Nvidia Unveils H200 GPU, Bolstering AI Capabilities

OpenAI Secures More Chips, AMD Develops Competing MI300X Chip

Nvidia has unveiled the H200, a new graphics processing unit (GPU) designed for training and deploying artificial intelligence (AI) models. The new GPU is an upgrade from the H100, which has been in high demand among AI developers.
The H200 is the first GPU to use a new, faster memory spec called HBM3e, which allows it to perform 'inference', or using a large model after it's trained to generate text, images or predictions, nearly twice as fast as the H100.
OpenAI, which uses Nvidia's chips to train its AI models, has received a coveted order of H100 chips and is expecting more soon. The procurement of more chips signals that more sophisticated AI models will be ready in the near future.
Nvidia's new H200 GPU and the GH200 Grace Hopper 'superchip', which combines the H200 GPU with an Arm-based processor, will be used in several upcoming supercomputers, including the JUPITER supercomputer in Germany, which will become the 'world’s most powerful AI system' when it's installed in 2024.
AMD is developing a new chip, the MI300X, which is similar to the H200 in that it has additional memory over its predecessors. This chip is expected to compete with Nvidia's H200.