growth, demonstrating broad-based improvement.
Options market signals: market is not fully invested, with demand for upside protection emerging
Observations from Brian Garrett, a colleague at Goldman Sachs, further support this view. He points out that whenever the S&P 500 attempts to break to new highs, doubts about intraday strength quickly emerge; the index's upside volatility is nearly twice its downside volatility; and the implied correlation between S&P 500 call option volatility and spot prices is notably elevated. Garrett argues these are not typical behaviors of a "fully invested" market, and the upcoming earnings season could amplify this dynamic.
From an options structure perspective, the contrast between single-stock call skew and put skew similarly confirms the cautious tone—some investors are purchasing "just in case" upside insurance rather than actively adding positions to bet on gains. Flood believes that, taken together, these signals suggest the market's defensive positioning leaves room for potential upside moves, especially against the backdrop of high-profile IPOs materializing and a strong earnings season.