July's crowded trade unwind leaves S&P 500 2% from highs

The S&P 500 held within 3.5% of its July range, less than 2% from highs, as Goldman Sachs logged its largest gross exposure cut since late 2022.

"After a genuinely parabolic run in high-velocity trading, a hammer came down through consensus positioning over the past month," Tony Pasquariello, head of Goldman Sachs hedge fund coverage, said. "I lean toward the view that the fever has broken."

The S&P 500's average daily move stayed under 1 percent, while Goldman's flagship momentum basket swung nearly 10 percent daily. The bank's TMT momentum basket was up 145 percent year-to-date on June 22 before suffering its worst drawdown on record, then rebounded 17 percent in a single session. Korean stock leveraged ETF assets fell to $15 billion from $53 billion at the June peak. Fundamental long/short clients' momentum factor leverage dropped to the 28th percentile of the past year's range.

The episode marks a structural shift: the market no longer rewards crowded trades or leverage. With roughly $1 trillion in AI capital expenditure still in the system, the question is whether hyperscalers can convert that spending into returns. Meta offered no clear evidence of AI-driven revenue, while Microsoft showed capex converting into product revenue at scale and Amazon delivered re-accelerating AWS growth with expanding cloud margins.

Index Calm Masks Record Positioning Washout

Global tech exposure saw its largest selloff in more than five years. Goldman's prime brokerage recorded its biggest total exposure reduction since late 2022. Asian fundamental long/short funds, which posted record first-half performance, suffered their worst monthly drawdown in a decade, while the KOSPI surged 18 percent overnight.

Equal-weight S&P 500, low-volatility S&P 500, and the S&P 500 excluding AI names all hit record highs this week, showing how narrow the selloff was. The Nasdaq 100 sits 8 percent below its June peak but remains up 12 percent year-to-date, with the index's price-to-earnings ratio at the low end of its recent range.

AI Trade Shifts From Narrative to Returns

The AI trade's problem has moved from narrative to return on investment. Hyperscaler bond credit spreads narrowed after this week's earnings, a sign that investors see the $1 trillion AI capex cycle as increasingly productive. But differentiation is now sharp: names that can show AI revenue conversion are rewarded, while those that cannot face valuation pressure.

The FOMC's communication style has also shifted toward less transparency, adding friction for long-duration growth stocks most sensitive to long-end Treasury yields. Traders now face fewer explicit signals from the Fed, which raises the cost of positioning in rate-sensitive AI and technology names.

The Nasdaq 100 has fallen in six of the past nine months but is still up 9 percent point-to-point. The bull market remains intact, but the path forward is harder. July demonstrated that the market does not reward crowding and does not forgive leverage.

This article is for informational purposes only and does not constitute investment advice.