Nvidia Geforce RTX 5080: Specs, Release Date and What We Know So Far
After months of agonizing anticipation, Nvidia has finally announced the RTX 5080, along with the rest of the Blackwell lineup, including the RTX 5090, RTX 5070 Ti and RTX 5070 at CES 2025. We'll finally be able to get our hands on the next-generation graphics card on January 30. Until then, Nvidia has revealed the full specs of the card so we can get a rough idea of what to expect when it makes its way into a gaming PC near you.
Nvidia RTX 5080 release date
The Nvidia GeForce RTX 5080 launches January 30, 2025, along with its bigger sibling, the RTX 5090. Nvidia has also announced the RTX 5070 and RTX 5070 Ti, though those don't have a definite release date – though we can expect them by March.
As for the laptop version of the RTX 5080, Nvidia claims availability will 'start in March', though that is going to largely depend on the laptop manufacturers. it could be April before we see the likes of Alienware, MSI and Asus work the RTX 5080 into their next-generation laptops.
Nvidia RTX 5080 price
When Nvidia unveiled the RTX 5080, it revealed a starting price of $999 for the RTX 5080, with third party cards likely being much more expensive, depending on how fancy their coolers and features are.
While I don't know how likely it'll be to get an RTX 5080 for $999 when it hits the street, it is a significantly lower launch price than the RTX 4080, which launched for $1,199 back in 2022. That's surprising, when you consider the RTX 5090 saw a price jump from $1,599 to $1,99 – also a starting price.
As for the lower-tier cards, the RTX 5070 Ti will start at $749, with the RTX 5070 starting at $549.
Getting a gaming laptop with an RTX 5080 is going to be quite a bit more expensive, of course, as you're buying an entire system instead of a single component. During the keynote at CES, Nvidia claims systems will start at $2,199, with more premium systems likely getting a substantial price bump. With these gaming laptops, though, keep in mind that they'll be much less performant than the equivalent desktop GPU. My general rule of thumb, without seeing testing, is that the laptop GPU is the equivalent of two tiers down. So, for instance, the RTX 5090 mobile will likely perform at the level of the desktop RTX 5070, with the RTX 5080 likely matching a desktop RTX 5060 – even if that hasn't even been announced yet.
Nvidia RTX 5080 specs
The Nvidia GeForce RTX 5080 is built on the Blackwell architecture that Nvidia's been using to power Supercomputers for the past year or so. While I'm not lucky enough to to test a data center GPU in Cyberpunk, Nvidia is making some lofty claims about the performance of this architecture, especially when it comes to AI performance, which is important for upscaling in modern PC games.
This graphics card sports 10,752 CUDA cores across 84 Streaming Multiprocessors (SMs). That's actually a raw increase over the RTX 4080, which only sported 9,728 shaders. Assuming each Blackwell-based CUDA core has a significant IPC improvement over their last-gen counterparts, this increase in cores could mean significantly better performance.
Of course, each SM has more than just CUDA cores. Nvidia hasn't released the chip layout, but assuming Blackwell has a similar layout to Ada Lovelace, each SM should have 4 Tensor Cores, which would make for a total of 336 Tensor Cores. Each SM also features a RT Core, which powers Ray Tracing. Nvidia is claiming a theoretical 1,801 TOPS of AI performance through the Tensor Cores and 171 Teraflops of ray tracing performance through the RT cores.
Finally, the RTX 5080 sports 16GB of GDDR7 memory on a 256-bit bus. Because the RTX 5000 series are the first graphics cards to ever use GDDR7, I have no idea what impact this will have on performance, but it should be much faster than the GDDR6X on the RTX 4080 – though only time will tell.
Nvidia RTX 5080 performance
When Jensen Huang took the stage at CES 2025 with his flashy new jacket, he made some lofty claims about RTX 5090 performance, and even claimed that the RTX 5070 would match the RTX 4090. he supported these claims with benchmarks using the new DLSS 4, which coincidentally won't run on RTX 4000 cards, so you should take them with a grain of salt.
The truth of the matter is that I have no idea how fast these graphics cards are, and I won't have a clear picture until I get them in the lab to actually test them in a controlled setting. Nvidia also made really lofty claims of gen-on-gen performance when it launched the RTX 4080, and that didn't turn out so well for Team Green. Luckily, with the RTX 5080 launching on January 30, we won't have to wait long to see what they have in store.
Jackie Thomas is the Hardware and Buying Guides Editor at IGN and the PC components queen. You can follow her @Jackiecobra