NVIDIA H100 80GB PCIe Gen5 data center GPU with Hopper architecture. 80GB HBM2e delivers 2,039 GB/s bandwidth at 700W TDP. Transformer Engine accelerates FP8 training — the gold standard for GPT-scale LLM training and HPC workloads.
H100 80GB PCIe
Hopper GH100
80 GB HBM2e
2,039 GB/s
34 TFLOPS
495 TFLOPS
1,979 TFLOPS
1,979 TOPS
PCIe 5.0 x16
700W
Dual-slot, FHFL
None (PCIe only)
NVIDIA H100 80GB PCIe is the premier choice for hyperscale AI training clusters. With the Transformer Engine delivering near-double throughput on FP8 precision, a single H100 node can replace an entire rack of previous-generation A100 servers. Ideal for GPT-style LLM training, diffusion models, genomic sequencing, and computational fluid dynamics.
Need NVIDIA H100 80GB PCIe?
Enterprise hardware — configured, tested, deployed. Contact us for pricing and availability.
Request Quote