Kovara
Loading your page…
You can continue using the navigation while this page loads.
You can continue using the navigation while this page loads.
GeForce RTX 4090 Founders / reference has 70% less memory than H100 SXM (24 GB vs 80 GB). Its memory bandwidth is 70% less (1.008 vs 3.35 TB/s). In the cloud, the median on-demand GeForce RTX 4090 Founders / reference price is $0.33 per GPU-hour against $3.85 for the H100 SXM. Per gigabyte of GPU memory, that is $0.014 vs $0.048 per hour.
Manufacturer figures as recorded by Kovara. Higher memory and bandwidth are highlighted; neither is a benchmark of your workload.
| GeForce RTX 4090 Founders / reference | H100 SXM | |
|---|---|---|
| Manufacturer | NVIDIA | NVIDIA |
| Architecture | Ada Lovelace | Hopper |
| Memory | 24 GB GDDR6X | 80 GB HBM3 |
| Memory bandwidth | 1.008 TB/s | 3.35 TB/s |
| Power (TDP) | 450 W | 700 W |
| Form factor | PCIe consumer | SXM |
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.
| GeForce RTX 4090 Founders / reference | H100 SXM | |
|---|---|---|
| Lowest price | $0.16 | $1.99 |
| Median price | $0.33 | $3.85 |
| Highest price | $0.74 | $12.29 |
| Median per GB of memory | $0.014 | $0.048 |
| Public offers (all terms) | 9 | 59 |
| Providers | 4 | 21 |
Lowest tracked price per provider. Open a provider to see every offer, region and term.
Right now the cheapest on-demand GeForce RTX 4090 Founders / reference Kovara tracks is $0.16 per GPU-hour and the cheapest H100 SXM is $1.99. Medians are $0.33 and $3.85. The cheaper GPU per hour is not always cheaper per job: a faster or larger GPU can finish sooner or need fewer units.
The H100 SXM has more memory: 80 GB against 24 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.