Microsoft plans to announce the launch of the new product at its Ignite event in Seattle in November, according to reports.Its first customized chip designed for artificial intelligence computing, thereby getting rid of the reliance on NVIDIA to provide GPU acceleration.
Due to the explosive market demand for AI computing applications, NVIDIA's GPUs for AI acceleration have been in short supply, and there is also the possibility of over-reliance on them. Therefore, Microsoft plans to build its own AI computing processing chips to reduce the need to rely on external resources.
This artificial intelligence computing processing chip is internally referred to as "Athena". It will be unveiled at the Ignite event at the earliest and will be compared with the NVIDIA H100 GPU acceleration performance. It has also been tested with OpenAI.
This processing chip began to be developed in 2019. In addition to hoping to reduce dependence on NVIDIA and reduce procurement costs, it will also serve as a bargaining chip with NVIDIA, but it does not mean that it is preparing to end its partnership with it.
By investing in its own artificial intelligence computing processing chips, Microsoft will be able to compete with AWS and Google Cloud, thereby accelerating the operation of cloud services at a lower cost and higher efficiency, while also accelerating the development of artificial intelligence computing capabilities driven by cloud services.
