Goldman warns momentum wipeout enters 5th week as AI narrative shifts

A historic unwind of momentum strategies is accelerating into its fifth week, with Goldman Sachs warning that AI and tech positioning remains far from cleared.

Momentum strategies have unwound for five straight weeks as AI-related positioning faces a structural repricing, Goldman Sachs strategist Benny Quek said. The shift marks a rapid transition from fear of missing out to fear of holding, with leveraged exchange-traded funds bearing the brunt of the collapse.

"AI and tech positioning is far from clean, and we expect elevated volatility to persist through the third quarter," Quek, a derivatives strategist at Goldman Sachs, said in a weekend note. The warning comes as multiple macro pressures — geopolitical conflict, oil price swings, interest rate uncertainty, and resurgent tariff risks — compete to dominate market pricing.

The unwind has been most acute in Korea, where leveraged ETF assets halved to roughly $26.5 billion from a $53 billion peak, with leverage exposure as a share of free-float market capitalization dropping to 2.1% from 3.3%. The KOSPI has fallen for five consecutive weeks, a pattern Quek described as an "unsettling parallel" to China's CSI 1000 selloff in 2014-2015. In the U.S., a high-versus-low momentum stock pair trade dropped 8% in a single session last week, underscoring the fragility of crowded positions.

The shift matters because the AI trade's center of gravity is moving from capital expenditure headlines to profitability and return on invested capital. Goldman estimates $489 billion in AI-related bond supply has been issued year to date, with 40% coming from hyperscale cloud operators — equivalent to 1.5 times the full-year 2025 forecast. Credit spreads on those names are widening, raising the cost of the debt that has funded the AI buildout. The S&P 500 has added $31 trillion in market capitalization since December 2022, while Goldman's base-case estimate for AI's total addressable value stands at $9 trillion, suggesting current prices already discount a significant portion of AI's potential upside.

Low Correlation Masks Index-Level Risk

Despite the surge in individual stock and factor-level volatility, implied correlation among S&P 500 components remains near historic lows. Quek said this divergence provides structural support for long volatility positions — if correlation mean-reverts, index-level swings could amplify sharply. Goldman's risk appetite indicator remains elevated, signaling that market sentiment has not yet fully priced in the converging macro risks.

The VIX has yet to reflect the full extent of the dislocation beneath the surface, with single-stock volatility running well above what the index level suggests. That gap, combined with crowded momentum positioning, creates what Quek described as a setup where "the index looks calm but the components are screaming."

Capital Rotates Toward Asia Value

While the momentum unwind has punished high-beta tech names, capital rotation is visible in Asian markets. Indonesia's Jakarta Composite Index has rebounded 16% from its lows, and India recorded its strongest month of foreign inflows in the region. The software sector has outperformed semiconductors, with the ratio of software stocks to semiconductor stocks breaking above its 50-day moving average — a signal that money is rotating out of the most crowded AI trades.

Gold is also drawing attention. CFTC futures positioning is increasing, a metric that historically leads spot prices, while China's June gold imports rose to a two-year high, signaling persistent safe-haven demand.

Quek said the approach for the current environment is a "barbell" strategy — pairing defensive assets with selective offensive positions to navigate a market that is directionally uncertain but structurally volatile. With U.S. midterm elections approaching, historical patterns suggest equity markets tend to trade sideways before elections and strengthen afterward, offering a potential timeline anchor for the current turbulence.

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