NVIDIA GeForce RTX 5090 Founders / reference vs NVIDIA L40 48GB
GeForce RTX 5090 Founders / reference has 33% less memory than L40 48GB (32 GB vs 48 GB). Its memory bandwidth is 107% more (1.792 vs 0.864 TB/s). In the cloud, the median on-demand GeForce RTX 5090 Founders / reference price is $0.54 per GPU-hour against $0.86 for the L40 48GB. Per gigabyte of GPU memory, that is $0.017 vs $0.018 per hour.
Specifications
Manufacturer figures as recorded by Kovara. Higher memory and bandwidth are highlighted; neither is a benchmark of your workload.
| GeForce RTX 5090 Founders / reference | L40 48GB | |
|---|---|---|
| Manufacturer | NVIDIA | NVIDIA |
| Architecture | Blackwell | Ada Lovelace |
| Memory | 32 GB GDDR7 | 48 GB GDDR6 |
| Memory bandwidth | 1.792 TB/s | 0.864 TB/s |
| Power (TDP) | 575 W | 300 W |
| Form factor | PCIe consumer | PCIe dual-slot FHFL |
Cloud prices today
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 5090 Founders / reference | L40 48GB | |
|---|---|---|
| Lowest price | $0.25 | $0.79 |
| Median price | $0.54 | $0.86 |
| Highest price | $0.99 | $1.25 |
| Median per GB of memory | $0.017 | $0.018 |
| Public offers (all terms) | 7 | 13 |
| Providers | 3 | 5 |
Where to rent each
Lowest tracked price per provider. Open a provider to see every offer, region and term.
Questions
Which is cheaper to rent, the GeForce RTX 5090 Founders / reference or the L40 48GB?
Right now the cheapest on-demand GeForce RTX 5090 Founders / reference Kovara tracks is $0.25 per GPU-hour and the cheapest L40 48GB is $0.79. Medians are $0.54 and $0.86. The cheaper GPU per hour is not always cheaper per job: a faster or larger GPU can finish sooner or need fewer units.
Which has more memory?
The L40 48GB has more memory: 48 GB against 32 GB. Memory decides which models and batch sizes fit on one GPU.
Are these benchmark results?
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.