NVIDIA and Microsoft announced a multi-year partnership to build artificial intelligence supercomputers in the cloud, adding tens of thousands of NVIDIA GPUs to Microsoft Azure. the new agreement makes Azure the first public cloud to incorporate NVIDIA’s full AI stack – including its GPUs, networking and AI software.
By enhancing Azure’s infrastructure with NVIDIA’s full AI suite, more enterprises will be able to train, deploy and scale AI – including large, advanced models.
In a statement, Manuvir Das, vice president of enterprise computing at NVIDIA, said, “Advances in AI technology and industry adoption are accelerating, with breakthroughs in the underlying model sparking a wave of research, fostering new startups, and enabling new enterprise applications.”
Azure instances already feature NVIDIA A100 GPUs paired with NVIDIA Quantum 200Gb/s InfiniBand networks, and future instances will integrate NVIDIA H100 GPUs and NVIDIA Quantum-2 400Gb/s InfiniBand networks.
Meanwhile, NVIDIA’s software suite, NVIDIA AI Enterprise, has been certified and supports Microsoft Azure instances with A100 GPUs, and support for Azure instances with H100 GPUs will be added in a future software release.
Azure enterprise customers will also receive NVIDIA’s full suite of AI workflows and software development kits optimized for Azure, and the two companies will work together to optimize Microsoft’s DeepSpeed deep learning optimization software.
This new deal also gives NVIDIA the opportunity to use Azure virtual machines to advance its generative AI research.
