Microsoft Showcases Existing AI Data Centers as OpenAI Expands Globally

Microsoft CEO Satya Nadella has announced the deployment of the company’s first massive AI system, or “AI factory,” across its Azure global data centers. This move comes as OpenAI, Microsoft’s partner and competitor, is expanding its own data center footprint with high-profile deals with Nvidia and AMD.

The AI factory, a cluster of over 4,600 Nvidia GB300 rack computers equipped with the Blackwell Ultra GPU chip, is connected via Nvidia’s InfiniBand networking technology. Microsoft promises to deploy hundreds of thousands of these GPUs globally, positioning itself as uniquely capable of handling frontier AI workloads.

Microsoft's AI Infrastructure

Each AI factory consists of more than 4,600 Nvidia GB300 rack computers, each featuring the highly sought-after Blackwell Ultra GPU. These systems are interconnected using Nvidia’s InfiniBand, a super-fast networking technology that Nvidia acquired through its $6.9 billion purchase of Mellanox in 2019.

“This is the first of many such Nvidia AI factories that will be deployed across Microsoft Azure’s global data centers to run OpenAI workloads,” Nadella tweeted, sharing a video of the deployment. Microsoft already operates more than 300 data centers in 34 countries, giving it a significant advantage in the race to meet the demands of advanced AI models.

Industry Context and Implications

OpenAI, which has committed an estimated $1 trillion to build its own data centers, recently signed deals with both Nvidia and AMD. OpenAI CEO Sam Altman has indicated that more data center expansions are on the horizon, further intensifying the competition in the AI infrastructure space.

Microsoft’s existing data center network and its new AI factories are designed to handle the next generation of AI models, which could have hundreds of trillions of parameters. The company aims to leverage its extensive infrastructure to support the growing computational needs of AI research and development.

Future Outlook

Microsoft CTO Kevin Scott is set to provide more details on the company’s AI infrastructure plans at TechCrunch Disrupt, scheduled for October 27 to October 29 in San Francisco. As the AI arms race continues, Microsoft’s robust data center capabilities position it as a key player in the evolving landscape of artificial intelligence.

References

• While OpenAI races to build AI data centers, Nadella reminds us that Microsoft already has them | TechCrunch