Some of the world’s biggest asset managers are sounding the alarm about risks to the artificial intelligence boom, even as Wall Street’s top CEOs talk up the potential windfall from the market.

Pacific Investment Management Co.’s Chief Executive Officer Emmanuel Roman and Bain Capital’s chair John Connaughton are among a growing number of investment managers warning of bubble risks in AI and an eventual downturn after the sector’s breakneck expansion.

“At some point in time, there’s going to be a downturn, and then things are going to break,” Roman said at the 2026 Global Financial Leaders’ Investment Summit in Hong Kong on Tuesday. A slump could be triggered by anything from rising US interest rates or war in the Middle East to a euro-zone crisis, he said.

They add their names to a growing list of asset managers expressing doubts about AI, although many of these executives also say there are plenty of opportunities for careful investors.

Howard Marks, co-chair of Oaktree Capital Management, said that a hint of skepticism was emerging around AI, which he called a “crowded trade.”Capital Group’s Jody Jonsson said there was a disconnect between what AI companies expect and the actual pace of adoption.

The world’s biggest alternative asset managers and securities firms have been pouring money into AI at an unprecedented pace, betting that surging demand will underpin returns even as they warn the scale of the buildout raises the risk of capital being misallocated. But many of the world’s biggest funds are now on watch for ripples in the market, after enduring a wild year during which the AI stock rally has been punctuated by painful downturns.

Wall Street bank bosses, overwhelmed with potential business from hyperscalers desperate to raise capital, are striking a more bullish tone.

Discussing the technology’s supercycle, Goldman Sachs Group Inc.’s David Solomon said “there will be bumps in the road,” adding that people need to figure out how to manage that. But he added in an interview with Bloomberg Television that capital markets remain active despite short-term volatility, with investors “still focused on the trajectory of AI demand, compute demand and capital markets are open and receiving.”

Morgan Stanley’s chief Ted Pick said at the conference that capital expenditure on AI could reach $10 trillion to $15 trillion. cementing the technology’s standing as the most significant economic transformation since the dawn of the internet. He added he also said there’s “a potential crowding out of the non-AI needs of great companies.”Their bullish comments come as Goldman and Morgan Stanley scored the top roles on Anthropic PBC’s upcoming initial public offering, taking the same positions they had for SpaceX’s record-shattering $86.2 billion listing in June.

“Some of this skepticism today is probably healthy,” said Blackstone Inc. President Jon Gray, also sounding a note of optimism. He added that the broader AI trend points to more investment and substantially greater demand for capital.

“What is happening is very powerful, and I think underestimating the productivity gains that are coming and the change in consumer and businesses, that’s a mistake,” he said.

Pimco, which helped supercharge the AI debt boom with billions of dollars in lending in 2025, cut positions just in time to skirt losses and extract a profit, people familiar with the matter said last month. With a series of well-timed sales, interest rate hedges and a decision to skip later AI bond deals, Pimco has so far managed to duck the billions of dollars in hits that fixed-income investors now face after a flood of debt overwhelmed the market and sent prices tumbling.

“There’s definitely bubbles that that will pop,” Bain’s Connaughton said. “So from our standpoint, we look at the valuations to be very high, but the real question on AI is: Are you going to get a return on that valuation? “The warnings about AI from asset managers are becoming harder to ignore. Franklin Templeton Inc. Chief Executive Officer Jenny Johnson said the AI investment cycle may have overextended, pointing to a structural shift in the economics of the largest technology platforms as the central risk.

“We’re at that have-we-gone-too-far” stage of the cycle, Johnson said. “I think part of the issue, especially with the hyperscalers, is you have taken these big cash-flowing companies that used to be capital-light businesses, and suddenly they’re capex-heavy.”

The rising tide of fear about AI has sapped a historic rally in tech stocks, pushing the Nasdaq 100 into a technical correction after it sank 10% from a peak. Tech giants including Oracle Corp., Meta Platforms Inc. and Microsoft Corp. have all charted a course for multi-billion dollar bond issues, highlighting the huge funding needs in the sector.