Amazon said it spent $31.4 billion on capital expenditures in the second quarter, up from $24.3 billion three months earlier, and signalled that this pace will hold through the remainder of the year. The guidance lifts the company’s full-year 2025 capex outlook to roughly $118 billion, compared with an implied $105 billion target given at the start of the year. Chief Financial Officer Brian Olsavsky told analysts that Amazon Web Services remains the main driver of the outlays as the company races to add data-center capacity and roll out its own AI accelerators such as the Tranium family of chips. The spending jump comes even as Amazon seeks to narrow a supply shortfall in power and real estate for its cloud infrastructure. Amazon’s decision underscores an industry-wide investment surge. Updated forecasts compiled from recent earnings calls put combined 2025 capital spending by Amazon, Microsoft, Alphabet and Meta above $344 billion, and at least $400 billion in 2026, as the so-called hyperscalers expand fleets of servers and networking gear to train and run generative-AI models. Gartner expects total global data-center investment to reach about $475 billion this year, 42 % more than in 2024. While the scale of the outlays has revived concerns about capital intensity, share-price gains across the group suggest investors largely endorse the view that the spending will translate into faster growth in profitable cloud and AI services.
No quieren quedarse atrás. Grandes tecnológicas disparan el gasto en IA: más de US$344.000 millones en un año https://t.co/PpeFPlTuIJ
💰The money going into capital projects at big tech firms is now massive. Just last quarter, the "Magnificent Seven" (the top 7 tech cos) threw down over $102.5 billion on datacenters. 📈 Big Tech will pour $400B into AI hardware and data-center buildouts in 2025, dwarfing https://t.co/BFzhYAPyQe https://t.co/6f16vbV0xA
This year companies will spend $400bn on the infrastructure needed to run artificial-intelligence models. Predictions of the final bill are uniformly enormous https://t.co/iIZT6SDp74