AMD on Tuesday unveiled its most advanced GPU for artificial intelligence, the MI300X, which the company says will start shipping to some customers later this year. The announcement represents the strongest challenge yet to Nvidia, which currently dominates the market for AI chips with over 80% market share.
The MI300X is based on AMD’s new CDNA 3 architecture, which is designed specifically for AI workloads. It features up to 192GB of high-bandwidth memory and can support up to 128 AI accelerators. This makes it significantly more powerful than Nvidia’s rival H100 chip, which has 120GB of memory and supports up to 96 accelerators.
AMD says the MI300X will be ideal for a wide range of AI applications, including natural language processing, computer vision, and machine learning. It will also be used in data centers and supercomputers, where it will help to accelerate the development of new AI models and applications.
The launch of the MI300X is a significant development in the AI chip market. It is the first time that AMD has offered a chip that can compete with Nvidia’s high-end offerings. This could help to shake up the market and give customers more choice.
It remains to be seen how well the MI300X will perform in the real world. However, AMD’s announcement is a positive sign for the company and for the AI chip market as a whole. It suggests that AMD is serious about competing with Nvidia in this important market.