Comparison · GPU vs GPU · read 21:35 UTCH100 vs RTX 5090
Specifications, today's prices from 101 providers, price per unit of memory and compute, and who rents each cheapest. Better values are highlighted; lower is better for prices.
Side by sideSpecs, prices, value
Highlighted cells are the better of the two: more memory and compute, lower prices, lower cost per unit.
| H100 | RTX 5090 | Difference | |
|---|---|---|---|
| Memory & compute | |||
| VRAM | 80 GB | 32 GB | +150% |
| Memory bandwidth | 3.35 TB/s | 1.79 TB/s | +87% |
| FP16 tensor | 989 TFLOPS | — | — |
| FP32 | 67 TFLOPS | 104 TFLOPS | -36% |
| Power | 700 W | 575 W | +22% |
| Prices today | |||
| Cheapest on-demand | $1.35 | $0.24 | +474% |
| Median on-demand | $3.17 | $0.69 | +359% |
| Highest on-demand | $8.98 | $1.88 | +378% |
| Cheapest spot | $0.67 | $0.10 | +567% |
| Cheapest reserved | $1.08 | $0.38 | +184% |
| Current offers (n) | 189 | 58 | +226% |
| Providers with a fresh offer | 76 | 25 | +204% |
| Value | |||
| Per GB of VRAM · hour (cheapest) | $0.017 | $0.007 | +130% |
| Per GB of VRAM · hour (median) | $0.040 | $0.022 | +84% |
| Per PFLOP FP16 · hour (median) | $3.201 | — | — |
| Per TB/s · hour (median) | $0.945 | $0.385 | +145% |
| Trend | |||
| 30-day median change | +3.7% | +29.7% | — |
| 90-day median change | +9.2% | — | — |
90 days · median on-demand price · rebased to 100How the two have moved
Both medians indexed to 100 ninety days ago, so the slopes compare even though the price levels differ.
Cheapest providers today · on-demand · current offersWho rents each one cheapest
- 1
GPUBankTrust 931× $1.35 - 2
GcoreTrust 801× $1.71 - 3
LiumTrust 408× $1.79
- 1
SaladTrust 671× $0.24 - 2
Clore.aiTrust 461× $0.33 - 3
HyperAI1× $0.35
Verdict · computed from the numbers aboveChoose one, choose the other
Choose the H100 if…
- your model needs more than 32 GB per GPU: 80 GB against 32 GB.
- you want choice: 76 providers have a fresh offer today against 25.
Choose the RTX 5090 if…
- the hourly price matters most: median $0.69 against $3.17, 78% less.
- you serve LLMs: memory bandwidth per dollar is better ($0.385 per TB/s-hour against $0.945).
Questions about H100 vs RTX 5090
Which is cheaper to rent, the H100 or the RTX 5090?
Today the cheapest current on-demand H100 is $1.35 per GPU-hour and the cheapest RTX 5090 is $0.24; medians are $3.17 and $0.69. Cheapest excludes offers marked unavailable and counts one floor per marketplace.
How much memory do the H100 and RTX 5090 have?
80 GB of HBM3 on the H100, 32 GB of GDDR7 on the RTX 5090, at 3.35 and 1.79 TB/s.
Where do these numbers come from?
From 247 current offers read on the providers' own pages and APIs over the last 45 days (reference 21:35 UTC), or in our October 2026 market study for the few providers our collector does not read. Specifications come from manufacturer datasheets. Every price on the GPU pages links to its receipt and shows its read date.