The cheapest RTX 5090 on the table right now is $0.133 per GPU-hour at Vast.ai (spot), out of 16 offers from 11 providers. The RTX 5090 carries 32 GB of VRAM on NVIDIA's Blackwell generation. That is peer/marketplace capacity — rented from other users' machines, and not directly comparable to a dedicated instance.
On-demand pricing for the same card spans $0.250 to $1.19 per GPU-hour — a 4.8× spread between the cheapest and the most expensive provider for identical silicon. That gap is the entire reason this table exists.
Cheapest by pricing model: on-demand $0.250 at Salad Cloud · spot $0.133 at Vast.ai.
| Provider | Tier | $/GPU-hr | Pricing | Node | Commit | In stock | |
|---|---|---|---|---|---|---|---|
| Vast.ai | marketplace | $0.133 | spot | 2× GPU | — | yes | rent |
| Salad Cloud | marketplace | $0.250 | on-demand | 1× GPU | — | yes | rent |
| Vast.ai | marketplace | $0.347 | spot | 1× GPU | — | yes | rent |
| Vast.ai | marketplace | $0.413 | on-demand | 2× GPU | — | yes | rent |
| Vast.ai | marketplace | $0.414 | on-demand | 1× GPU | — | yes | rent |
| Vast.ai | marketplace | $0.439 | spot | 4× GPU | — | yes | rent |
| Vast.ai | marketplace | $0.466 | on-demand | 4× GPU | — | yes | rent |
| Novita | dedicated | $0.560 | on-demand | 1× GPU | — | — | rent |
| Akash | marketplace | $0.570 | on-demand | 1× GPU | — | no | rent |
| GPU.ai | dedicated | $0.590 | on-demand | 1× GPU | — | yes | rent |
| Octa | marketplace | $0.640 | on-demand | 1× GPU | — | yes | rent |
| Runpod Community | marketplace | $0.690 | on-demand | 1× GPU | — | yes | rent |
| Lium.io | marketplace | $0.750 | on-demand | 8× GPU | — | yes | rent |
| Spheron | dedicated | $0.780 | on-demand | 1× GPU | — | — | rent |
| Beam | dedicated | $1.09 | on-demand | 1× GPU | — | yes | rent |
| Runpod | dedicated | $1.19 | on-demand | 1× GPU | — | yes | rent |
How much does an RTX 5090 cost per hour?
As of the latest scrape, $0.133 per GPU-hour at Vast.ai is the cheapest RTX 5090 offer across 11 tracked cloud providers. The cheapest on-demand rate is $0.250 per GPU-hour at Salad Cloud, and the most expensive on-demand rate is $1.19 at Runpod.
Which cloud is cheapest for the RTX 5090?
Vast.ai at $0.133 per GPU-hour on spot capacity. Prices move constantly, so this page is regenerated from the live scrape every 15 minutes. Note that marketplace tiers are peer hardware and are not directly comparable to dedicated capacity.
Is spot RTX 5090 capacity cheaper?
Yes — the cheapest spot RTX 5090 is $0.133 per GPU-hour at Vast.ai, -47% versus the cheapest on-demand rate of $0.250. Spot instances can be reclaimed by the provider at any time, so they suit checkpointed training and batch inference rather than long-lived services.
How many GB of VRAM does the RTX 5090 have?
32 GB, on the Blackwell architecture. Multi-GPU instances multiply that: an 8× RTX 5090 node exposes 256 GB of GPU memory in total.