NVIDIA H100 SXM vs NVIDIA H200
What each costs to rent by the hour across the providers we track — NVIDIA H100 SXM on 39 providers, NVIDIA H200 on 53.
As of , the NVIDIA H200 has a median rate card of $4.54 per GPU-hour versus $3.52 for the NVIDIA H100 SXM — $1.02 (23%) more for 76% more memory.
Side by side
| NVIDIA H100 SXM | NVIDIA H200 | |
|---|---|---|
| Architecture | Hopper | Hopper |
| VRAM | 80 GB | 141 GB |
| Form factor | SXM | SXM |
| Median rate card, per GPU-hour | $3.52 | $4.54 |
| Cheapest listed rate | $1.06 Google Cloud, spot, GPU only captured | $2.09 Beam, on-demand, from captured |
| Median per GB of VRAM, per hour | $0.044 | $0.032 |
| Providers listing it | 39 | 53 |
Providers that list both
| Provider | NVIDIA H100 SXM | NVIDIA H200 | H200 vs H100 SXM |
|---|---|---|---|
 Seeweb | $2.12 captured | $2.91 captured | +$0.79 (+38%) |
 Bitdeer AI | $2.36 from captured | $2.76 from captured | +$0.40 (+17%) |
 HPC-AI.COM | $2.42 from captured | $2.68 from captured | +$0.26 (+11%) |
 Taiga Cloud (Northern Data Group) | $2.60 from captured | $3.00 from captured | +$0.40 (+15%) |
 Atlas Cloud | $2.95 captured | $3.50 captured | +$0.55 (+19%) |
 Civo | $2.99 captured | $3.49 captured | +$0.50 (+17%) |
 HexGrid | $2.99 Secure Cloud captured | $4.20 Secure Cloud captured | +$1.21 (+40%) |
 FPT AI Factory | $2.99 Southeast Asia captured | $6.60 Japan captured | +$3.61 (+121%) |
 Geodd | $3.00 captured | $2.72 captured | −$0.28 (−9%) |
 Gcore | $3.06 Bare Metal GPU Clusters captured | $3.56 Bare Metal GPU Clusters captured | +$0.51 (+17%) |
 Hyperbolic | $3.19 from captured | $3.99 from captured | +$0.80 (+25%) |
 Hyperstack | $3.20 captured | $3.99 captured | +$0.79 (+25%) |
 Canopy Wave | $3.25 captured | $4.00 captured | +$0.75 (+23%) |
 Jarvislabs | $3.49 captured | $4.59 captured | +$1.10 (+32%) |
 Barrack AI | $3.59 captured | $4.45 captured | +$0.86 (+24%) |
 Crusoe | $3.90 captured | $4.29 captured | +$0.39 (+10%) |
 Modal | $3.95 GPU only captured | $4.54 GPU only captured | +$0.59 (+15%) |
 RunPod | $3.99 Secure Cloud captured | $5.29 Secure Cloud captured | +$1.30 (+33%) |
 Together AI | $3.99 captured | $5.99 captured | +$2.00 (+50%) |
 GPU.ai | $3.99 from captured | $5.29 from captured | +$1.30 (+33%) |
 Neysa | $4.39 from captured | $4.72 from captured | +$0.33 (+8%) |
 Paperspace | $4.41 GPU Droplets captured | $4.47 GPU Droplets captured | +$0.06 (+1%) |
 Nebius | $4.50 captured | $5.40 captured | +$0.90 (+20%) |
 Verda | $4.51 captured | $5.89 captured | +$1.37 (+30%) |
 CoreWeave | $6.16 captured | $6.30 captured | +$0.15 (+2%) |
 Amazon Web Services | $6.88 us-east-1 captured | $7.91 us-east-1 captured | +$1.03 (+15%) |
 Google Cloud | $9.80 us-west8, GPU only captured | $9.31 us-west1, GPU only captured | −$0.48 (−5%) |
 Oracle Cloud (OCI) | $10.00 captured | $10.00 captured | same |
 Microsoft Azure | $11.06 eastus captured | $10.60 westus2 captured | −$0.46 (−4%) |
Medians are median rate cards — each provider once, at its cheapest on-demand rate, community capacity and marketplace floors out — the same figure as on each card's page; the cheapest listed rate may be spot or a community tier and is labelled as such. Compare providers directly on the overview, or see each GPU's full table: NVIDIA H100 SXM, NVIDIA H200. On your own page: embed this comparison (/embed/gpus/h100-sxm-vs-h200, an iframe, no script).
Data updated: