RTX 4080 vs RTX 5090
The cheapest RTX 4080 ($0.11) costs 50% less per hour than the cheapest RTX 5090 ($0.22).
Per unit of performance, the RTX 5090 is 14% cheaper: $0.24 versus $0.27 per A100-equivalent hour.
The RTX 5090 has 32 GB of memory against 16 GB on the RTX 4080.
| Measure | NVIDIA GeForce RTX 4080 | NVIDIA GeForce RTX 5090 |
|---|---|---|
| Lowest per GPU-hour | $0.11 | $0.22 |
| Median per GPU-hour | $0.44 | $0.87 |
| On-demand from | $0.12 | $0.34 |
| Spot from | $0.11 | $0.25 |
| Reserved from | — | $0.22 |
| Lowest per A100-equivalent hour | $0.27 | $0.24 |
| Live offers | 38 | 405 |
| Providers | 15 | 84 |
| Memory | 16 GB GDDR6X | 32 GB GDDR7 |
| Memory bandwidth | 717 GB/s | 1,792 GB/s |
| FP16 / BF16 | 97.5 TFLOPS | 209.5 TFLOPS |
| FP8 | 194.9 TFLOPS | 419 TFLOPS |
| Performance vs A100 80GB | 0.4× | 0.92× |
| Architecture | Ada Lovelace | Blackwell |
| Power | 320 W | 575 W |