GPU comparison
H100 vs MI300X: specs and rental prices
NVIDIA H100 against AMD Instinct MI300X: 80 GB against 192 GB of memory, lowest listed on-demand price $2.50 against $2.59 per GPU-hour. Specs from NVIDIA and AMD, prices as of Oct 25, 2026.
In short
- The MI300X lists 192 GB of GPU memory against 80 GB for the H100: 2.4× higher.
- Memory type is HBM3 on the H100 and HBM3 on the MI300X.
- The MI300X lists 5,300 GB/s of memory bandwidth against 3,350 GB/s for the H100: 1.58× higher.
- The MI300X lists 2,614.9 TFLOPS of FP16 / BF16 tensor throughput against 1,979 TFLOPS for the H100: 1.32× higher.
- Maximum power is listed at 700 W for the H100 and 750 W for the MI300X.
- Not every figure can be compared: at least one vendor page does not publish: GPU-to-GPU interconnect.
- The lowest listed on-demand price is $2.50 per GPU-hour for the H100 (Hyperstack) and $2.59 for the MI300X (DigitalOcean).
- Per GB of GPU memory that is $0.031 per hour for the H100 and $0.013 for the MI300X.
This summary is generated from the tables below. Prices were last checked Oct 25, 2026. A listing not re-checked for 7 days is greyed and left out of every lowest-price figure; after 14 days it is removed from the page.
Specifications side by side
| Specification | H100 SXM | MI300X | Difference |
|---|---|---|---|
| GPU memory | 80 GB | 192 GB | MI300X 2.4× higher |
| Memory type | HBM3 | HBM3 | Same |
| Memory bandwidth | 3,350 GB/s | 5,300 GB/s | MI300X 1.58× higher |
| FP8 tensor throughput, with sparsity | 3,958 TFLOPS | 5,229.8 TFLOPS | MI300X 1.32× higher |
| FP16 / BF16 tensor throughput, with sparsity | 1,979 TFLOPS | 2,614.9 TFLOPS | MI300X 1.32× higher |
| TF32 tensor throughput, with sparsity | 989 TFLOPS | 1,307.4 TFLOPS | MI300X 1.32× higher |
| FP64 throughput | 34 TFLOPS | 81.7 TFLOPS | MI300X 2.4× higher |
| Maximum power | 700 W | 750 W | MI300X 1.07× higher |
| GPU-to-GPU interconnect | 900 GB/s | Not published | — |
Compared versions: H100 SXM and MI300X. The figure in the stronger cell is shaded; power is shown without a winner. Two throughput figures are compared only when both vendors state them on the same basis (both with sparsity, or both dense), and “derived” means a total the vendor publishes divided by the GPU count it publishes. Full tables with sources: H100 specifications · MI300X specifications.
Rental prices
| Rental price | H100 | MI300X | Difference |
|---|---|---|---|
| Lowest listed on-demand price | $2.50 per GPU-hour · Hyperstack · checked Oct 25, 2026 | $2.59 per GPU-hour · DigitalOcean · checked Oct 25, 2026 | MI300X 1.04× higher |
| Lowest listed spot price | $0.79 per GPU-hour · Nebius · checked Oct 25, 2026 | $1.11 per GPU-hour · Microsoft Azure · checked Oct 25, 2026 | MI300X 1.4× higher |
| Price per GB of memory | $0.031 per GB of GPU memory per hour | $0.013 per GB of GPU memory per hour | H100 2.32× higher |
| Fresh listings | 26 from 12 providers | 6 from 4 providers | — |
Prices cover every version and instance size the providers list, so they are not tied to the versions compared above; the price per GB of memory divides the lowest fresh on-demand price by the memory of the compared version. Listed prices are not like for like across providers. See every listing: H100 prices · MI300X prices.
More on these GPUs
Cite this page
Last updated Oct 25, 2026. Figures may change as records are updated — include the date you retrieved them.
| APA | Data Center Census. (2026). H100 vs MI300X. Retrieved from https://datacentercensus.com/gpu-prices/compare/h100-vs-mi300x |
|---|---|
| MLA | “H100 vs MI300X.” Data Center Census, Oct 25, 2026, https://datacentercensus.com/gpu-prices/compare/h100-vs-mi300x. |
| Link | https://datacentercensus.com/gpu-prices/compare/h100-vs-mi300x |
Republishing a figure or table? Attribute it to Data Center Census with a link to this page. The underlying data is downloadable as JSON and CSV; the press kit has standing figures and contact details.