Gartner 称 AI 基建到 2030 年需 8.5 万亿美元收入，缺口近 4.9 万亿

Trendingnow1UnitedHealthcare and Humana are ending Medicare Advantage plans for nearly 1 million people2United Airlines CEO—who made $32 million last year—calls for a national wealth tax and a nearly 100% inheritance tax: 'I would pay them'3No wonder the White House wants Warsh to cut: A 1% Fed rate hike costs the Treasury an extra $95 billion a year in national debt interest1UnitedHealthcare and Humana are ending Medicare Advantage plans for nearly 1 million people2United Airlines CEO—who made $32 million last year—calls for a national wealth tax and a nearly 100% inheritance tax: 'I would pay them'3No wonder the White House wants Warsh to cut: A 1% Fed rate hike costs the Treasury an extra $95 billion a year in national debt interestAIData centersThe AI boom is nearly $5 trillion short of paying for itself, Gartner saysBy Joshua HongJoshua HongNews FellowBy Joshua HongJoshua HongNews FellowOctober 21, 2026, 11:36 AM ETDemonstrators wave signs during a nationwide protest against AI data center expansion in Imperial, California on July 18, 2026. SANDY HUFFAKER / AFP via Getty ImagesAdd Fortune on Google for similar content.

There isn’t an endless stream of revenue for AI as early adopters once thought. The gap between what’s being spent on the AI buildout and the revenue projected to come back is turning into a multitrillion-dollar chasm.Recommended Video

Data-center owners need to generate $8.5 trillion in cumulative revenue by the end of the decade to justify their investments, according to a press release last week by global research and advisory firm Gartner. But the firm forecasts that no more than $3.6 trillion of enterprise and consumer AI revenue can be attributed to those data centers over the same period—leaving a $4.9 trillion gap.

According to a 2026 market study by trade association BOMA, data centers accounted for approximately $41.2 billion—or nearly 46%—of private office construction in the country in 2025. A decade earlier, data centers only represented less than 5%.

Gartner warns that more than $3.5 trillion of AI infrastructure spending could be exposed to impairment—meaning it would have to be written down—by the end of 2030 if AI providers keep spending at their current pace. Further out, the firm says more than $4 trillion in asset value could face write-down risk through 2032.

“In 2026, AI providers are expected to spend nearly $1 trillion on AI infrastructure, and are forecast to invest an additional $6 trillion between 2027 and 2030,” Will Sommer, senior director analyst at Gartner, said in the press release. “However, our research indicates that enterprise and consumer AI revenue will be insufficient to support profitable returns on all planned AI capital expenditures during this period.”

That $8.5 trillion figure assumes data-center owners need to cover announced capital spending and financing costs while also earning a 12% return on invested capital, which Gartner treats as the hurdle rate for the investment. Gartner doesn’t expect that 12% return to be reached until the end of 2032 at the earliest.

The math doesn’t lie

According to the analysis, cumulative global AI capital spending is forecast to reach $7.5 trillion from 2025 through 2030. Additionally, annual AI capital spending is projected to approach $2 trillion by 2030, which represents well over 1% of global GDP.

That proves a stark difference compared to what the investment looks to receive in return. Total revenue from AI software, services and consumer products is only expected to reach $4.4 trillion by 2030—and Gartner estimates that no more than 30% of the revenue will be booked against the underlying capital investments. That means much of the remaining cash is going to companies that are selling the applications, models and other services that are built on the infrastructure investment.

In other words, businesses might be paying for AI-powered software and services, but only a fraction of that will ultimately support the data center and computing chips that are used to deliver its service.

“Product leaders should not view AI capital write-downs as an infrastructure problem alone,” Gartner’s study read. “Even if a product organization owns no AI infrastructure, provider consolidation, price volatility, model discontinuations, service rationalization, and sovereign AI requirements can materially affect product economics, roadmap execution, and customer commitments. Product leaders should therefore design AI-dependent products assuming that today’s AI supply landscape will become materially more concentrated through the end of the decade.”

Not a good look for neoclouds

The research firm also expects the economics to change depending on the provider. Large cloud companies like Amazon, Microsoft and Google parent Alphabet have diversified businesses and established cash flow sources that will aid in their AI investments. But smaller, specialized cloud infrastructure providers—called neoclouds—could be more exposed.

“Neoclouds must soon earn returns on AI-optimized IaaS and associated platforms to sustain operations,” Gartner’s press release read. “By 2030, Gartner predicts that cost pressure and a lack of a differentiating business model will force over 80% of neocloud providers to cease independent operations.” Fortune Daily breaks the traditional barrier between audience and newsroom. The show transforms Fortune’s trusted reporting into actionable, conversational, and entertaining insights for an emerging class of business leaders. Watch here.About the AuthorBy Joshua HongNews Fellow

Joshua Hong is a News Fellow at Fortune covering data, AI, cybersecurity, energy, and retail. See full bioAdd Fortune on Google for similar content.