GPU comparison
L40S vs H100: specs and rental prices
NVIDIA L40S against NVIDIA H100: 48 GB against 80 GB of memory, lowest listed on-demand price $1.50 against $2.50 per GPU-hour. Specs from NVIDIA, prices as of Oct 25, 2026.
In short
- The H100 lists 80 GB of GPU memory against 48 GB for the L40S: 1.67× higher.
- Memory type is GDDR6 on the L40S and HBM3 on the H100.
- The H100 lists 3,350 GB/s of memory bandwidth against 864 GB/s for the L40S: 3.88× higher.
- The H100 lists 1,979 TFLOPS of FP16 / BF16 tensor throughput against 733 TFLOPS for the L40S: 2.7× higher.
- Maximum power is listed at 350 W for the L40S and 700 W for the H100.
- Not every figure can be compared: at least one vendor page does not publish: FP64 throughput; GPU-to-GPU interconnect.
- The lowest listed on-demand price is $1.50 per GPU-hour for the L40S (Crusoe) and $2.50 for the H100 (Hyperstack).
- Per GB of GPU memory that is $0.031 per hour for the L40S and $0.031 for the H100.
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 | L40S | H100 SXM | Difference |
|---|---|---|---|
| GPU memory | 48 GB | 80 GB | H100 1.67× higher |
| Memory type | GDDR6 | HBM3 | Different |
| Memory bandwidth | 864 GB/s | 3,350 GB/s | H100 3.88× higher |
| FP8 tensor throughput, with sparsity | 1,466 TFLOPS | 3,958 TFLOPS | H100 2.7× higher |
| FP16 / BF16 tensor throughput, with sparsity | 733 TFLOPS | 1,979 TFLOPS | H100 2.7× higher |
| TF32 tensor throughput, with sparsity | 366 TFLOPS | 989 TFLOPS | H100 2.7× higher |
| FP64 throughput | Not published | 34 TFLOPS | — |
| Maximum power | 350 W | 700 W | H100 2× higher |
| GPU-to-GPU interconnect | Not published | 900 GB/s | — |
| CUDA cores | 18,176 | Not published | — |
Compared versions: L40S and H100 SXM. 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: L40S specifications · H100 specifications.
Rental prices
| Rental price | L40S | H100 | Difference |
|---|---|---|---|
| Lowest listed on-demand price | $1.50 per GPU-hour · Crusoe · checked Oct 25, 2026 | $2.50 per GPU-hour · Hyperstack · checked Oct 25, 2026 | H100 1.67× higher |
| Lowest listed spot price | $0.98 per GPU-hour · CoreWeave · checked Oct 25, 2026 | $0.79 per GPU-hour · Nebius · checked Oct 25, 2026 | L40S 1.25× higher |
| Price per GB of memory | $0.031 per GB of GPU memory per hour | $0.031 per GB of GPU memory per hour | Same |
| Fresh listings | 10 from 7 providers | 26 from 12 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: L40S prices · H100 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). L40S vs H100. Retrieved from https://datacentercensus.com/gpu-prices/compare/l40s-vs-h100 |
|---|---|
| MLA | “L40S vs H100.” Data Center Census, Oct 25, 2026, https://datacentercensus.com/gpu-prices/compare/l40s-vs-h100. |
| Link | https://datacentercensus.com/gpu-prices/compare/l40s-vs-h100 |
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.