Loading your page…
You can continue using the navigation while this page loads.
You can continue using the navigation while this page loads.
Every public NVIDIA GeForce RTX 5090 Founders / reference offer Kovara tracks on TensorDock, normalised to US dollars per GPU-hour, with regions, terms and how it compares with the rest of the market.
Sorted by price per GPU-hour, including spot, reserved and serverless terms. Open an offering to see its source, availability evidence and to request it.
| Offering | Location | Price / GPU-hour | Terms | Last checked |
|---|---|---|---|---|
| OFR-0253 | Location not specified | $0.60 | On Demand · Marketplace GPU VM | 29 Sept 2026 |
Prices come from public provider sources and are converted to USD per GPU-hour. They can change at any time; check the offering page for the source and date.
Lowest on-demand GeForce RTX 5090 Founders / reference price by provider, across every location. TensorDock is highlighted.
| Provider | Lowest on-demand / GPU-hour | Offers | vs. market median |
|---|---|---|---|
| SaladCloud | $0.25 | 4 | 45% below |
| TensorDock | $0.60 | 1 | 31% above |
| Runpod | $0.99 | 1 | 116% above |
Market median: $0.46 per GPU-hour across 6 on-demand offers from 3 providers.
Manufacturer specifications recorded by Kovara. See the full profile for sources and related models.
Kovara currently tracks 1 public offer for the NVIDIA GeForce RTX 5090 Founders / reference on TensorDock. On demand, prices run from $0.60 to $0.60 per GPU-hour. Prices are normalised to US dollars per GPU per hour so they can be compared across providers.
Its lowest on-demand price is 31% above the median on-demand GeForce RTX 5090 Founders / reference price across the 3 providers Kovara tracks ($0.46), and it ranks 2 of 3 by lowest on-demand price. Terms, region and capacity differ between offers, so compare like with like.
No. A published price shows what a provider charges, not that GPUs are free right now. Kovara shows availability evidence separately on each offering and can help you confirm capacity through a quote request.