Two of the names of new AMD Radeon graphics cards appear to have been confirmed. A piece of code found on Github reveals that, at least internally, AMD plans to stick with its current naming for at least this generation.
The new AMD Radeon RX 8800 and 8600 are directly referenced in a patch for AMD’s ROCm software. This is a software stack meant for artificial intelligence projects, an area where the company is keen to make a splash. While the best graphics card options for AI have typically been based on Nvidia GPUs, AMD has been pushing ahead to ensure it’s not entirely left behind.
As ROCm is developed openly, any updates and patches can be easily scanned for any tidbits on upcoming hardware. Found by user @johnnync13 on X (formerly Twitter), the patch appears to be based on focused on introducing FP8 (8-bit floating point) support to AMD GPUs. This data format is already supported by Nvidia’s Ada and Hopper GPUs for AI computing, providing more precision than INT8 (8-bit integer) data.
AMD’s next lot of hardware will presumably come with better support for AI projects, but in terms of gaming, we’ve already seen leaks concerning these upcoming GPUs.
According to one recent report, the AMD Radeon RX 8800 XT could match the Nvidia GeForce RTX 4080 in some areas, including ray tracing. Performance should also see some gains over the last generation, with hopes that the top new GPU will be priced around the $500 mark.
AMD’s current GPUs in this price bracket have been received well in our RX 7800 XT review, as well as our RX 7900 GRE review, but Nvidia is still dominating in terms of market share. With CES 2025 coming up, we could see these new GPUs being fully revealed very soon, and we’re also expecting to see Nvidia lift the lid on the RTX 5000 series.
While AMD is gearing up for its next graphics cards, it’s completely taken over the best gaming CPU space with the Ryzen 7 9800X3D in the meantime. The GPU market might be very green with Nvidia, but AMD has painted the heart of PC gaming bright red.