The RTX 4090 is already the fastest graphics card in the world, making it a great option not only for gamers, but also creators, AI applications, and professionals. However, Nvidia is now working on its next-gen GPU lineup, and a new leak has now seemingly revealed a few details about the RTX 5090, which is expected to be the top dog in the RTX 50 series.

The information comes from reputable tipster Panzerlied over at the Chiphell forums, and offers us some insight into Nvidia's next-gen Blackwell series graphics processors. Most notably, the tipster claims that the RTX 5090 will be 1.7x faster (than the RTX 4090), which should make it an impressive upgrade. The post also suggests that the next-gen flagship will pack 50 percent more 'scale' (believed to be cores), 78 percent more (L2) cache, 52 percent more memory bandwidth, and 15 percent higher frequency than its predecessor.

If the rumors are accurate, the RTX 5090 could come with 192 SMs or 24,576 FP32 cores, given that the RTX 4090 offers 128 SMs and 16,384 FP32 cores. In terms of memory bandwidth, the RTX 5090 could come with 1,532 GB/s of effective bandwidth as opposed to the 1,008 GB/s in the 4090. The rumored 78 percent additional cache could mean that the next-gen card would have 128MB of L2 cache compared to the 72MB in the RTX 4090.

Finally, the post also claims that the frequency of the new card will go up by 15 percent, which suggests that the RTX 5090 could have a boost clock of around 2.9GHz. For comparison, the RTX 4090 has a base clock of 2.23 GHz and a boost clock of 2.52 GHz. Overall, the rumors paint the picture of a very powerful card, but given that the RTX 50 series launch is still a long way away, it would be prudent at this stage to take all the leaks and rumors with a generous pinch of salt.