Morgan Stanley sees multiple factors feeding a slowdown in AI spending: cash is a concern, pricing is accelerating, and speculative risk is rising
Getty Images
AI stocks have been the bread and butter for investors of late, fattening the portfolios of high school teachers and massive hedge funds alike. But there are signs that the mania over generative artificial intelligence is moderating, especially on Wall Street.
The bellwether S&P 500 RavenPack AI Sentiment Index is only up 6% so far this year, well behind the S&P 500 Index, which is humming along at a 14.2% clip in 2025. That recent performance seems out of line for the AI sector.
Since ChatGPT was rolled out in November 2022, AI data center ecosystem stocks have accounted for roughly 75% of S&P 500 returns, 80% of earnings growth, and 90% of capital expenditures growth, according to data from. Perhaps not so coincidentally, since that date, the S&P 500 has gained a robust 90%, which is why the investment giant refers to AI and high stock returns over the past three years as a “one-note narrative.”
Join 500,000+ readers who start their day with Quartz.
By subscribing, you agree to our Terms of Service and Privacy Policy.
Now that narrative is slowing down, and Morgan Stanley says the AI fireworks show is petering out.
“While most bullish outlooks are premised on economic re-acceleration, it’s difficult to ignore the market’s reliance on AI [capital expenditures],” Lisa Shalett, chief investment officer at Morgan Stanley Wealth Management, said in the research note this week. “In market-pricing terms, we believe we’re closer to the seventh inning than the first, and several developments indicate we may be entering the later phases of the boom.”
Why? Morgan Stanley identifies multiple factors contributing to a slowdown in AI spending.
Cash is a concern: Morgan Stanley noted that AI hyperscaler free-cash-flow growth “has turned negative.”
Pricing is accelerating: Price competition in the "monopoly-feeder businesses” appears to be speeding up.
Speculative risk is rising: As a topper, recent deal-making “smacks of speculation and vendor-financing strategies of old,” the report said.
Market experts say AI spending is pausing, but it is not expected to lag over the long haul.
"AI spending isn’t falling off, it’s shifting gears,” said Maja Vujinovic, CEO and co-Founder of Digital Assets FG Nexus. “Last year was a sprint, with tech giants racing to buy chips and build data centers. Now it’s more of a marathon: Companies want efficiency and clear returns on what they’ve already built, so spending growth looks slower.”
That’s why Morgan Stanley is right to wave a “caution flag,” Vujinovic said.
“The big players will keep investing, but smaller AI firms without steady contracts, access to power, or real cash flow could struggle as the easy money phase ends,” Vujinovic said. “For investors, this is no longer about buying every ‘AI stock,’ it’s about picking the few that can prove they’re delivering real results."
Morgan Stanley offers advice on the steps investors should take as AI expenditures approach a breaking point.
“We can try to focus on economic broadening, but this bull still depends on its GenAI foundation,” Shalett said. “Consider selling small-cap, unprofitable tech and low-quality meme stocks, while being highly selective in consumer names.”
Morgan Stanley also advises investors to add “real assets” to their portfolios.
“[That means] gold, REITs, energy infrastructure, and industrial and agricultural commodities, to complement passive U.S. equity benchmark weighting, and favoring selection in large-cap, quality U.S. names,” Shalett wrote.
Taking a breather to assess AI stocks is a good idea right now, other market mavens say.
“Exponential growth stocks like AI today are hard to value,” said Arie Brish, a business professor at St. Edwards University in Austin, Texas, and a former tech executive. “Will they grow 300% or only 200%? This is anyone's guess.”
Historically, inflated stocks such as AI today tend to underdeliver, Brish said.
“That doesn't mean they will not make it. All it means is they’re likely to be overpriced," he said. "So do your homework and look at the value proposition, company management experience, and potential competition.”
