The mighty American consumer may have finally fulfilled its match, and it is a huge rectangular box that houses very few people inside, but a huge nest of servers, storage systems and networking devices. Consumer expenditure in the American service industry is so large that they confuse and represent the mind About two thirds of the gross domestic product. In order to paraphrase the long -time coffee chain Dunkin, America runs with expenses.
But this mighty American consumer has slow seasons, and the summer of 2025 seems to be one of them. There are several interconnected factors, namely, namely, Longest job growth Now much smaller than previously thought and the effects of artificial intelligence on the workforce. But these hulking rectangular boxes, the massive data centers that sprout across the country, develop into a huge magnet for dollars in a way that makes consumers accessible themselves.
Huge technology companies spent so much for data centers in 2025 that their expenses now contribute more to economic growth in the United States than the consumer expenditure that was considered for a long time as the nation Economic engine. If you assume that the expenses for data centers correspond to the AI investments that are defined as capital, which is used for information processing devices and software, the pattern is clear: a lot of money flows into a concentrated area and the result of this is uncertain.
MicrosoftPresent GooglePresent AmazonAnd Meta are the main players who invest on a stolen levels to create and update data centers that support the exponential demand for AI calculation performance, and these four companies alone Prediction of a record 364 billion US dollars From capital investments in 2025. In combination, the so-called great tech giants spent more than $ 100 billion for data center projects over the past three months Calculated by the Wall Street JournalChristopher Mims.
All of these expenses have to affect the economy. Analyst estimates from the Renaissance macro research show that the dollar value has previously contributed to GDP growth by AI data center editions that exceeds the overall effect of all US consumer expenses -for the first time that this has ever occurred.
So far this year, AI Capex, which we define as information processing devices plus software, has expanded more about GDP growth than the expenditure of consumers. pic.twitter.com/d70fx2lxaw– Renmac: Renaissance -Macro research (@renmacllc) July 30, 2025
Or as Rusty Foster, author of the widespread media blog Today in tabsIt exhibits: “Our economy could only be three AI calculus centers in a trench coat.” This is reminiscent of the classic comedic device of several children who wear a long jacket that pretends to be an adult, as is still shown in Netflix Bojack HorsemanWhen “Vincent Adultman”Keep the illusion successful for several appointments with Princess Carolyn. But then the bladder banged or the trench layer came off.
Why does that happen now?
Several forces drive this unprecedented investment wave. The boom in generative AI and advanced major language models – technologies that require large amounts of arithmetic resources – forced tech giant to increase its physical infrastructure quickly. Data from McKinsey Projects between 2025 and 2030 have to invest companies a worldwide a Remarkable 6.7 trillion dollars in new data center capacities to keep up with AI demand.
The AI data center editions have grown with at least 10 times since 2022 Well-known business blogger Paul Kedrosky Estimation that it corresponds to 2% of the total USBIP. “Schatz, Ai Capex eats the economy,” he writes, and argues that Ai Capex is so great that it “influences economic statistics, promotes the economy and approaches the Railroad boom”.
The Torsten Slok from Apollo Global Management without getting involved in the Capex question of the data centers The KI boom exceeded the market value of the Tech boom in the late 90swhich became known as the “dotcom bubble” after speculative mania was planned and a recession was used.
Kedrosky makes a similar point and contrasting investments from the entire financial history, in particular the telecommunications boom from 2020 in connection with the 5G/fibre technology and the railway boom of the 19th century when the United States recorded a transport revolution. “The investment expenditure for AI calculus centers is probably about 20% of the top expenditure for railways as a percentage of GDP, and it is still rising quickly,” writes Kedrosky. “And we have already reached the highlight of telecommunications during the Dotcom bubble for decades.” Noah Smith, a widespread replacement for economics, asks the obvious question: “Will data centers bring the economy to crash?”
The impact on the broader economy
This increase in technical investments had profoundly emaciated consequences. Without the building of AI Data Center, GDP could actually be ill in the face of uncertain macroeconomic conditions. Therefore, the expenses for data centers may have abolished or postponed a recession.
Flooding money into the AI infrastructure are redirected from other sectors, including risk capital, traditional production and even startups with consumers. In contrast to historical infrastructure booms such as those for railways or telecommunications, AI data centers are short-lived, quickly prepared and require continuous hardware upgrades. The thinking of this investment pattern can remain volatile and capital hungry in the coming years.
When AI redefines the industry, the capital flow into the physical backbone of this technology – VAST -CONTRACTION Center – excited old assumptions about what drives America’s economy. The consumer expenses are absolutely still immense, but are not stepped with the extraordinary scale and speed of the investment by tech giant to lead in the AI era. The trajectory indicates that the US economy is not so much shaped by the purchasing power of its population in 2025, but by the relentless arms around AI calculation capacity-a-to-time, technically led growth engine.
For this story, Assets Used generative AI to help with a first draft. An editor checked the accuracy of the information before publication.
