The announcement comes as Nvidia's primary GPU rival, AMD, recently announced its own AI-oriented chip, the MI300X, which can support 192GB of memory and is being marketed for its capacity for AI inference. Other tech giants like Google and Amazon are also designing their own custom AI chips for inference. The GH200's larger memory capacity allows larger AI models to fit on a single system, potentially reducing the need for multiple systems or GPUs.
Key takeaways:
- Nvidia has announced a new chip, the GH200, designed to run artificial intelligence models, as it aims to maintain its dominance in the AI hardware space against competitors like AMD, Google, and Amazon.
- The GH200 has the same GPU as Nvidia's current highest-end AI chip, the H100, but pairs it with 141 gigabytes of memory and a 72-core ARM central processor. It is designed for the scale-out of the world's data centers.
- The new chip is particularly designed for inference, a computationally expensive process in working with AI models that requires a lot of processing power. The GH200's larger memory capacity allows larger AI models to fit on a single system.
- The GH200 will be available from Nvidia's distributors in the second quarter of next year. The announcement comes as AMD, Nvidia's primary GPU rival, recently announced its own AI-oriented chip, the MI300X.