Microsoft’s artificial intelligence ambitions have had the company inking deals with a ton of partners recently and the company is adding NVIDIA to its list to deliver AI to even more businesses soon.
According to a press release from NVIDIA, Microsoft, and the GPU manufacturer have agreed to partner bring the Grace Blackwell Superchip to Microsoft Azure customers in the near future.
The press release follows Microsoft’s CEO Satya Nadella attending NVIDIA’s GPU Technology Conference where he dropped the announcement of the two companies coming together to promote NVIDIA’s “generative AI and Omniverse” ambitions.
Together with NVIDIA, we are making the promise of AI real, helping to drive new benefits and productivity gains for people and organizations everywhere. From bringing the GB200 Grace Blackwell processor to Azure, to new integrations between DGX Cloud and Microsoft Fabric, the announcements we are making today will ensure customers have the most comprehensive platforms and tools across every layer of the Copilot stack, from silicon to software, to build their own breakthrough AI capability.
Satya Nadella, Chairman and CEO, Microsoft.
AI is transforming our daily lives – opening up a world of new
For some context, NVIDIA’s Grace Blackwell Super chip is the NVIDIA Grace CPU and the NVIDIA Blackwell GPU, which together form the NVIDIA Grace Blackwell Superchip. This superchip is designed for large-scale AI and computing tasks, boasting significant performance capabilities:
- 40 petaflops of AI performance, which is a measure of the processor’s ability to perform one quadrillion floating-point operations per second1.
- 864GB of HBM3E memory, providing a vast amount of high-bandwidth memory to support large and complex AI models1.
- 16TB/sec of memory bandwidth, ensuring rapid data transfer rates that are essential for high-performance computing tasks1.
Furthermore, Microsoft will be among the first companies to support NVIDIA’s Grace Blackwell GB200 and Quantum-X800 InfiniBand networking technologies in its Azure cloud platform services portfolio.
As for Microsoft’s part in the deal, it will be tasked with providing Azure NC H100 v5 virtual machine technology aimed at supporting mid-level AI training with a bit of inference tasking coupled. NVIDIA’s VM stack will consist of NVIDIA’s H100 NVL platform paired with dual H100 94GB PCIe Tensor Core powered GPUs and the company’s Multi-Instance GPU technology.
Microsoft will support these new NVIDIA technologies with its own enhancements to Microsoft 365 powered by Copilot. Combined, Microsoft and NVIDIA promise enhanced inference predictions, improved user experiences and productivity as well as the flexibility to target specific industries that can include health care, factory workers, and biotechnology companies.
