| Header Cell - Column 0 | RTX 5080 | RTX 5090 | RTX 5070 Ti | RTX 5070 |
|---|---|---|---|---|
| CUDA cores | 10752 | 21760 | 8960 | 6144 |
| Boost clock (GHz) | 2.62 | 2.41 | 2.45 | 2.51 |
| Base clock (GHz) | 2.3 | 2.01 | 2.30 | 2.16 |
| Tensor core TOPS | 1801 | 3352 | 1406 | 988 |
| Ray tracing core TFLOPS | 171 | 318 | 133 | 94 |
| Memory | 16 GB GDDR7 | 32 GB GDDR7 | 16 GB GDDR7 | 12 GB GDDR7 |
| Memory bus width | 256-bit | 512-bit | 256-bit | 192-bit |
| Total Graphics Power (watts) | 360 | 575 | 300 | 250 |
| Required system power (PSU wattage) | 850 | 1000 | 750 | 650 |
| Power connector | 1x 450 W PCIe Gen 5 OR 3x PCIe 8-pin adapter | 1x 600 W PCIe Gen 5 OR 4x PCIe 8-pin adapter | 1x 300 W PCIe Gen 5 OR 2x PCIe 8-pin adapter | 1x 300 W PCIe Gen 5 OR 2x PCIe 8-pin adapter |
| Price | $999 | $1,999 | $749 | $549 |
| Release date | Jan 30, 2025 | Jan 30, 2025 | Feb 2025 | Feb 2025 |
The second of Nvidia's RTX Blackwell series of graphics cards, the RTX 5080, was revealed at the Nvidia CES 2025 keynote, alongside the RTX 5090, RTX 5070 Ti, and RTX 5070.
With the promise of twice the performance of the RTX 4080 and a price that matches the RTX 4080 Super, the second-string RTX Blackwell GPU looks pretty tasty.
In real terms, however, it's the might of DLSS 4 and Multi Frame Generation that is going to be held up as the feature which gives it that double RTX 4080 gaming performance, and only then in the games which support it.
The RTX 5080 release date puts it alongside the RTX 5090 with both cards launching on the same day, January 30, 2025. Like the RTX 5090 I would expect we'll see AIB cards being released on January 31, with traditionally chonky shrouds.
This has been the most controversial thing about the rumoured specs; the GPU make up of the RTX 5080. The RTX 5080 specs look remarkably like those of the RTX 4080, and you've got to be hoping that Blackwell architecture is able to deliver something special—and not just DLSS 4—when it comes to gaming frame rates, because this doesn't look special. Especially the now-confimred specs of the 256-bit bus and just 16 GB VRAM.
The original RTX 4080 launched at a ludicrous $1,200. A price so ludicrous, in fact, that I had to just go look it up again to make sure I was remembering it right. That changed with the RTX 4080 Super after significant backlash over the first card's pricing, and that came out at $999. Thankfully Nvidia hasn't gone overboard on the pricing of the RTX 5080 and has stuck with the $999 price point.
Latest about Nvidia GeForce RTX 5080
Intel is reportedly in talks to spin off its chip factories into a partnership with arch rival TSMC and now I think I've seen everything
News Next week, cats and dogs living together.
Nvidia's RTX 5070 Ti GPU officially goes on sale February 20 and the RTX 5070 is go for March 5
News So, the 5070 has indeed been delayed a little.
Zotac beats those dastardly GPU scalpers by selling RTX 50-series graphics cards to actual gamers courtesy of their Discord channel
News So long as it's not just using this as a way to fluff its engagement numbers, and really is putting GPUs in the hands of loyal customers, we're here for it.
TSMC reportedly plots 2027 start date for its 3 nm US fab, but will that be in time to save next-gen GPUs from tariffs?
News Tariff fears said to be accelerating TSMC's 3nm plans.
Nvidia's RTX 5070 graphics card rumoured to be delayed from February until March but that could actually be good news
News A bit more time to build up decent stock levels before launch?
Surely not again: Worrying analysis shows Nvidia's RTX 5090 Founders Edition graphics card may be prone to melting power connectors
News After the melting RTX 4090 debacle, surely Nvidia is all over this stuff?
There is one thing making me excited about the new Nvidia RTX 5070 Ti and no, it isn't Multi Frame Generation
Push it If the GPU at the heart of the RTX 5070 Ti can deliver the same level of overclocking it does with the RTX 5080 we could have a champ on our hands.
I'm genuinely stunned by the overclocking performance of the RTX 5080, and curious as to why Nvidia left so much headroom
OC hero Maybe I won the silicon lottery, but the GB203 sure looks like there was a lot of performance left on the table.
