H100 cloud comparison
The same sourced rows arranged for a buying decision.
GPU Pricing
Current NVIDIA H100 price per hour, CT-H100 median, sourced per-provider rates, and the spot and reserved context around the list price.
NVIDIA H100 price per hour is sourced from 6 provider price pages, each linked and dated below. The CT-H100 Index condenses those public on-demand H100 rates into one citable daily median; the table shows the spread behind it.
Median sourced on-demand rate as of Jul 28, 2026
Change in the CT-H100 median since the prior daily point
As of Jul 28, 2026
Sourced rates
6 sourced provider rows, lowest H100 cost per hour first. Every row links the public price page it was observed on.
| Provider | Region | List $/GPU-hr | H100e $/hr | Availability | No change in | Source |
|---|---|---|---|---|---|---|
| Hyperstack | Multi-region | $3.20 | $2.40 | Available Limited | 20 days | Open price page |
| Crusoe | Multi-region | $3.90 | $3.90 | Available Limited | 61 days | Open price page |
| Lambda | Multi-region | $4.29 | $4.29 | Available | 61 days | Open price page |
| CoreWeave | us | $6.16 | $6.16 | Available Limited | 10 days | Open price page |
| AWS | us-east-1 | $6.88 | $12.29 | Quota Gated | 36 days | Open price page |
| Azure | eastus | $12.29 | $12.29 | Quota Gated | 73 days | Open price page |
Public on-demand list prices from provider price pages — not negotiated quotes, reserved terms, or spot rates. Each row carries its own source link and price date. How we source, label, and date evidence: methodology.
See every reviewed provider rowRead
The table shows the public on-demand slice of the H100 market.
Spot, reserved, committed, and capacity-block pricing are different products. Evidence rules live in methodology.
Open the CT-H100 IndexNext
The same sourced rows arranged for a buying decision.
Turn any rate in the table into a monthly estimate with utilization and overhead.
Check whether listed H100 capacity is actually orderable.