Nvidia has announced the availability of the GH200 super chip, a more powerful version of its high-end processors for generative AI use. As the demand for running large AI models continues to grow, the GH200 is designed to handle the most complex generative AI workloads, including large language models, recommender systems, and vector databases. It will feature the same GPU as the H100, Nvidia's current most powerful and popular AI offering, but with triple the memory capacity. The GH200 is expected to be available in the second quarter of 2024. At the same time, Nvidia has not disclosed the price for the GH200, its H100 line currently sells for approximately$40,000.

Running complex AI models requires powerful GPUs to handle the computations necessary for tasks like generating text or images. Even with Nvidia's H100 chips, some models must be split among multiple GPUs to run efficiently. Nvidia's dominance in generative AI-capable GPUs has given it a near monopoly in the market, with cloud providers like AWS, Azure, and Google relying on its H100 Tensor Core GPUs. These providers often offer additional services to help clients leverage large language models and differentiate themselves.