Goldman Sachs: Asian Institutional Caution on AI Trades Has Overshot Fundamentals
nashnova research
Goldman partner John Flood, after a week-long Asia client tour, says institutional skepticism toward AI trades has exceeded what fundamentals justify — the July momentum sell-off left deep scars, yet earnings data suggest this defensive stance may be missing real upside.
Why are Asian institutions so cautious?
The root cause is July's historic momentum-stock sell-off — some Asian hedge funds lost more than 20 percentage points of year-to-date gains in a single month, leaving visible psychological scars.
This means → even though these funds are still up for the year (Goldman data shows Asian hedge funds returning roughly 21% YTD), traders' risk appetite was locked in place by that drawdown.
In plain terms = they made money but got badly spooked — better to leave gains on the table than face another crash.
How light are positioning and sentiment?
Retail investor sentiment remains bearish; Goldman's sentiment indicator has fallen back into negative territory, the lowest since March.
Nasdaq 100 futures short interest is up 35% since mid-June; U.S. fundamental long-short funds carry a net leverage of just 49.8%, sitting at the 6th percentile over the past year.
This means → both retail and institutional positioning is extremely light — the market isn't "fully short," it's "huddled in a corner, watching."
How much has AI-sector exposure shrunk?
U.S. information-technology gross and net allocations hit a five-year high in June, then dropped sharply — now at the 43rd and 71st percentiles over the past year, respectively.
Goldman's broad U.S. AI equity basket (a group of stocks tracking broad AI beneficiaries) sees its long-short ratio at 3.54, down from the one-year high of 4.27, sitting at the 38th percentile.
Mutual funds were meaningfully underweight AI-exposed stocks in Q2, with the largest trims in Alphabet and Meta.
Are fundamentals actually weak?
Q2 S&P 500 EPS grew roughly 30% year-over-year; hyperscalers (Amazon AWS, Microsoft Azure, Google Cloud and peers) and AI infrastructure companies posted 54% YoY earnings growth, contributing about 50% of the index's quarterly EPS gain.
Strip out energy, and the rest of the S&P 500 still grew EPS 14% YoY.
This means → the earnings improvement has breadth — it is not a handful of tech giants carrying the index alone, which is precisely the condition the cautious camp feared was missing.
What is the options market saying?
Goldman's Brian Garrett flags three anomalies: skepticism surges every time the S&P 500 approaches a new high; rally volatility is nearly double sell-off volatility; call implied volatility shows an unusually high realized correlation with spot prices.
In plain terms = some investors are buying "just-in-case" upside insurance rather than adding real long exposure — not the behavior of a market that believes in the rally.
This reflects a market stuck between fear of missing out and fear of losing money.
What could break the stalemate?
Flood argues Americas investors likely underestimate the real adoption pace of Chinese open-source AI models, while Asian investors likely overestimate it — the truth probably lands somewhere in between.
An incoming IPO pipeline could serve as the catalyst that draws both institutions and retail back into the market.
This means → whether today's defensive positioning converts into actual upside — under the twin push of high-profile IPOs and a strong earnings season — will be the key test of Goldman's "excessive caution" call.
市场有风险，内容仅供研究参考，不构成投资建议。