NVIDIA H100

Introducing the NVIDIA H100, NVIDIA's latest data center GPU built for AI and HPC convergence. It is engineered for universal workloads, offering servers with various H100 configurations to meet different performance and efficiency requirements, enabling both innovation and optimization.

NVIDIA H100 80GB HBM3

GPUs
8x NVIDIA H100 80GB HBM3
CPUs
2x Intel(R) Xeon(R) Platinum 8462Y+ | 64 cores
Memory
2.05 TB
NVMe Total
61.93 TB
NVMe Disks
9
Location
Washington

On demand

$1.85per card-hour

Interruptible

$1.65per card-hour