Company

NVIDIA

NVDAcompute

NVIDIA designs the GPUs and AI accelerators that dominate model training and inference, and has extended that position into networking (via its Mellanox/InfiniBand and NVLink lines) and full-rack systems (NVL72, Grace Blackwell). Its data-center segment now accounts for the large majority of revenue, driven by hyperscaler capex on AI infrastructure. CEO Jensen Huang (see jensenhuang) sets the annual GPU architecture cadence that the rest of the compute stack largely builds around.

Recent