- Gartner flagged a widening gap between data center capex and actual enterprise and consumer AI revenue
- The result could be more than $3.5 trillion in impairment charges by 2030
- A major market correction could dramatically reshape the neocloud sector
The AI data center gold rush is looking less like a boom than a balance-sheet time bomb. Gartner Senior Director Analyst Will Sommer warned current capital spending is “unsustainable,” with the industry potentially facing more than $3.5 trillion in impairment charges by 2030 if today’s buildout pace continues – a correction that could also force more than 80% of neoclouds to abandon independent operations.
“The problem isn’t that demand is low," Sommer told Fierce. Instead, earning revenue off that demand is challenged by fragmented existential plays for market share, he said. “If company A puts prices up too high, company B will take that market share. ... Enterprises are having a really hard time getting value out of this stuff and consumers are relatively unwilling to pay for it.”
To come to its conclusion, Gartner examined the capital spending plans announced by providers against its own infrastructure forecasts as well as its forecasts for consumer and enterprise spending, Sommer said.
“When you put all that money together and you compare it against how much is projected to be spent on data centers, the lines don’t cross until, at the earliest, the 2030s,” Sommer said.
Not all revenue generated by AI will go back into data centers, Gartner found. Most AI service providers don’t own data centers so they earn revenue off the compute power that another company is hosting.
All told, Gartner forecasted the total enterprise and consumer AI revenue that can be attributed to AI data centers to reach no more than $3.6 trillion. That number is less than half the $8.5 trillion in cumulative revenue the firm calculated AI data centers would need through 2030 to cover announced spending and expected returns on investment.
Data center spending spree
Gartner isn’t the only one with doubts about where things are headed. Analysts, including Berstein’s Mark Moerdler, have been questioning hyperscalers about the disconnect between spending and revenue on recent earnings calls.
The heightened concerns come toward the end of a year in which data center capex is expected to near $1 trillion dollars.
Dominant providers Amazon, Alphabet and Microsoft have collectively spent nearly $250 billion in the first half of the year alone, and all have said they expect spending to accelerate in the second half. There’s also a long tail of smaller and neocloud providers – including CoreWeave, Nebius, Lambda, Nscale and more – hustling to bring more capacity online.
Interestingly, Sommer noted the average lifespan of the chips inside data centers is expected to be five to six years. That means that for new capacity brought online now, a pricey refresh cycle will hit around the same time Sommer predicted revenue might finally begin catching up to expenses.
What does this mean for the market?
The nearly inevitable result of the current situation is a massive reset, and neoclouds are the providers most exposed if revenue falls short of expenditures. Sommer predicted a wave of consolidation could sweep through the market, leaving “no more than a few” providers of compute and services controlling the majority of the market.
Asked which neoclouds will survive, Sommer said that’s impossible to predict. But he did say sovereign plays could be a savior for some.
“I think that a lot of the neoclouds that do continue as independent operations past the end of the decade will be sovereign-aligned entities,” he said, highlighting local compute initiatives in Europe, the Middle East and Asia. “There are a number of geopolitical areas that are implementing rules, laws or incentives to support a sovereign tech ecosystem.”
Read more about the data center market on Fierce Network
Hyperscalers face growing investor doubts over AI data center spending spree
About 700 data centers are being built in U.S. as public outcry increases
Amazon's AI infrastructure bill just got $20B bigger
Microsoft Azure crosses $100B mark as growth hits four-year high