NVIDIA's H100 GPUs for AI

There's like a 85% chance that OpenAI, that chat GPT was modeled and created on these A100 NVIDIA chips. The H100 is up to 10 times faster than the previous A100 for the types of calculations that are used in GPT-3 processing. And for example, the Amazon Web Services announced the EC2 Ultra Cluster P5 which can scale 20,000 interconnected H100 GPUs in one machine.