Microsoft CEO Satya Nadella has unveiled the company’s first large-scale AI system, or “AI factory,” powered by Nvidia’s Blackwell Ultra GPUs, marking a major step in expanding its AI infrastructure globally.
In a post on X (formerly Twitter), Nadella shared a video of the system in action, calling it the “first of many” Nvidia-powered AI factories that will soon be deployed across Microsoft Azure’s 300+ global data centers to run OpenAI workloads.
Each AI factory consists of over 4,600 Nvidia GB300 rack computers equipped with the sought-after Blackwell Ultra GPU and connected via Nvidia’s high-speed InfiniBand networking technology. The company said it plans to deploy hundreds of thousands of Blackwell Ultra GPUs worldwide as part of its infrastructure expansion.
The announcement comes shortly after OpenAI signed separate data center deals with Nvidia and AMD, reinforcing growing competition, and collaboration, between the two partners. OpenAI CEO Sam Altman has reportedly secured commitments worth over $1 trillion to build additional data centers, with more expected soon.
By highlighting its massive existing network, Microsoft aims to remind the industry that it already possesses a robust global data center footprint capable of handling frontier AI workloads and training next-generation models with “hundreds of trillions of parameters.”
Related: Microsoft Expands Renewable Energy Push with 100 MW Solar Deal in Japan
More details about Microsoft’s AI infrastructure strategy are expected later this month when CTO Kevin Scott speaks at TechCrunch Disrupt 2025, taking place October 27–29 in San Francisco.
