J.P. Morgan: True diversification from AI is hard to find

Gabriela Santos, J.P. Morgan Asset Management’s chief market strategist for the Americas, warned that genuine diversification away from the artificial intelligence trade is difficult because AI-linked capital spending now reaches most asset classes, including equities, fixed income and private markets.

She pointed to a summer unwind in momentum that hit AI-linked stocks most sharply in July and continued into August. “You can be really really bullish AI and still need to think really really carefully about portfolio construction,” Santos said, adding that attention to position sizing, leverage and diversification remains necessary even for investors bullish on an extended earnings cycle.

J.P. Morgan strategists created an AI factor basket to measure how closely different assets track the broader AI trade. The firm’s tests showed many assets moving together, which reduces the number of independent sources of return available to investors.

Santos identified U.S. Treasuries, gold, core real estate and European equities as the areas still providing genuine diversification. Other traditional offsetting assets no longer behaved as reliable diversifiers in the bank’s analysis.

She said the scale of AI-related capital spending has changed market behavior within and across sectors. Companies often grouped together-hyperscalers, chipmakers and software firms-are starting to diverge, with some members of a sector rallying while others lag, making old sector groupings less useful for balanced portfolios.

The cushion bonds provided after the 2008 financial crisis has shifted. In the decade following the crisis, low yields meant fixed income often offered recession protection on its own. Santos noted that competition for capital, supply disruptions, inflationary pressure and rate volatility have returned, weakening the single-asset protection that bonds once offered.

Santos said portfolio responses will need to be flexible as the shape of the AI buildout continues to evolve. She urged investors to reconsider position sizes and leverage and to include a broader set of inflation-resistant holdings alongside traditional bond allocations.

How AI-related capital spending develops through the rest of the year will influence which assets stay correlated with the AI trade and which can offer independent returns, Santos said.