Investors Just Pulled $32 Billion From U.S. Stocks. Should You Be Selling, Too?

U.S. equity funds recorded more than $32 billion of outflows in one week, the largest weekly withdrawal in nine months, amid Brent crude above $110 per barrel, the 10-year Treasury yield nearing 5%, and expectations for a potential Fed hike. The article flags a potential AI-spending slowdown as a risk to large-cap earnings forecasts, but argues the flows reflect rotation out of large caps rather than indiscriminate selling, with small-cap and broad-market funds retaining interest. S&P 500 earnings growth is still expected to be well into the double digits this year and next, supporting a constructive long-term stance on U.S. equities despite near-term volatility.

Analysis

The relevant signal is not the absolute equity-fund redemption but the apparent large-cap-to-broader-market rotation. If that persists, cap-weighted S&P exposure faces a relative headwind from duration-sensitive mega-cap technology, while equal-weight and small-cap indices can outperform despite a flat-to-down headline index. The mechanical risk is concentrated: NVDA and other AI infrastructure leaders carry enough index weight that even modest earnings-multiple compression can overwhelm improving breadth.

Over the next 1-3 months, the key transmission channel is real yields rather than fund flows. A sustained 10-year yield near or above 5% raises the discount rate applied to long-duration AI cash flows and pressures hyperscaler capex returns; that would force investors to distinguish revenue-recognition beneficiaries from companies priced on multi-year infrastructure assumptions. NVDA is vulnerable to a double reset if customers defer incremental accelerator orders and its valuation de-rates, though a mere safety-policy discussion is not evidence of an order reduction.

The contrarian case is that passive outflows are supply, not a fundamental signal, and forced selling in liquid mega-caps can create an entry point if earnings estimates remain intact. Broad earnings expectations are insufficient validation: watch 2027 EPS revisions, cloud-provider capex guidance, and semiconductor lead times. A break in real yields or a capex-guidance cut would falsify the rotation-long thesis quickly; conversely, stable estimates plus easing yields would favor re-risking AI leaders.

Key Decisions for Investors

• Express rotation rather than outright equity beta: long RSP / short SPY in equal dollar size for a 1-3 month horizon. Target 3-5% relative outperformance; exit if the 10-year yield falls materially and mega-cap growth reasserts leadership.
• Keep NVDA exposure hedged into the next hyperscaler capex updates: buy 3-6 month NVDA put spreads or reduce cash exposure rather than initiate an unhedged short. The catalyst is an order or capex revision, not AI-safety headlines; abandon the hedge if major customers reaffirm accelerating accelerator spend.
• Add a tactical long IWM versus QQQ only if small-cap relative strength persists through the next inflation and labor releases. The trade requires evidence that yields are stabilizing; renewed inflation upside and higher real yields would disproportionately damage leveraged small-cap balance sheets.
• Set an alert on U.S. 10-year yields above 5% and on upward revisions to inflation expectations. Either condition warrants reducing high-multiple software and AI-infrastructure exposure because multiple compression can precede earnings revisions by several weeks.
• No actionable position in MORN or GETY from this signal alone; neither has a direct, disclosed sensitivity to the flow/AI-capex mechanism. Reassess only on company-specific earnings or guidance data.