RTX 5090 vs RTX 5080: Is the Extra $1,000 Worth It?

NVIDIA's RTX 5090 and RTX 5080 launched the same day — January 30, 2025 — at $1,999 and $999 respectively. The actual gap between them is more lopsided than the price tag suggests.

The hardware gap

The 5090 packs 21,760 CUDA cores against the 5080's 10,752 — barely short of double. It pairs that with 32GB of GDDR7 on a 512-bit bus versus the 5080's 16GB on a 256-bit bus, so memory bandwidth roughly doubles too. The 5080 actually boosts slightly higher (2.62 GHz vs. 2.41 GHz) — the gap is almost entirely core count and memory width.

Real-world scaling

That translates into a genuinely large gap at high resolution. GamersNexus's testing put the 5090 as much as 62% ahead of the 5080 in Final Fantasy XIV at 4K, with the 5090's lead ranging from roughly 30% to nearly 69% across their 4K test suite depending on the title.

Power draw

The 5090's 575W TDP against the 5080's 360W is a 215W gap — NVIDIA recommends a 1000W system PSU for the 5090 versus 850W for the 5080.

Who should actually buy which

If you're building around a 4K high-refresh monitor, particularly with ray tracing and path tracing pushed to their limits, the 5090's lead is large enough to matter. If you're gaming at 1440p, you are very likely to be CPU-bound before you're GPU-bound on a 5080 — the 5090's extra headroom goes largely unused at that resolution. The 5090 also functions as a prosumer/AI-workload card given its 32GB VRAM pool.

What we wouldn't recommend

Buying the 5090 purely "to be safe" for a 1440p setup. For a straight gaming build tied to a 1440p or lower monitor, that extra $1,000 buys headroom you're unlikely to touch for the life of the card. Neither card should be assumed to be in stock at MSRP — always check current retailer pricing before buying.