AI stocks have come under pressure as investors lose patience with endless valuation debates Photo: MarketWatch photo illustration/iStockphoto
Referenced Symbols
Nicholas Colas, co-founder of DataTrek Research, called the current market action evidence of an “AI burnout trade.” In other words, investors are getting sick of projecting payoffs from investments in technology with uncertain, far-off returns — and they’re also tired of debating what that means in the here and now.
“Should an investor in the S&P 500 care about the next 36 months or the next decade? We all know the answer is the latter, but fretting over the former can be (or, frankly, is) exhausting,” Colas wrote.
Investors are also finding it challenging to “cut through the noise surrounding artificial intelligence — from the grand promises of transformative growth and the massive costs of data-center build-outs, to concerns about AI displacing jobs and the steady stream of new chip or model announcements that arrive almost daily,” said Luke Rahbari, the co-portfolio manager of the Rational Equity Armor Fund.
He agreed that “many have grown weary of trying to look years ahead at the ultimate payoff.”
For the last several years, the market has rewarded companies when they announce a new, larger model, a larger data center, or a bigger AI-infrastructure commitment. But that’s changing as investors are demanding clearer proof that capital spending can translate into sustainable revenue growth, StoneX analyst Yi Fu Lee told MarketWatch.
Lately, there has been consternation around reports of financial metrics from leading private AI companies OpenAI and Anthropic. The market can show “little mercy when expectations and reported figures diverge, and AI-related stocks quickly began to show cracks,” Lee added.
Peter Berezin, the chief economist at BCA Research, told MarketWatch that there’s also concern that spending on AI tokens — the units of data that AI models process, and that providers bill for — is slowing. That could pose challenges for AI companies. According to his estimates, AI companies will need $10 trillion in annual sales to generate a “decent rate of return” on their spending on AI data centers. which has become the poster child for AI borrowing needs, are down to their lowest point since July.
