Quantity The capital flowing into AI data center projects is staggering. Last week, Microsoft, Alphabet, Meta and Amazon said their 2025 capital expenditures would be around $370 billion, and they expect that number to continue rising in 2026. Microsoft was the biggest spender last quarter, committing nearly $35 billion to data centers and other investments, accounting for 45 percent of its revenue.
Rarely, if ever, has a single technology consumed so much money in such a miniature period of time. Warnings of an AI bubble are growing louder by the day, but whether or not it eventually crashes, the madness is already reshaping the U.S. economy. Harvard economist Jason Furman estimates that investments in data centers and software processing technologies have contributed to this. almost all of the US GDP growth in the first half of 2025.
Today we look at the impact of data centers on three key areas: public markets, jobs and energy.
Paycheck
The US stock market is booming, mainly thanks to artificial intelligence. Since ChatGPT launched in November 2022, AI-related stocks have been included 75 percent of the returns of the S&P 500 and 80% profit growth, says JPMorgan’s Michael Cembalest. The question now is whether this growth will be sustainable, given that tech companies continue to spend heavily on AI infrastructure.
Earlier this year, tech giants were mostly financing their AI projects with cash they had on hand. As financial journalist Derek Thompson he noticedAmerica’s ten largest public companies started 2025 with historically high free cash flow margins. In other words, their business was so profitable that they had billions of dollars to spend on Nvidia GPUs and building data centers.
This trend has largely continued throughout 2025. Alphabet, for example, told investors last week that its capital expenditures this year will be as much as $93 billion, up from the previous estimate of $75 billion. But it also said revenue was up 33 percent year-on-year. In other words, Silicon Valley both spends more and earns more. That means everything is fine, right?
Not exactly. First, tech giants seem to be using them accounting tricks to make their finances look better than they actually are. Much of the AI investment goes to Nvidia, which releases recent versions of its GPUs roughly every two years. However, companies like Microsoft and Alphabet now estimate that their chips will last six years. If they need to upgrade sooner to stay competitive – which is likely – it could eat into their profits and undermine their overall performance.