Nvidia introduced the H100 graphics processor, a major accelerator for generative-AI computing.
The H100 is based on Nvidia’s Hopper architecture and was announced in 2022. It is designed for data-center workloads, including training and running large artificial-intelligence models. Its tensor cores and high-speed memory help process the matrix operations central to modern machine learning.
Demand for AI accelerators helped Nvidia become one of the world’s largest companies by market capitalization. The company’s earlier success in gaming graphics processors provided the technology base, while CUDA software helped build a broad developer ecosystem.
The H100 is not a conventional consumer graphics card. It is a data-center accelerator, and its role is often confused with that of Nvidia’s GeForce products, which target personal computers and gaming.