AI bubble fears are spiking among professional investors.
Bank of America's latest Global Fund Managers Survey polled 166 managers overseeing $400 billion in assets. A third named an AI equity bubble as the top market threat. That's up from just one in 10 a month earlier.
It's the first time in the survey's history that an AI bubble ranked as the biggest risk.
More than 50% already believe AI stocks are in a bubble. A record 60% said global stocks are overvalued overall. Just a month earlier, only 41% thought stocks were in bubble territory.
Those are huge jumps in concern over a short period.
Here's the twist: fund managers are worried, but they're not backing off.
Bank of America strategists noted that "positioning shows investors see returns outweighing risks." Fund managers are the most net overweight on stocks since January - before Trump's tariff announcements sparked major volatility.
Translation? They think AI stocks are probably overvalued and potentially in a bubble. They're buying them anyway.
The fear of missing out is winning over caution.
Emerging markets became the preferred region for global equity investors in October. Fund managers also stayed overweight on European stocks.
In Europe, banks were the most overweighted sector. Just 4% of those polled see downside ahead for banking, even after huge first-half gains.
Bank of America's strategists summed it up: "AI-bubble tail risks [are] not enough to dampen optimism."
Concerns about an AI bubble rose to the number one market risk, "but not by enough to dampen a more positive equity market outlook overall."
Lewis Grant, senior portfolio manager at Federated Hermes, explained what's really happening.
This year's stock rally "has been primarily sentiment driven, with fundamentals an afterthought," Grant told CNBC.
"FOMO challenges the resolve of even the most disciplined, and there are plenty of reasons to argue that this time is different."
He pointed out that AI is genuinely "era defining" and the rally is led by "established, well capitalised, mega-cap companies." The story is compelling.
But Grant added a warning: "Fundamentals and valuations can only be ignored for so long."
This survey captures a dangerous moment in markets. Professional investors - people who manage billions - are saying out loud that stocks look overvalued and AI might be in a bubble.
Then they're buying anyway.
That's classic late-cycle behavior. Nobody wants to be the one who sold too early and missed the last leg of gains. So they stay in, even while acknowledging the risks.
The shift from 10% calling AI a bubble risk in September to 33% in October is dramatic. That's a tripling of concern in one month. Yet positioning got more bullish, not less.
This pattern has happened before. In 1999, many investors knew dot-com valuations were insane. They kept buying because stocks kept going up. In 2007, concerns about housing and financial sector leverage were widespread. Markets kept rising until they didn't.
When professional money managers tell you they see bubble risks but are staying overweight on stocks because "returns outweigh risks," that's a flashing yellow light.
It doesn't mean the bubble pops tomorrow. These situations can persist for months or even years. But the longer they go, the more vulnerable markets become to a trigger event.
Fund managers are in a tough spot. They're paid to generate returns. If AI stocks keep climbing and they're underweight, they underperform and lose clients. If they stay invested and the bubble pops, they lose money but so does everyone else.
Career risk makes staying invested the safer choice, even when valuations look stretched.
Grant's comment about fundamentals being an afterthought is key. Markets driven by sentiment rather than earnings growth are fragile. When sentiment shifts, there's no fundamental support to catch the fall.
AI is transformative technology. That part isn't hype. But transformative technology can still be overvalued. The internet changed the world, but many 1999-era internet stocks still went to zero.
The survey shows professional investors are increasingly nervous. A third now call AI the biggest risk. Over half think AI stocks are in a bubble. A record 60% say stocks overall are overvalued.
Yet they're buying more aggressively than any time since January. That disconnect between perception and action is what makes markets dangerous right now.
For everyday investors, this is important information. When the people managing $400 billion tell you they see bubble risks but are too afraid of missing out to reduce exposure, you're getting a warning and a roadmap of how things could unwind.
The optimism might be justified. AI might deliver returns that make today's valuations look cheap in hindsight. But when FOMO is overpowering discipline among professional managers, individual investors should be extra careful about chasing momentum.