Software King of the World Microsoft said it is “derisking” its AI build-out costs by contracting GPU capacity to customers before it splurges.
It said Microsoft Cloud revenue topped $50bn for the first time, up 26 per cent year on year, while total revenue hit $81.3bn, up 17 per cent.
Vole claimed it added nearly one gigawatt of data centre capacity this quarter and saw a 10x jump in Microsoft 365 Copilot daily users.
That is a lot of tin being bolted into racks, so naturally, the money question came up, and Vole tried to sound like it is not freewheeling.
Microsoft’s chief financial officer, Amy Hood, said: “The majority of the capital that we’re spending today, and a lot of the GPUs that we’re buying, are already contracted for most of their useful life.”
She said Microsoft 365 subscriptions pay back faster than Azure, where usage mixes general compute and GPU-heavy work, so Microsoft is selling GPU contracts for the full useful life.
Vole is banging on about “tokens per watt per dollar”, which means squeezing utilisation and cutting total cost of ownership through silicon, systems and software. The company said it is seeing a six-year useful life for servers and GPUs in its data centre fleet, which is either clever engineering or optimistic accounting, depending on how you look at it.
Microsoft, chief executive Satya Nadella said: “We use software to continuously run even the latest models on the fleet. That’s what gives us [longevity]. That’s why we think about ageing the fleet constantly. It’s not about buying gear one year – it’s about each year you ride Moore’s Law, you add and use software, and then you optimise across all of it.”
After the Maia 200 AI acceleration hardware plug-in, Nadella was asked to focus more on silicon and delivered a full-stack sermon.
“When you have a new workload, you can start innovating end-to-end between the model and the silicon and the entire system. It’s just not even about the silicon,” Nadella said.
