The NVIDIA H100 (Hopper, 80 GB HBM3) rents from $1.59/hr, averaged across 12 cloud providers, updated 2026-06-30. On-demand pricing has been roughly flat over the past 38 days.
The workhorse of the current AI build-out — strong FP16/FP8 training and inference, and the most widely available data-center GPU. Hopper · 80 GB HBM3 · 3.35 TB/s memory bandwidth · launched 2022.
▮ on-demand /hr · ▮ spot /hr
| Date | On-demand | Spot |
|---|---|---|
| 2026-06-30 | $3.55 | $1.59 |
| 2026-06-29 | $3.62 | $1.64 |
| 2026-06-28 | $3.62 | $1.80 |
| 2026-06-27 | $3.58 | $1.70 |
| 2026-06-26 | $3.59 | $1.69 |
| 2026-06-24 | $3.60 | $1.76 |
| 2026-06-23 | $3.58 | $1.77 |
| 2026-06-19 | $3.77 | $1.82 |
| 2026-06-18 | $3.67 | $1.96 |
| 2026-06-17 | $3.63 | $1.73 |
| 2026-06-15 | $3.63 | $1.89 |
| 2026-06-14 | $3.58 | $1.83 |
| 2026-06-13 | $3.63 | $1.80 |
| 2026-06-12 | $3.72 | $1.91 |
| GPU | On-demand /hr | Spot /hr | Memory | |
|---|---|---|---|---|
| H100 | $3.55 | $1.59 | 80 GB HBM3 | |
| B200 | $6.77 | $3.51 | 192 GB HBM3e | H100 vs B200 → |
| H200 | $4.17 | $2.81 | 141 GB HBM3e | H100 vs H200 → |
| MI355X | $8.60 | $— | 288 GB HBM3e | H100 vs MI355X → |
| MI325X | $3.26 | $— | 256 GB HBM3e | H100 vs MI325X → |
| MI300X | $3.10 | $1.20 | 192 GB HBM3 | H100 vs MI300X → |
The H100 averages $3.55/hr on-demand, $1.59/hr spot, and $3.09/hr reserved across 12 cloud providers (updated 2026-06-30).
The lowest tier is spot at about $1.59/hr on average; reserved (1–3 year commitment) runs ~$3.09/hr. Compare live per-provider rates on the Bargo GPU price tracker.
At on-demand $3.55/hr, one H100 runs about $2,592/month (730 hours); reserved pricing lowers that to roughly $2,256/month.
On-demand, the H100 is $3.55/hr vs the B200 at $6.77/hr — cheaper per hour. See the full H100 vs B200 comparison for specs and cost-per-job context.
Spot ($1.59/hr) is interruptible spare capacity at a discount to on-demand ($3.55/hr); use spot for checkpointed or fault-tolerant jobs and on-demand or reserved for long, uninterrupted runs.
Cross-provider averages from public cloud GPU listings. Informational only.