U.S. technology companies have collectively invested about $155 billion in artificial-intelligence projects since the start of 2025, according to industry estimates cited on Monday. The outlays span new data-center capacity, semiconductor purchases and other infrastructure needed to train and run increasingly complex AI models. The surge in spending is straining the nation’s power grid. Industry research and media reports indicate utilities are struggling to keep pace with demand as electricity requirements for hyperscale data centers climb. To secure long-term supplies, large technology firms are now exploring on-site nuclear solutions, including small modular reactors, in areas that would be impractical for other industrial uses. Economists note the investment wave is contributing to a form of “job-light” expansion, in which gross domestic product and capital spending rise even as traditional employment measures remain subdued. Analysts add that while AI revenues are accelerating, free cash flow at some firms is narrowing as they shift to more asset-heavy operating models to accommodate AI’s energy and hardware needs.