Nvidia recently announced their new RTX 50-series GPUs claiming double the performance with the RTX 5090 and saying that the more modest RTX 5070 could equal the performance of the last-generation RTX 4090.
The RTX 4090 debuted in October 2022 and has an eye-watering price tag of $1,600 and is sold out almost everywhere. In contrast, the RTX 5070 will be released in late February with a suggested retail price of $550.
On paper, the RTX 5070 has a third less memory bandwidth, half the physical memory, and barely a third of the CUDA core count with comparable boost clock speeds than the RTX 4090.
However, Nvidia’s claims largely center around the enhanced support for AI-driven upscaling using the latest generation of Tensor cores, with the RTX 5070 featuring multi-frame generation.
Until testing is done, it is unclear how good this next-generation card is or how it stacks up against the RTX 4090, but Nvidia’s slides suggest that the RTX 5070 may be less than 20% faster than the RTX 4070.
Numerous factors, including DLSS 4 support and its new architecture, process node, and DLSS-heavy marketing, make it challenging to say for sure how good the RTX 5070 really is.
It seems unlikely that the RTX 5070 can match the RTX 4090's performance in most games, and heavy multi-frame generation can introduce latency issues and visual artifacts unlikely to be loved by everyone.
Nonetheless, the RTX 5070's reduced power usage and new architecture might enable it to close the performance gap with the RTX 4090, making it appropriate for small form-factor gaming PCs.
If you're currently using the RTX 4090, it is unlikely that the RTX 5070 will measure up to its performance, so beware of the FOMO of a new generation.
Wait for real-world testing to settle this performance debate until the RTX 5070 is released next month.