Expenditure on AI data centers are so massive that it took a larger part of GDP growth than purchases – and it could crush the American economy

But this powerful American consumer has slow seasons, and summer 2025 seems to be one of them. There are several interconnected factors, namely the recent growth of jobs which now seems much smaller than what did not think it before and the impact of artificial intelligence on the workforce. But these rectangular boxes that have grown, the massive data centers that grow across the country, emerge as a giant magnet for dollars in a way that rivals consumer spending.

Giant technological companies have spent so much in data centers in 2025 that their expenses are now contributing to American economic growth that consumer spending has long considered the country’s economic engine. If you make the reasonable assumption that expenditure in data centers equivalent to the capital expenses of the AI, defined as the capital deployed for equipment and information processing software, the model is clear: a ton of money flows in a concentrated area, and the result is uncertain.

Microsoft, Google, Amazon and Meta are the main players who invest at narcotic levels to build and upgrade the data centers that support the exponential IA computer power demand, with these four companies alone providing for a record of 364 billion dollars in capital investment in 2025. Wall Street Journal‘S Christopher Mims.

All these expenses must have an impact on the economy. The estimates of the Macro -Rearissance macro -research analysts indicate that so far in 2025, the value in dollars has contributed to the growth of GDP by the expenditure of the AI data center has exceeded the total impact of all American consumption expenses – the first time.

Until now this year, AI CAPEX, which we define as information processing equipment and software has added more to GDP growth than consumer expenditure. pic.twitter.com/d70fx2lxaw– Rememac: Renaissance Macro Research (@renMacllc) July 30, 2025

Or like Rusty Foster, author of the media blog widely read Today in tabssaid: “Our economy could be of three AI data centers in a trench.” This recalls the classic comic device of several children wearing a long jacket, pretending to be an adult, as described memorably in Netflix Bojack riderWhen “Vincent Adultman” managed to maintain the illusion of several dates with Princess Carolyn. But then the bubble broke out, or the trench coated.

Why is it going now?

Several forces stimulate this unprecedented wave of investment. The BOOM of the generative AI and advanced models of large language – technologies that require large amounts of IT resources – have forced technology giants to quickly increase their physical infrastructure. McKinsey’s data are planning that between 2025 and 2030, world companies will have to invest 6.7 billions of remarkable dollars in a new capacity in the data center to meet the demand for AI.

Expenditure on the AI data center have increased at least 10 times since 2022, with the The well-known commercial blogger Paul Kedrosky estimates that he approaches 2% of the GDP of the United States in itself. “Honey, Ai Capex eats the economy,” he wrote, arguing that the Capex is so great that it “affects economic statistics, strengthening the economy and begins to approach the railway boom”.

Kedrosky makes a similar point, contrasting Capex booms in all financial history, including the 2020 telecommunications boom linked to 5G / Fiber technology and the 19th century railway boom while the United States has adopted a transport revolution. “Capital spending on AI data centers are probably around 20% of advanced expenses on railways, percentage of GDP, and it is still increasing quickly,” writes Kedrosky. “And we have already exceeded the peak of decades in telecommunications spending during the Dotcom bubble.” Noah Smith, a largely read economy sub-droseman, asks the obvious question: “Will data centers crush the economy?”

This increase in technological investments has had deep consequences downstream. Without the Data Center Building Construction SIR, the GDP could have contract in the face of uncertain macroeconomic conditions. Thus, expenditure of the data center may have distant – or postponed – a recession.

Silver floods in AI infrastructure are diverted from other sectors, including venture capital, traditional manufacturing and even consumer-oriented startups. Unlike booms in historical infrastructure such as those for railways or telecommunications, AI data centers are short -lived, rapid depreciation and require continuous material upgrades – the suggestion of this investment model can remain volatile and hated for years to come.