H100 · providers
Read 12 November 2026
Who lists an H100, and where each sits
All providers with an H100 row
22 providers| Provider | Type | Standing | Lowest $/GPU-hr | Against median | Sizes | Spot | Reserved |
|---|---|---|---|---|---|---|---|
| Voltage Park | GPU cloud | Rank 1 of 19 | $1.99 | -49% | 1× | 0 | 0 |
| Vultr | GPU cloud | Rank 2 of 19 | $2.30 | -41% | 8× | 0 | 1 |
| Hyperstack | GPU cloud | Rank 3 of 19 | $2.50 | -36% | 1× | 1 | 3 |
| Massed Compute | GPU cloud | Rank 4 of 19 | $2.73 | -30% | 1× 2× 4× 8× | 0 | 0 |
| RunPod | GPU cloud | Rank 5 of 19 | $2.89 | -26% | 1× | 0 | 0 |
| Thunder Compute | GPU cloud | Rank 6 of 19 | $2.95 | -24% | 1× | 0 | 0 |
| Civo | GPU cloud | Rank 7 of 19 | $2.99 | -23% | 1× 8× | 0 | 0 |
| OVHcloud | GPU cloud | Rank 8 of 19 | $2.99 | -23% | 1× 2× 4× | 0 | 0 |
| Runcrateours, live | AI cloud (ours) | Rank 9 of 19 | $3.58 | -8% | 1× | 0 | 0 |
| Crusoe | GPU cloud | Rank 10 of 19 | $3.90 | +0% | 8× | 0 | 0 |
| Together AI | GPU cloud | Rank 11 of 19 | $3.99 | +2% | 8× | 1 | 2 |
| DigitalOcean (Paperspace) | GPU cloud | Rank 12 of 19 | $4.41 | +13% | 1× | 0 | 1 |
| Nebius | GPU cloud | Rank 13 of 19 | $4.50 | +15% | 8× | 1 | 0 |
| Verda | GPU cloud | Rank 14 of 19 | $5.09 | +30% | 1× | 2 | 0 |
| CoreWeave | GPU cloud | Rank 15 of 19 | $6.16 | +58% | 8× | 2 | 0 |
| Amazon Web Services | Hyperscale cloud | Rank 16 of 19 | $6.88 | +76% | 1× 8× | 0 | 0 |
| Microsoft Azure | Hyperscale cloud | Rank 17 of 19 | $6.98 | +79% | 1× 2× 8× | 3 | 0 |
| Oracle Cloud | Hyperscale cloud | Rank 18 of 19 | $10.00 | +156% | 1× | 0 | 0 |
| Google Cloud | Hyperscale cloud | Rank 19 of 19 | $11.06 | +184% | 8× | 2 | 0 |
| TensorDock | GPU marketplace | Not ranked: marketplace figure | $2.25 | -42% | 1× | 0 | 0 |
| Vast.ai | GPU marketplace | Not ranked: marketplace figure | $2.00 | -49% | 1× | 0 | 0 |
| Lambda | GPU cloud | Announced, takes effect 2026-11-16 | $3.79 | -3% | none now | 0 | 0 |
The rank counts only rows where the provider sets the price and that are in effect today. The difference to the median compares the provider's lowest figure with $3.90, the median of the 19 provider lows.
Providers: common questions
How many providers list an H100?
22 providers have at least one H100 row of any kind. 19 of them are ranked on a current on-demand price they set themselves. The rest are marketplace figures, announced prices or spot and reserved rows only.
Which providers are not ranked, and why?
TensorDock and Vast.ai publish only marketplace or "from" figures set by hosts. Lambda has announced prices that start later (2026-11-16). A provider can also be ranked and show a marketplace or announced row beside it; those extra rows are not counted.
Who is lowest and who is highest?
Voltage Park is lowest at $1.99 per GPU-hour and Google Cloud is highest at $11.06, among 19 ranked providers.
Does a provider page include spot and reserved prices?
Yes, when the provider has any. 12 spot rows and 7 reserved rows exist across all providers; each appears on its provider's page beside that provider's own on-demand price, and on the commitments page.
Disclosure
H100 Rate Card is published by the Runcrate team, who also run Runcrate, an AI cloud. Runcrate is listed in the table for the H100, marked "ours, live". It is sorted by price like every other row, with no pinned position. Other providers' prices are read from their own public pricing pages. We have no paid placement.