According to a recent report by The Information Technology website, citing people familiar with the matter, Microsoft is preparing to launch a new chip that it specifically designed to train and operate artificial intelligence systems.
The new chips will be unveiled at the Microsoft Ignite IT and AI conference in mid-November.
Reports earlier this year revealed that the American company was working on designing and developing these chips due to the great demand for artificial intelligence technology.
Microsoft is trying to reduce its heavy reliance on Nvidia, the world's largest maker of artificial intelligence chips, which is facing a rise in orders and a shortage in supplies.
Microsoft's chips will be similar to NVIDIA's and will be primarily used in data center servers.
Microsoft currently uses Nvidia GPUs in data center servers to run and develop advanced AI software for its customers, such as: B. OpenAI (owner of the ChatGPT bot) as well as our own software and projects.
GPU chips are used in artificial intelligence due to their ability to efficiently perform parallel computing operations.
These chips are usually designed primarily to speed up graphics and screen displays, but they also have enormous computing power and can be used in the field of artificial intelligence to process large amounts of data in parallel, such as image and video analysis, machine learning, and further big data analysis, and implement parallel operation. , higher speed and better performance of the CPU chip.