J.P. Morgan's AI Diversification Warning Has a Catch

If you're an investor who thinks you've hedged your AI exposure, J.P. Morgan's chief market strategist for the Americas has some uncomfortable news for you. Gabriela Santos said on CNBC this week that true diversification from the artificial intelligence trade is now incredibly hard to find. Her shortlist of actual hedges? Treasuries, gold, and Europe. That's it.

The timing makes sense. We're more than a year into an AI capex buildout that has reshaped how money moves through global markets, and the concentration risk is getting harder to ignore.

The Timeline

Let's back up a bit. The AI trade started as a chipmaker story in early 2023, then morphed into a data center story, and by 2025 it had become something close to a market-wide phenomenon. Nvidia alone added roughly $2 trillion in market value during that stretch. Microsoft, Google, Amazon, and Meta collectively poured over $200 billion into AI infrastructure last year, with projections for 2026 pointing even higher.

Santos's warning, delivered on February 22, 2026, isn't that AI stocks will crash. Her point is more subtle and arguably more troubling. The capex buildout has grown so massive that its effects now ripple through nearly every asset class. Equities, fixed income, private markets, even commercial real estate in certain tech-heavy metros. Nothing operates independently of AI anymore.

So when investors think they're diversifying by buying a broad index fund or a private credit vehicle, they're often just buying AI exposure in a different wrapper. The correlation numbers back this up. During last year's AI-driven selloff in August, supposedly defensive sectors dropped almost as much as tech did. That's not diversification. That's just paying different fees for the same bet.

The Impact

Here's the thing, Santos says the only true diversifiers left are boring ones. US Treasuries, which move on interest rates rather than chip shipments. Gold, which has its own geopolitical and monetary drivers. And European equities, which are largely trading on a different economic cycle than the US.

That's a hard sell for investors who've watched US tech dominate global returns for two straight years. But it's also a necessary reality check. The S&P 500's top 10 holdings now make up over 35% of the index, a concentration level not seen since the 1960s. And a huge chunk of that weight is directly tied to AI optimism.

I'm not entirely convinced gold is the answer here, honestly. It's up roughly 40% since early 2025, and chasing it after that run feels risky. But the logic is sound: if AI stocks correct because of a growth scare, gold and Treasuries tend to rally on rate cut expectations. Europe is the more interesting call, especially with the Euro STOXX 50 trading at a meaningful discount to US peers despite solid earnings.

What changed this week is that a major Wall Street player finally said the quiet part out loud. Investors can't assume their diversified portfolios are actually diversified. The question worth asking: how many people are truly positioned for an AI sentiment shift?

The Outlook

So what does this mean going forward? Watch the next few earnings cycles closely. If the hyperscalers report even slightly disappointing AI revenue growth, the ripple effects won't stay contained to tech stocks. They'll hit every corner of the market that's been touched by the AI buildout narrative, which is most of it.

Nvidia reports again in late May, and that's the next obvious flashpoint. But the bigger signal to watch is the bond market, specifically the 10-year Treasury yield. If it starts dropping meaningfully while equities stay elevated, that tells you the AI trade is losing its grip as the primary market driver.

Santos's advice, in practical terms, is that investors should consider tilting new money toward Treasuries, gold, and European equities rather than assuming their current mix provides protection. That doesn't mean abandoning AI stocks entirely, but it means acknowledging the honest limits of diversification in this market.

Color me skeptical that many investors will actually rebalance this aggressively. US equities have rewarded concentration for too long, and behavioral inertia is powerful. But the smart money starts positioning before the crowd does.

Time will tell, though, whether Santos's caution reads as prescient or premature. Either way, she's given investors something more valuable than a stock pick. She's given them a warning they can actually act on.