“We continue to increase our investments in AI across both capital and talent to meet the massive opportunity ahead- Satya Nadella, Microsoft’s CEO.
American tech giants announced their spending on AI infrastructure last week and the numbers sound astronomical in every form.
The money is being pumped AI projects believed to hold the key to a profitable future.
Meta, Alphabet and Microsoft are looking to invest billions of dollars in AI over the next year, focused on building chips, data centres and talent.
The tech behemoths see huge opportunities in AI and are willing to stake huge fortunes to build every essential support system.
“We continue to increase our investments in AI across both capital and talent to meet the massive opportunity ahead,” said Satya Nadella, Microsoft’s CEO.
Increased Investments
In the last quarter ending 30 September, Microsoft reported expenditure totaling $34.9bn, a huge increase from $24 billion in the previous quarter.
The company has outlined plans to build two major AI “superclusters,” one of which — named Prometheus — is expected to reach 1 gigawatt of compute capacity when online in 2026. The company is also targeting over 1.3 million graphics-processing units (GPUs) by year-end as part of its infrastructure push.
Alphabet describes its approach as “full-stack”. In its Q3 2024 earnings communication,CEO Sundar Pichai highlighted investment in “data centres, chips, and a global fibre network” as central to its infrastructure strategy. He also emphasised that Google’s data-centres now deliver nearly four times more computing power per unit of electricity than five years ago, underlining the scale and efficiency of the build-out.
Microsoft CEO Satya Nadella, in a recent discussion, pointed out that the bottleneck in the AI infrastructure race is shifting: “compute is not the bottleneck … electricity is,” referencing the enormous power demands of scaled-up AI systems.
The company announced a capital expenditure hitting US $19.4 billion for the 3rd quarter of 2025. While there is no mention of exactly how much is set for AI development, a huge chunk is data centres, servers, next-gen chips, networking, and other infrastructure a large portion of which Meta says is for its AI efforts.