NVIDIA GeForce RTX 4090 vs NVIDIA GeForce RTX 5080
Our index puts the NVIDIA GeForce RTX 4090 roughly 29% ahead of the NVIDIA GeForce RTX 5080. This is an estimate for ranking, not a measured benchmark, and the gap varies considerably by title and by resolution.
| Specification | NVIDIA GeForce RTX 4090 | NVIDIA GeForce RTX 5080 |
|---|---|---|
| Manufacturer | NVIDIA | NVIDIA |
| Architecture | Ada Lovelace | Blackwell |
| Relative performance index | 270 (better) | 210 |
| Video memory | 24 GB (better) | 16 GB |
| Memory type | GDDR6X | GDDR7 |
| Memory bus | 384-bit (better) | 256-bit |
| Release year | 2022 | 2025 (better) |
| Board power | 450 W | 360 W (better) |
| Suggested PSU | 850 W | 850 W |
| Hardware ray tracing | Yes | Yes |
| Upscaling support | DLSS 3, FSR 3, XeSS | DLSS 4, FSR 3 |
| DirectX | 12 Ultimate | 12 Ultimate |
Green marks the stronger figure for that row only — for power draw and PSU requirement, lower is treated as better. Performance indexes are CanMyPcPlay estimates; read how they are derived.
Which should you pick?
- Check video memory, not just speed. A faster card with less VRAM can perform worse in recent titles at high texture settings.
- Check power and clearance. Confirm your PSU and case can take the card — the build checker does this.
- Upscaling support differs. DLSS is NVIDIA-only; FSR and XeSS run more broadly.