Kovara
Loading your page…
You can continue using the navigation while this page loads.
You can continue using the navigation while this page loads.
H200 SXM has 194% more memory than L40S 48GB (141 GB vs 48 GB). Its memory bandwidth is 456% more (4.8 vs 0.864 TB/s). In the cloud, the median on-demand H200 SXM price is $4.47 per GPU-hour against $1.50 for the L40S 48GB. Per gigabyte of GPU memory, that is $0.032 vs $0.031 per hour.
Manufacturer figures as recorded by Kovara. Higher memory and bandwidth are highlighted; neither is a benchmark of your workload.
| H200 SXM | L40S 48GB | |
|---|---|---|
| Manufacturer | NVIDIA | NVIDIA |
| Architecture | Hopper | Ada Lovelace |
| Memory | 141 GB HBM3E | 48 GB GDDR6 |
| Memory bandwidth | 4.8 TB/s | 0.864 TB/s |
| Power (TDP) | 700 W | 350 W |
| Form factor | SXM | PCIe dual-slot |
Public offers Kovara tracks, normalised to US dollars per GPU-hour. Prices use on-demand terms where available; spot and reserved offers are on each GPU’s price pages. Lower is highlighted.
| H200 SXM | L40S 48GB | |
|---|---|---|
| Lowest price | $2.60 | $0.97 |
| Median price | $4.47 | $1.50 |
| Highest price | $10.60 | $2.25 |
| Median per GB of memory | $0.032 | $0.031 |
| Public offers (all terms) | 41 | 38 |
| Providers | 14 | 9 |
Lowest tracked price per provider. Open a provider to see every offer, region and term.
Right now the cheapest on-demand H200 SXM Kovara tracks is $2.60 per GPU-hour and the cheapest L40S 48GB is $0.97. Medians are $4.47 and $1.50. The cheaper GPU per hour is not always cheaper per job: a faster or larger GPU can finish sooner or need fewer units.
The H200 SXM has more memory: 141 GB against 48 GB. Memory decides which models and batch sizes fit on one GPU.
No. This page compares manufacturer specifications and public prices. Real performance depends on the model, precision, software and interconnect. Kova can help you estimate cost for a specific workload.