Roughly $750 billion. That’s how much Meta Platforms Inc., Amazon.com Inc., Microsoft Corp. and Google owner Alphabet Inc. are expected to devote to capital expenditures this year alone.
That’s more than the entire gross domestic product of Ireland or Belgium. More than half of what all of the companies in the S&P 500 Index spent in total last year. Enough to cut a $5,500 check for every household in the US .
There was a time when John McWilliams, part of the research group at real estate services firm Cushman & Wakefield, would’ve found it impossible to imagine four companies spending that much, primarily on data centers, in a single year. “We’ve never seen it,” he said.
McWilliams and his team have since crunched the numbers . The cost of construction materials such as steel and cement has skyrocketed because of trade wars, fuel expenses and geopolitical conflicts. The AI boom itself has driven up expenses for everything from chips to copper. The price of short-term memory used in data centers is more than five times what it was just a year ago, data compiled by Bloomberg Intelligence show. Demand for equipment that generates and carries electricity is so high because of the building frenzy that suppliers have backlogs stretching beyond 2030.
Local opposition and regulatory setbacks are adding to those woes, forcing developers to redraw their plans and resort to costlier alternatives for resources like power. Just last month, Oracle Corp. issued a force majeure notice on a giant New Mexico data center after a gas pipeline that’s key to the project ran into regulatory delays.
Virtually every major category of costs associated with building a data center is ballooning.
“When you think about it in that context, it doesn’t seem as crazy as the number might suggest,” McWilliams said. “These are extremely expensive facilities to develop.”
This explains why trillions of dollars’ worth of investments, debt and jobs now hinge on the AI data center build-out. The unprecedented level of spending has reverberated across all corners of the financial world, flooding everything from private credit to bonds and equity markets. That’s despite high-profile names on Wall Street, including Michael Burry of Big Short fame and Bridgewater Associates founder Ray Dalio, warning of an AI bubble that will eventually burst .
Cushman & Wakefield, which advises clients on data center construction and issues reports on the market, provided Bloomberg with an exclusive breakdown of costs for building the shell of a single-gigawatt AI data center (enough to power roughly 750,000 homes at any given moment) in Abilene, Texas , where some of the world’s largest computing sites are being built. Bloomberg Intelligence developed its own estimates for the costs of the hardware within a gigawatt-sized site. Based on the high end of that analysis, here’s what a developer can expect to pay:
Altogether, building and housing a theoretical AI data center in the Abilene area capable of handling a gigawatt of computing power could cost a developer roughly $30 billion.
Over the next four years, more than 60 gigawatts of this data center capacity is expected to be added in the US alone, tripling current levels in the nation that’s already home to the most computing power in the world.
Add the fact that these build-outs are occurring globally, from Brazil to China to the UK and Europe, and suddenly the estimates that the AI boom will cost trillions of dollars make sense. The unprecedented levels of spending are rippling across entire sectors, supercharging markets for natural gas turbines, power transformers, construction, real estate, hardware, networking equipment and everything in between.
The staggering costs also underscore the high stakes being wagered in this boom. These multibillion-dollar data centers are being built on the assumption that demand for AI systems will continue to grow, and that the chips housed within their walls will pay dividends for years to come. Given some of the world’s most powerful AI model developers have yet to turn a profit, that demand outlook is far from certain.
In warning of a bubble, Burry has pointed to, among other things, the so-called circular deals in which hardware makers such as Nvidia Corp. are directly partnering with and investing in key customers such as OpenAI.
“All the capital being spent and lent by the richest companies on Earth,” Burry said, won’t be enough to keep the bubble from eventually bursting.