NVIDIA H100 SXM vs AMD Instinct MI300X
What each costs to rent by the hour across the providers we track — NVIDIA H100 SXM on 42 providers, AMD Instinct MI300X on 8.
As of , the NVIDIA H100 SXM has a median rate card of $3.53 per GPU-hour versus $2.60 for the AMD Instinct MI300X — $0.93 (26%) more for 58% less memory.
Side by side
| NVIDIA H100 SXM | AMD Instinct MI300X | |
|---|---|---|
| Architecture | Hopper | CDNA 3 |
| VRAM | 80 GB | 192 GB |
| Form factor | SXM | OAM |
| Median rate card, per GPU-hour | $3.53 | $2.60 |
| Cheapest listed rate | $1.06 Google Cloud, spot, GPU only captured | $1.11 Microsoft Azure, spot captured |
| Median per GB of VRAM, per hour | $0.044 | $0.014 |
| Providers listing it | 42 | 8 |
Providers that list both
| Provider | NVIDIA H100 SXM | AMD Instinct MI300X | MI300X vs H100 SXM |
|---|---|---|---|
 Seeweb | $2.09 captured | $2.60 captured | +$0.51 (+24%) |
 RunPod | $4.39 Secure Cloud captured | $2.39 Secure Cloud captured | −$2.00 (−46%) |
 Paperspace | $4.41 GPU Droplets captured | $2.59 GPU Droplets captured | −$1.82 (−41%) |
 Oracle Cloud (OCI) | $10.00 captured | $6.00 captured | −$4.00 (−40%) |
 Microsoft Azure | $11.06 eastus captured | $6.00 eastus2 captured | −$5.06 (−46%) |
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, AMD Instinct MI300X. On your own page: embed this comparison (/embed/gpus/h100-sxm-vs-mi300x, an iframe, no script).
Data updated: