Nvidia's new RTX 5090, 5080, 5070 Ti and 5070 announced at CES – but is there a meaningful performance bump?

Gamers and professional digital creatives rejoice – Nvidia has announced its new 5000 series of GPUs, with CEO Jensen Huang revealing all at this year's CES tech event.

If spec sheets are to be believed, the top dog of the series – the RTX 5090 – could well become the best graphics card for gaming. Scrap that, it'll be the best graphics card out there, full stop. But we won't be able to say for sure until the 5090 releases on 30 Jan at the princely sum of $1999 (along with the 5080 at $999). And we can test the RTX 5070 Ti ($749), and the RTX 5070 ($549) in February.

With GPUs, it's all about price and performance. And at first glance, the 5000 series certainly promises a lot in both categories. Take the 5090 price tag: it was rumoured that it would sell at $2,500 late last year, so its $2,000 price tag is welcome news (however, whether that price rumour was leaked by Nvidia to make us all fell good about the final price is anyone's guess. $2000 is, after all, a big investment for a consumer GPU).

Price

For the rest of the series, all their prices are between $50 and $100 cheaper than the 4000 series equivalent – while promising twice the performance, according to Huang. In fact, the 5090 is the only one that will sell more expensive than its 4000 counterpart (see below). However, this is all in a perfect world, where scalpers don't exist and retail prices of GPUs are actually what you pay. The real cost of these bad boys are yet to be defined.

• RTX 4090: $1,599 | RTX 5090: $1,999
• RTX 4080: $1,199 | RTX 5080: $999
• RTX 4070 Ti Super: $799 | RTX 5070 Ti: $749
• RTX 4070 Super: $599 | RTX 5070: $549

Performance

When it comes to performance, the 5000 series promises a lot – the aforementioned doubled up performance compared to the 4000 series high on the list. You may know of the underwhelming 4000 series, with the 4070 and 4070 Ti having to be upgraded to 'Super' iterations, largely due to public outrage over the disappointing initial offering. With the announced promises, it doesn't look like there will be a similar debacle this time around.

The full series of GPUs will include a PCIe 5.0 interface, a 16-pin power connector, DisplayPort 2.1a, GDDR7 memory and DLSS 4 technology. Looking at the flagship 5090, you'll get 32GB of GDDR7 VRAM, and shiny new 5th Gen Tensor and 4th Gen Ray Tracing Cores. Add to that a 512-bit memory bus, a PCIe 5.0 x16 interface, and a zippy 1792 GB/s of memory bandwidth with a memory speed upwards of 23.8 Gbps, and you have yourself a GPU beast. Probably more of a beast than most people will actually need.

With these specs, the 5090 is better equipped for handling 8K than 4K, and as there are no games that run at 8K, this GPU is squarely aimed at professional creative studios – not casual (or even professional) gamers.

For the other iterations in the 5000 series, twice the performance for $50 to $100 less than the last generation sounds pretty great. Lets see how they all perform in February.

For more on Nvidia's new tech, read our impressions of the new DLSS 4 and the AI-driven Multi-Frame Generation tech set to improve gaming and 3D modelling.