Glossary term
Glossary term
Infrastructure and Serving
AI Supercomputing is high-performance infrastructure built to train and run large language models and generative AI workloads at scale, delivering the speed needed for complex reasoning and real-time enterprise inference.
xAI's Colossus cluster in Memphis runs over 100,000 Nvidia H100 GPUs to train Grok models.
Meta's Research SuperCluster (RSC) and OpenAI's clusters on Microsoft Azure power training of Llama and GPT models.
Nvidia DGX SuperPOD and the Stargate project announced in January 2025 are flagship AI supercomputing platforms.