The AI trade isn't unravelling; winners are outpacing losers: JP Morgan AM

J.P. Morgan Asset Management's Raisah Rasid says recent AI-stock pullbacks reflect greater scrutiny of returns on capital, not doubts about the technology's long-term prospects. She favors high-quality investment-grade bonds, especially at the short-to-intermediate end of the yield curve, citing elevated yields as a source of income and a cushion against volatility.

Analysis

The investable distinction is between AI adoption and the returns earned on AI infrastructure: continued spending can support semiconductor and data-center demand while still weighing on equity multiples if monetization, utilization, or cash-flow conversion disappoints. That leaves a wider range of outcomes for AI-exposed equities than for the technology thesis itself. Second-order pressure could reach data-center power and equipment suppliers if customers slow or sequence projects; that is a conditional risk, not evidence of cancellations.

Short- to intermediate-duration investment-grade bonds can provide carry with less rate sensitivity than long-duration credit, but they are not a hedge against every drawdown: a broad risk-off move can widen spreads, while renewed inflation or higher policy-rate expectations can reduce bond returns. The near-term signal is positioning and repricing; over the next 1–3 months, earnings commentary on AI capex, utilization, and monetization matters more than broad adoption claims. Over 6–18 months, evidence that returns justify infrastructure spend should determine whether equity multiples stabilize. The contrarian risk is that investors overreact to a returns debate even as spending translates into durable cash flows. Verify current yields, spreads, and issuer fundamentals before sizing; the article provides no valuation or market-level data.

Key Decisions for Investors

• Consider shifting marginal risk from high-beta AI equity exposure toward short- to intermediate-maturity, higher-quality investment-grade credit, subject to spread and issuer-level checks. Treat this as a volatility and carry allocation, not a direct AI hedge.
• Avoid a broad short of AI equities on this evidence alone. Use upcoming earnings and company guidance to test capex payback, utilization, and monetization; favor reducing exposure where spending rises without corresponding evidence of returns.
• Monitor data-center power and equipment suppliers as a second-order watchlist: project delays or revised customer spending plans would strengthen the downside case, while sustained orders would weaken it.
• Falsify the cautious equity thesis if companies show improving cash-flow conversion and credible returns on AI investment; reassess the bond allocation if investment-grade spreads widen materially or rate expectations move higher enough to erase carry.