NVIDIA H200 cloud pricing per GPU-hour
141GB HBM3e — top-end training/inference. Public on-demand list prices across 5 cloud providers, normalized to USD per single GPU per hour. Updated 2026-06-10 — see methodology.
| Provider | Instance | On-demand $/GPU-hr | 1-yr committed | Interconnect |
|---|---|---|---|---|
| NebiusBEST | HGX H200 | $4.50 | $2.93 | 3.2 Tbps InfiniBand |
| CoreWeave | HGX H200 | $6.31 | $4.25 | 3.2 Tbps InfiniBand |
| AWS | p5en.48xlarge | $7.91 | $6.69 | 3.2 Tbps EFAv3 |
| Oracle (OCI) | BM.GPU.H200.8 | $10.00 | $6.50 | 3.2 Tbps RoCEv2 (cluster network) |
| Google Cloud | a3-ultragpu-8g | $10.75 | $6.45 | 3.2 Tbps RoCE |
Public list prices and published reference rates; actual costs vary by region, commitment, and negotiated contract. Verify with each provider before purchasing.
Looking for cheaper GPUs?
SponsoredMarketplace and budget providers often undercut the list prices above. These are partner links — how this works.
On-demand & spot GPUs (incl. H100/H200) billed by the second — often the cheapest way to start.
GPU marketplace with the lowest rates anywhere — rent spare capacity from verified hosts.
Dedicated NVIDIA clusters and on-demand instances with no minimums.
Developer-friendly GPU droplets and notebooks — great for fine-tuning and inference.
FAQ
How much does an NVIDIA H200 cost per hour in the cloud?+
As of 2026-06-10, public on-demand list prices for the NVIDIA H200 range from $4.50 (Nebius) to $10.75 (Google Cloud) per GPU-hour — a 58% spread between the cheapest and most expensive provider.
Which cloud provider has the cheapest NVIDIA H200?+
Nebius currently lists the lowest public on-demand rate at $4.50/GPU-hr (HGX H200). Committed or reserved rates are lower on most providers.