The AI boom is no longer just a software story. It is also a construction, finance, electricity, and labor story, because the systems behind modern artificial intelligence require vast data center infrastructure.
Microsoft, Alphabet, Meta, and Amazon reported that their 2025 capital expenditures would total roughly $370 billion, with expectations that spending will keep rising in 2026. That scale is already changing how investors read the market, how utilities plan for power, and how companies decide where to put capital.
AI spending is becoming an economic engine
The largest technology companies are pouring money into AI data centers at a pace rarely seen around a single technology. Microsoft was the biggest spender last quarter, putting nearly $35 billion into data centers and other investments. That was equivalent to 45 percent of its revenue.
Warnings about an AI bubble are growing louder, but the immediate impact is already visible. Harvard economist Jason Furman estimates that investment in data centers and software processing technology accounted for nearly all of US GDP growth in the first half of 2025.
That does not mean every part of the economy is benefiting equally. It means that a large share of recent momentum is tied to one concentrated buildout: the infrastructure needed to train and run AI systems.
Public markets are leaning on AI
The US stock market has been strongly shaped by AI since ChatGPT launched in November 2022. According to JPMorgan’s Michael Cembalest, AI-related stocks have accounted for 75 percent of S&P 500 returns and 80 percent of earnings growth.
At the start of this year, large technology companies were mostly funding their AI projects with cash they already had. Financial journalist Derek Thompson pointed out that the ten largest US public companies began 2025 with historically high free cash flow margins. Their core businesses were producing enough money to fund Nvidia GPUs and data center buildouts.
That pattern has continued in important ways. Alphabet told investors last week that its capital expenditures this year would be as much as $93 billion, up from a previous estimate of $75 billion. It also reported that revenue was up 33 percent year over year.
Still, high revenue does not remove every risk. A significant portion of AI investment flows to Nvidia, which releases new versions of its GPUs approximately every two years. Microsoft and Alphabet are currently estimating that their chips will last six years. If they need to upgrade earlier to remain competitive, profits could come under pressure.
Some companies are also seeking new funding channels. Meta recently announced a $27 billion deal to develop a cluster of data centers in Louisiana through a special purpose vehicle. Last week, Meta said it also raised another $30 billion in debt by selling corporate bonds.
The energy problem is getting harder to ignore
AI data centers are power-intensive because they can contain tens of thousands of GPUs. Those chips can collectively perform trillions of operations during an AI training run, and that computing creates heat that must be managed through cooling systems.
The result is intense pressure on the US energy grid. The issue is not only that data centers use power. It is that the US is not building enough grid capacity to support all of the facilities now being built.
Zachary Krause, an energy analyst at East Daley Analytics who has studied the data center industry, warned that some facilities could be completed with computing equipment in place but without enough electricity available to run them.
When supply cannot keep up with demand, prices rise. In the first half of 2025, American utilities sought nearly $30 billion in rate increases, according to The New Yorker. Communities near data centers are especially exposed to that pressure.
The comparison with China also shows the scale of the challenge. Last year, the US deployed 49 GW of renewable energy infrastructure, according to the American Clean Power Association. China added 429 GW. The Chinese government is also reportedly offering generous energy subsidies to domestic tech giants like ByteDance and Alibaba.
OpenAI has warned the White House that limits on US electricity generation for AI development threaten the country’s ability to maintain its global lead in artificial intelligence.
Jobs are being reshaped by capital choices
The data center boom is arriving as the labor market softens. Private employers in the US added just 42,000 jobs in October, mostly in education and healthcare, according to ADP.
At the same time, major technology companies have been cutting workers while reporting record profits. Amazon announced last week that it would eliminate 14,000 corporate roles, with more cuts expected soon. Microsoft laid off about 15,000 people during two rounds of cuts in May and July.
It is tempting to describe this simply as AI replacing workers, but the source article presents a more complicated picture. There is some evidence that generative AI is eliminating entry-level roles in certain industries, including software engineering. Companies are also looking for ways to automate tasks that people currently perform.
Amazon estimated that it could avoid hiring 160,000 people in the US by 2027 by relying on robots, according to internal documents reviewed by The New York Times.
For now, however, the broader economic effect may be tied less to AI tools themselves and more to the infrastructure race behind them. Large companies and investors have limited capital each year. When more of that money goes toward data centers, less may flow to other areas. Manufacturing lost 3,000 jobs last month, according to ADP.
What the boom really signals
The AI data center boom shows how quickly a digital technology can become a physical economic force. It affects stock performance, corporate debt, energy prices, grid planning, and hiring decisions.
The biggest question is not only whether AI spending can keep rising. It is whether the surrounding economy can absorb the strain. The companies building the infrastructure are still earning heavily, but the power system, public markets, and labor market are already feeling the weight of the buildout.