By Britney Nguyen
‘Be careful.’ That’s one Wall Street pro’s advice as AI spurs a wave of debt offerings and consumer backlash.
If OpenAI stumbles, Baird’s Ted Mortonson says it will impact the market beyond the tech sector.
Almost $400 billion is being poured into artificial intelligence by technology giants this year, with hundreds of billions more on the way next year.
But the fate of the AI trade increasingly hinges on just one company, according to a Wall Street veteran, who sees possible pain ahead if cracks start to emerge.
That company is OpenAI, which was at the center of new multibillion-dollar agreements this past week, further solidifying its importance in the current AI boom that it arguably kickstarted with the release of ChatGPT almost three years ago. But while the deals signal confidence in the AI trade, some on Wall Street are worried that a debt bubble could form as technology companies feel pressured to spend or risk falling behind.
Oracle Corp. (ORCL) just issued $18 billion in bonds to fund its data-center buildout after striking a recent deal with OpenAI worth a reported $300 billion. And CoreWeave Inc. (CRWV), which on Thursday said it was expanding its agreement to provide OpenAI with AI infrastructure, is already a highly levered AI play, with a ratio of 5.1x net debt to earnings before interest, taxes, depreciation and amortization.
See more: Oracle’s stock surges toward its best day since 1999. These huge numbers explain why.
“They need money to keep this show rolling,” Ted Mortonson, managing director at Baird, said of OpenAI – adding that the business is “dramatically” showing negative free cash flow, while at the same time spending on things like its $6.5 billion acquisition of iPhone designer Jony Ive’s startup, io, which was announced in May.
Meanwhile, companies like Alphabet Inc. (GOOGL) (GOOG) and Microsoft Corp. (MSFT) were once “free-cash-flow machines,” according to Mortonson, but are now seeing degradation on this key metric as they intensify spending to stay ahead in the AI race. Altogether, the cloud titans are looking to spend about $380 billion on AI this year, and Mortonson sees that number going to $700 billion next year – or perhaps even more.
The only way to spend that much, he told MarketWatch, is through issuing debt. And that’s a potentially risky proposition, since so much of the AI boom is riding on OpenAI’s ability to keep spending and succeeding.
“The kingpin in all this is OpenAI, and if OpenAI stumbles, this AI trade is going to stumble,” said Mortonson, who’s spent more than 30 years at major sell-side firms and has parlayed his military and engineering backgrounds into a role advising clients on the technology market.
He called out Oracle’s $18 billion sale of investment-grade bonds earlier this week, which he described as “creative” given the inclusion of 40-year maturities despite the unpredictability of the tech industry. Clients, he said, are worried about the possibility of a “debt bubble” given the potentially exuberant assumptions underpinning recent deals.
Adding to the risk around OpenAI is that the company’s rise has had ripples beyond the tech sector – with Mortonson pointing to companies like Caterpillar Inc. (CAT) and Cummins Inc. (CMI) in the manufacturing and utilities industries that have been able to benefit from the AI wave.
“Let’s say OpenAI pauses on debt or can’t get debt financing,” Mortonson said. “The whole tentacles across the market start to atrophy.”
See more: Why Nvidia’s $100 billion investment in OpenAI signals a major transformation
Expectations are high, he added, especially with continuous press releases about AI data-center buildouts worth billions of dollars. Those releases are moving markets, but there’s uncertainty about how quickly revenue will materialize.
“Show me the money and show me the data center is what’s key,” Mortonson said. “It’s a race of ‘time-to-power,’” he noted, referring to the time it takes for the power grid to be connected to and start delivering electricity to a new facility. “It’s a race to get power hooked up.”
For now, if the companies don’t build, “they’re screwed,” Mortonson said – and that’s fueled the need to spend, even if it means issuing debt.
On the Gartner curve – a popular industry model that visualizes how technologies mature and get adopted over time – Mortonson said the tech industry is currently at the “peak of inflated expectations,” where “we’re priced to perfection and nobody’s looking around the corner at any of the downside.”
At some point, the generative AI trade will show a return on investment, Mortonson added – but that’s two to three years out.
‘Be careful’
Asked for the advice he’s giving investors now, Mortonson offered two words: “Be careful.”
Debt is just one of the four reasons why he’s cautious.
Another is that average utility prices are increasing across the country, meaning “consumers are getting whaled,” he said. At the same time, the rates of data-center buildouts “are historic,” and he sees that potentially making pressure on disposable income even worse. That’s created some consumer pushback to AI from people who “didn’t ask for” autonomous driving, crypto or generative AI, but who see rising household utility bills driven by the AI power crunch.
Bernstein analysts made a similar point in a Friday note about the U.S. power grid and the potential for it to fail as demand for power, which they said had been “flat” for a decade, is starting to inflect. The Bernstein team said they suspect consumers will soon start blaming electricity prices on utility companies, hyperscalers and the government. Both Mortonson and the Bernstein analysts also noted that U.S. politicians are taking notice of consumers’ concerns about costs for gas and power.
Wall Street does not know how to factor this blowback into expectations, Mortonson added.
See more: AI spending at Microsoft and Oracle is even higher than it looks, thanks to this accounting maneuver
He’s also worried about potential component shortages that could be a gating factor for AI advancements.
“Nvidia is pushing high-performance compute architecture like no company ever in history has pushed it,” Mortonson said. “So these are technologies that are just being mainstream right now, and the supply chain’s not ready for it.”
And then there’s what he perceives to be “friction” between how institutional clients are treating the current market and what retail traders are doing. His clients are worried about a pause and a correction going forward, and some are selling winners. But the market still seems to be functioning somewhat like a casino, in his view, with retail traders looking to profit by piling into stocks with heavy short interest.
“The market won’t stabilize until these retail investors that are kind of day traders are wiped out,” Mortonson said.
-Britney Nguyen
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
09-27-25 0823ET
Copyright (c) 2025 Dow Jones & Company, Inc.