RTX 5070 vs RTX 5080

One card is built for 1440p and value, the other for 4K headroom you'll pay handsomely for. Here's the honest gap between them, with real NZ pricing.

The mainstream 1440p card against the true high-end

The RTX 5070 and RTX 5080 sit either side of the biggest practical fork in NVIDIA's 50-series stack: one is built for 1440p gaming at a price most people can actually justify, the other is built for 4K and pushes into genuinely premium pricing. On paper the gap looks modest — in practice, and in your wallet, it isn't.

Specs, side by side

Performance gap

The 5080 has 75% more CUDA cores and a third more VRAM, and that shows up as roughly a 20-30% real-world performance lead, with the gap widening at 4K where the 5070's 12GB of VRAM and narrower memory bus start to become the limiting factor rather than raw shader throughput. At 1440p, the 5070 is comfortably capable of high frame rates in almost every current game at high-to-ultra settings, with DLSS 4's Multi Frame Generation available on both cards to push frame rates higher again in supported titles.

Price-to-performance

This is the part that actually matters for most buyers. At roughly double the price for 20-30% more performance, the RTX 5080 is objectively the worse deal on a pure dollars-per-frame basis. That's normal for a tier jump — the 5070 is, and is meant to be, the significantly better value card. The 5080 earns its price only if you specifically need what the extra VRAM and shader count unlock: consistent 4K gaming, or heavier content-creation/AI workloads where 16GB of VRAM avoids running out of headroom mid-project.

VRAM — the detail that actually bites

12GB is enough for 1440p gaming today and for the next couple of years in most titles, but some recent releases with high-resolution texture packs and ray tracing enabled are already pushing past 10-11GB at 1440p with everything maxed. It won't cripple the 5070, but it does mean dropping a texture setting a notch sooner than a 5080 owner would need to. If you also do any GPU-accelerated creative work — Premiere, DaVinci Resolve, Stable Diffusion/local AI image or video generation — the 5080's extra 4GB buys real headroom, not just a bragging-rights number.

Power and thermals

The 5070's 250W TDP means it runs happily in a wider range of existing PC cases and power supplies (a solid 650-750W PSU is plenty). The 5080's 360W draw wants a beefier PSU (850W+ recommended once you account for a modern CPU alongside it) and a case with genuinely good airflow — not a dealbreaker, but a real line-item cost if your current PC doesn't already meet that spec.

Verdict

Buy the RTX 5070 if: you're gaming at 1440p (or 1080p at very high frame rates), you want the best price-to-performance in NVIDIA's current lineup, and you don't do VRAM-hungry creative or AI work.

Buy the RTX 5080 if: you're gaming at 4K, you want meaningful headroom for the next few years of increasingly VRAM-hungry games, or you do serious content creation/local AI work where 16GB genuinely changes what you can run without hitting a wall.

Our honest take: for the vast majority of NZ gamers running a 1440p monitor, the RTX 5070 is the smarter buy — it's not close on value. Only step up to the 5080 if you specifically own or plan to own a 4K display, or your workload actually uses the extra VRAM.