Microsoft gave details about the rollout of its services for AI. One thing is for sure, Nvidia is the big winner from this turning point for technology.
For the past few months, AI has been on everyone’s lips in tech. Google, Microsoft, Discord, Snapchat, Meta, all companies seem to be jumping into Artificial Intelligence to integrate it into their services.
Of all these companies, Microsoft seems to be the one that invests the most in the field with its partner OpenAI. The firm has already relaunched a version of Bing, and plans to go much further. In the shadows, however, there is another partner that seems to be the big winner in this race, Nvidia.
How to make AI think
Microsoft this week announced the release of a new series of cloud virtual machines for Microsoft Azure. They are designed to develop and use generative AI applications such as ChatGPT or MidJourney.
To animate these machines, Microsoft integrates Nvidia H100 GPUs based on the Hopper architecture and dominated. Each machine integrates at least 8 and up to “thousands”, interconnected by Nvidia Quantum-2 at 3.2 Tb/s per virtual machine.
In all, Microsoft has spent hundreds of millions of dollars to build an AI supercomputer. One of the firm’s first clients is OpenAI, the developers of GPT-3, who will be able to train their future models with this infrastructure.
This is not the first time Nvidia has been chosen as a technical partner for artificial intelligence. The GPU manufacturer itself regularly highlights these issues for the consumer through technologies such as Maxine, DLSS or Video Super Resolution.
Today, if the industry takes a close look at the new applications invented with AI, we must not forget what technologies are used to run these applications. In the little game of the race for the smartest chatbot, it may be Nvidia that is going to be out of the game.
Do you use Google News (News in France)? You can follow your favorite media. Continue Frandroid on Google News (and Numerama).