J.P. Morgan's Santos Names Only 4 True AI-Trade Diversifiers, Leaving Out Bitcoin (BTC)

AI Capex Reaches Every Asset Class

Gabriela Santos, chief market strategist for the Americas at J.P. Morgan Asset Management, warned this week that real diversification away from the artificial intelligence trade has become scarce. Speaking on CNBC's “Closing Bell Overtime,” Santos argued that the AI capital-expenditure buildout has grown so large that its footprint now extends across nearly every asset class — public equities, fixed income and private markets alike. Her core message for allocators was that conviction in AI is not, by itself, a portfolio plan. “You can be really really bullish AI and still need to think really really carefully about portfolio construction,” she said, framing the caution as a matter of sizing and structure rather than direction. The warning follows a rough stretch for AI-linked equities: momentum in the trade began to fray in July and the selling carried through August, handing bulls a hard lesson in how crowded positioning behaves when sentiment turns. In our read of her remarks, the point is not that the AI earnings cycle is broken, but that concentration risk has migrated into every corner of a portfolio — which is why she pressed investors to watch position sizing, leverage and diversification even if they expect AI to keep powering an extended earnings cycle. Santos also flagged a structural problem for anyone relying on classic sector buckets. The AI buildout keeps changing shape, and hyperscalers, chipmakers and software companies — think AI-exposed names like CrowdStrike (CRWD) — increasingly diverge within their own groups instead of trading as a single block. A narrow sector proxy such as the XLE ETF already says little about what an AI-heavy fund actually owns, and that blurring makes top-down allocation frameworks less reliable than they were a cycle ago. Wall Street has echoed the concern, with at least one prominent investor describing the current market as effectively a single AI trade.

Four Assets Still Diversify

The mechanics behind the warning come from the bank's own research work. J.P. Morgan constructed an AI factor basket — a model portfolio designed to measure how tightly different assets and client portfolios track the broader AI trade — and Santos said the results show most assets now moving in the same direction at the same time. On that reading, the list of things that actually diversify has shrunk to a handful: US treasuries, gold, core real estate and European equities. Notably absent from that list is Bitcoin (BTC), the asset many allocators have pitched as a hedge against exactly this kind of concentrated equity exposure. Historically, the math was friendlier. Bonds reliably cushioned portfolios through recessions, and for roughly two decades after the global financial crisis, persistently low yields meant fixed income alone could do the hedging work. Santos argued that regime is over: competition for capital has returned alongside supply shocks, inflation and interest-rate volatility, which means investors now need inflation-resistant assets beyond bonds to complete their positioning. Whether the treasuries-gold-real-estate-Europe mix keeps working, she added, may hinge on how AI-related capital spending evolves over the remainder of the year — a variable she did not claim to forecast with precision. For crypto-adjacent allocators, the omission cuts both ways. Corporate vehicles that hold digital assets on balance sheets — the Ethereum treasury model among them — have marketed themselves as uncorrelated ballast, yet the summer unwind tested precisely that pitch. The open question is whether genuine non-correlation still exists anywhere outside the four asset classes she named. Readers tracking the market in real time can follow live spot and futures prices on Binance.

Bitcoin's Hedge Pitch Under Test

The thread tying Santos's warning to this market is a correlation question. From the ICO boom onward, digital assets have been marketed on uncorrelated-return promises, yet Bitcoin (BTC) — trading near $78,000 at the time of writing — is absent from the only four-asset diversification list a major bank's strategist now endorses. For crypto to claim the hedge demand that treasuries, gold, core real estate and European equities cannot absorb alone, BTC and ETH would need to hold up precisely when AI-heavy portfolios de-risk. Whether AI-related capital spending cools or extends through the rest of the year will arbitrate that claim — on Wall Street's terms and crypto's alike.