Momentum Strategies Collapse as AI Focus Shifts to ROI

Woofun AI reports that Goldman Sachs trader Benny Quek identifies a historic breakdown in momentum strategies, driven by an unresolved liquidation of positions in AI and technology sectors. This structural failure ensures that high volatility will persist through the third quarter, as market participants struggle to adjust to shifting narratives.

The deeper driver is a rapid shift in market sentiment, moving from a "fear of missing out" to a "fear of holding." This psychological pivot has occurred alongside five consecutive weeks of declines in the momentum factor. Structural pressures are intensifying, with geopolitical tensions, resurgent tariff risks, and fluctuating interest rates converging to disrupt pricing.

Notably, credit spreads for hyperscalable cloud providers have widened sharply, signaling increased stress on the entities underpinning the AI infrastructure boom.

Monitored by Woofun AI, the collapse of leveraged ETFs in the South Korean market illustrates the severity of this trend. The scale of these instruments has been cut in half, dropping from a peak of $53 billion to approximately $26.5 billion. Consequently, leverage exposure as a percentage of the market’s free float market cap fell from 3.3% at its peak to 2.1%. Daily rebalancing volumes for key tech giants Samsung and Hynix also plummeted, declining from 40% and 26% of their respective average daily trading volumes over the past month to just 15% and 14%. The KOSPI index has mirrored this downturn, falling for five consecutive weeks, a trajectory that draws unsettling comparisons to the trend of China’s CSI1000 index between 2014 and 2015.

Structurally, the U.S. market exhibits similar vulnerabilities. A hedge portfolio constructed from high-momentum and low-momentum stocks suffered an 8% decline on Friday alone last week. Although the portfolio still posted a 4% gain for the week, the magnitude of the single-day drop reveals the fragility of these positions. This volatility underscores the difficulty of maintaining directional bets in an environment where momentum factors are systematically failing.

A more critical variable is the fundamental shift in the AI narrative, which has moved from capital expenditure to profitability metrics. Market attention is now focused on profit margins, return on investment, and rising debt issuance levels rather than sheer spending power. Goldman Sachs estimates that bond supply related to AI since the beginning of the year has reached $489 billion. Of this total, 40% originates from hyperscalable cloud providers, a volume equivalent to 1.5 times the full-year forecast for 2025. This surge in debt issuance is accompanied by rapidly widening credit spreads, making the pressure of rising debt financing costs inevitable for these providers.

From a valuation perspective, the disconnect between market capitalization and potential value is stark. The S&P 500’s market cap has increased by $31 trillion since December 2022. In contrast, Goldman Sachs’ estimates for the potential value of AI under baseline, optimistic, and best-case scenarios are $9 trillion, $8 trillion, and $28 trillion, respectively. This discrepancy suggests that the current expansion in market cap has already largely exhausted AI’s potential upside, leaving little room for further multiple expansion without significant earnings growth.

Positions in AI, technology, and momentum strategies are far from being cleared, according to Quek, who notes they have not reached the "clean level" expected by the market. He favors adopting a "dumbbell strategy" to navigate this environment, allocating capital to both defensive assets and carefully selected aggressive positions. Technically, while volatility at the individual stock and factor levels has surged, the implied correlation among S&P 500 components remains extremely low. This divergence provides a structural reason to bet on stock index volatility, as low correlation means index-level volatility is currently underestimated. Goldman Sachs’ risk-on indicator remains high, indicating that overall market sentiment has not fully reflected these risks.

Rotation and divergence are creating structural opportunities in Asian markets despite the broader stress. The Indonesia Composite Index (JCI) has rallied 16% from its lows, while India has recorded the strongest monthly foreign capital inflows in the region. Funds are shifting from high-momentum, high-valuation sectors to more defensive and value-oriented assets. From a sectoral perspective, the valuation ratio of the software sector relative to semiconductors has broken above the 50-day moving average, indicating that internal structural rebalancing is still underway.

Gold signals further highlight the risk-averse turn in the market. CFTC gold futures holdings are increasing, a metric that historically leads spot gold prices.

Meanwhile, China’s gold imports in June rose to a two-year high, underscoring the potential support from risk-aversion demand. Quek also predicts that more headline-driven volatility will emerge as the U.S. midterms approach. Historical data shows that U.S. stock markets tend to consolidate before midterms and rally after they conclude, a pattern that may provide a temporary timeline for the currently stressed market.

Global stock markets closed flat last week, but this figure masks intense internal fluctuations. The market previously ignored ongoing geopolitical conflicts, oil price trends, and interest rate changes, but these three forces are now all entering the market’s focus, competing for pricing power. At the same time, the tariff issue has resurfaced, adding new uncertainty to already fragile market sentiment. Quek points out that volatility this year has been extremely taxing for investors, and the market is in a complex phase marked by overlapping narratives and difficult-to-predict trends.