JPMorgan’s Santos Says ‘AI Tentacle’ Risk Spans Bonds And Stocks

JPMorgan Asset Management’s Gabriela Santos warned that “AI-factor” concentration risk has spread beyond equities into fixed income, ratcheting up the need for investor caution even as the super-cycle thesis remains intact.

“You can be very bullish on all things AI and still need to think very carefully on portfolio construction,” Santos said Tuesday in a Bloomberg Television interview.

Santos, JPMorgan Asset’s chief market strategist for the Americas, said July’s correction in tech stocks highlighted the risk of crowded positioning. That month saw the Philadelphia Stock Exchange Semiconductor Index tumble 21%, the most since 2008, while South Korea’s Kospi—where chipmakers Samsung Electronics Co. and SK Hynix Inc. account for about half of the benchmark gauge—plunged 22%.

The upheaval underscored the importance of position sizing, leverage and diversification beyond AI, Santos said.

“That’s where it’s getting complicated. Because you can’t just think of traditional factors, or sectors or regions. Or even asset classes, because that AI tentacle is everywhere now,” she added. Treasuries, gold and core real estate are among the few areas offering different return streams, she said.

Santos said the AI buildout—which she estimates at $5.5 trillion in capital expenditures across public and private markets—is “very unique” because it is already visible in corporate profits. At the same time, she said, growth rates will eventually decelerate, and diversifying now is wise even if AI remains the dominant investment theme.

AI data center spending is on track to top over $900 billion in 2026, with projections for 2027 as high as $1.4 trillion, Goldman Sachs Group Inc. strategists estimated in June, underscoring the scale of the cycle that Santos described.

Santos’ concentration warning extends to bond portfolios. Investment-grade issuance has already hit a fourth consecutive record month, Santos said, with companies including Alphabet Inc. issuing 100-year notes. Moreover, multi-asset investors now face AI buildout exposure on both their equity and fixed-income sides simultaneously.

Hyperscalers’ debt offerings should be studied individually rather than being treated as a monolith, she said, citing the growing complexity of special-purpose vehicles backed by data-center leases as collateral.

(This story was produced with the assistance of Bloomberg Automation.)

This article was provided by Bloomberg News.