Chip design company Nvidia is partnering with Microsoft to construct a large-scale AI cloud computer, the firm announced on Wednesday. Described as a multiyear collaboration, the project will integrate Microsoft's Azure cloud computing system with Nvidia's A100 and H100 chips, its Quantum-2 InfiniBand network and Nvidia AI Enterprise software. The AI computer will be the "first public cloud" to do so, the companies said.
On Monday, Nvidia said that Microsoft Azure is the first to use the Quantum-2 InfiniBand platform, which performs data and simulation operations at high speeds. The A100 and H100 processors are the company's most powerful, and can be used in a variety of applications including science and manufacturing. The AI supercomputer will harness Microsoft's virtual machines to handle more complex research and automated functions for organizations.
Nvidia also says it will use Azure to develop advanced generative AI that provides machines with the ability to create new content like code, text video or images.