U.S. computing giants Nvidia and Microsoft have joined hands to develop artificial intelligence (AI) supercomputers.
In an official announcement on the 16th (local time), Nvidia and Microsoft decided to cooperate for many years to build one of the world’s most powerful AI supercomputers designed to handle the vast amount of computing tasks needed for AI learning.
Through this cooperation, the two companies plan to respond to the demand for AI supercomputers, which have been showing explosive growth recently. Nvidia plans to accelerate the development of generative AI models such as DALL-E by utilizing MS’s scalable virtual machine instance.
AI supercomputers based on the MS Azure cloud infrastructure use the Quantum-2 Ipiniband network platform, which provides tens of thousands of powerful GPUs for H100 and A100 data centers and 400Gb/s speeds.
Through this combination, more companies will be able to use the latest AI models as well as easily learn, distribute, and expand AI. In addition, the two companies are planning to jointly develop ‘Deep Speed’, Microsoft’s deep learning optimization software.
“The development of AI technology and industrial application are accelerating,” said Manuber Das, vice president of Nvidia’s enterprise computing division. “Through cooperation with MS, we will provide cutting-edge AI infrastructure and software necessary for companies and research institutes to make it easier to utilize AI’s innovative power.”