AI-linked stocks have suffered another week of selling, with investors embracing a broad risk-off stance on the sector—but not for the reasons most had anticipated.
The Global X Artificial Intelligence & Technology ETF (NYSE:AIQ) has now dropped nearly 10% since its early November peak, putting it on course for its first correction since April.
Concerns about a potential AI bubble have long centered on whether demand would justify the sector's explosive valuations.
But this week, a different picture has emerged.
Wall Street analysts now warn that demand is actually real and too strong, creating overwhelming pressure on physical infrastructure—data centers, transformers, turbines—that simply can't scale fast enough.
CoreWeave Sounds the Alarm: Demand For AI Far Exceeds Capacity
The wake-up call came from CoreWeave Inc‘s (NASDAQ:CRWV) latest earnings update.
Despite surging orders and a ballooning backlog, the company's stock sank 30% this week after admitting it cannot deliver infrastructure quickly enough.
"We continue to operate in a highly supply-constrained environment, where the demand for CoreWeave’s best-in-class AI cloud platform far exceeds available capacity," said CEO Mike Intrator during the company's earnings call.
CoreWeave added $25 billion in new revenue backlog in the third quarter alone. This lifted the total backlog to $55.6 billion—nearly double second-quarter levels and almost four times higher year-to-date.
But a delayed third-party data center buildout has already affected fourth-quarter expectations.
"There is no entity that has the capacity to deliver infrastructure globally to meet the demand," Intrator added. "We're doing everything possible to limit the damage associated with these delays—but the supply chain is overwhelmed."
To keep up, the company said it expects 2026 capital expenditures to be "well in excess of double" those of 2025.
Wall Street Rethinks AI's Next Phase
Jordi Visser, head of AI Macro Nexus at 22V Research, said CoreWeave's update marked a "regime change" for investors.
"The constraint has moved from capital to concrete," Visser said. "I expect a very different year in 2026. CoreWeave was a warning sign of a transition."
Visser highlighted that the AI sector has entered a delivery-constrained phase, where success depends less on innovation or spending and more on execution through supply chain stress.
"When do you hear a company say they have ‘insatiable' demand—and the stock falls 16% that day?"
Despite having the capital, the customers, and the orders, CoreWeave admitted it can't scale fast enough. "This is a systemic problem that the industry will have to deal with for the foreseeable future," management said.
Together, CoreWeave's $55.6 billion in backlogged revenue and Oracle Corp.'s (NYSE:ORCL) $455 billion add up to over $500 billion in contracted AI-related demand.
And that doesn't even include annual capex from giants like Microsoft Corp. (NASDAQ:MSFT), Amazon.com Inc. (NASDAQ:AMZN), Alphabet Inc. (NASDAQ:GOOGL) and Meta Platforms Inc. (NASDAQ:META) each spending $50–100 billion per year on AI infrastructure.
Execution, Not Spending, Will Define AI Winners in 2026
"The next AI investment phase will not be defined by who can spend the most," Visser said.
"It will be defined by who can execute through constraint." With the shift from a capital-constrained to a delivery-constrained environment, Visser believes future performance will be driven by execution risk, infrastructure readiness and project management—not just innovation.
"In 2026, the winners will be those who convert contracts into capacity while preserving margins amid rising costs and tighter timelines."
Investors must now consider whether AI's biggest risk isn't overspending or hype exhaustion—but failing to turn promise into product.
The sector's next phase won't reward the biggest spenders. It will favor those who can deliver under constraint.
Now Read:
Image: Shutterstock
© 2025 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.
