Mustafa Suleyman says Microsoft is "not able to build models in the very largest scale yet" but its "computation ramp is coming to enable us to do" it in 2026
Tech giant's AI chief says it will have the resources to build frontier systems later this year
Lead Source
More
Discussion
TechSnif Coverage
Microsoft Admits It Can't Build Top-Tier AI Models Yet
AI chief Mustafa Suleyman says Microsoft lacks the compute to train frontier models — but expects that to change in 2026.
Microsoft's AI boss Mustafa Suleyman is being refreshingly blunt: the tech giant still can't train models at the very largest scale. Not yet, anyway.
In comments reported by the Financial Times, Suleyman acknowledged Microsoft's current compute limitations but said the company's infrastructure ramp will close the gap. He expects Microsoft to have the resources needed to build frontier AI systems later this year.
The timeline matters. Microsoft has poured billions into AI infrastructure and its partnership with OpenAI, but actually training cutting-edge foundation models in-house has remained elusive. Suleyman is essentially saying 2026 is when Microsoft stops being just the checkbook behind someone else's models and starts competing at the frontier on its own terms.
Whether that compute ramp delivers on schedule will determine if Microsoft becomes an AI model builder — or stays the world's most expensive landlord for one.