By Mackenzie Tatananni
The artificial-intelligence bubble. It's a buzzy yet ambiguous term, and its very lack of definition means it might already appear deflated, depending on the metrics used to evaluate it.
That's according to Capital Economics, an independent research firm based in London. If the presence of an AI bubble is determined solely by the relative valuation of the market's technology sectors, "then any such bubble in the S&P 500 has already burst," John Higgins, the firm's chief market economist, argued on Wednesday.
Higgins noted that the premium attached to the valuation of the information technology sector has plunged since the end of last October. Before then, it had tended to grow in recent years, although not as quickly or dramatically as it did in the dot-com bubble of yesteryear.
His observation comes amid predictions that capital-intensity metrics will outstrip those from the dot-com era, as noted by Morgan Stanley analyst Todd Castagno last month.
Castagno expects hyperscalers' cash capital expenditure-to-sales ratios to reach roughly 34% this year and 37% in 2028, compared with 32% recorded around the turn of the 21st century. The magnitude of AI infrastructure spending have resulted in modeling challenges, Castagno added.
Still, Higgins contended Wednesday that forward price-to-earnings ratios for the IT sector were "now the smallest since the pandemic." While it's tempting to attribute this to recent weakness in the software and services industry group, which has been rocked by AI disruption fears, Higgins noted that the forward P/E for the semiconductors and semi equipment group and technology hardware and equipment group also have fallen.
However, there's a big disclaimer. Even if a bubble already has burst, "that doesn't preclude a big correction in the index if corporate earnings falter," Higgins wrote.
It's possible the risk has shifted to earnings rather than stock prices. "There may be a bubble in the earnings of the semiconductors and semiconductor equipment industry group, even if there isn't one in the price that investors are willing to pay for them," Higgins wrote.
If AI demand turns out weaker than expected, or if the war in Iran continues to the point where it disrupts trade, lofty profit predictions likely won't come true. If earnings fall, the stock price may crash to keep the ratio in balance. "But, so far, there is no sign of such a retreat," Higgins said.
Other Wall Street professionals have stressed earnings as an important driver. Ben Snider, the recently-appointed U.S. chief equity strategist at Goldman Sachs, noted in an interview last week that most of the S&P 500's return was driven by earnings in 2025.
"We spend a lot of time talking about valuations and bubble risk, but if you look at the long-term trajectory of the S&P 500, almost all of its return has been driven by earnings," Snider told Barron's.
Others on the Street don't deny the existence of a bubble, but argue it remains intact for now. Citi Research's macro strategy team contended in a report earlier this month that U.S. equity markets were, in fact, in a bubble, "but that for now there is enough liquidity to keep it going."
Intensifying return-on-investment concerns or AI disruption risks might be enough to cause it to erupt earlier, analysts added. In Citi's view, the negative wealth effect from bursting would likely generate a recession and lead to aggressive interest rate cuts.
Against this backdrop, the only safe sectors would be defensives -- think healthcare and staples -- or interest rate-sensitive sectors such as utilities and real estate, analysts with the firm argued.
Write to Mackenzie Tatananni at mackenzie.tatananni@barrons.com
This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
March 25, 2026 16:01 ET (20:01 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.