AMD Radeon RX 7000generating graphs created with the architecture RDNA3, will be officially unveiled on November 3, 2022, with a rumored launch date set for mid-December of that year. Added to this is the possibility that the red giant will do the same as NVIDIA and start releasing the highest models during 2023 to bring the mid and lower ones to market.

The highest known RX 7000 models, for now, are referenced by their dice (ten), Navi 31 and Navi 32 will use MCM design and TSMC’s 5nm manufacturing process. The RDNA 3 architecture aims to offer more than a 50% improvement in performance per watt compared to RDNA 2, but unfortunately for Radeon, the latest rumors don’t rate it very well compared to the RTX 40.

AMD seems to have decided to introduce more infinite cache in its future Radeon RX 7000, which could leave it with little room for improvement in aspects such as ray tracing performance. Here we find ourselves in the same situation as in recent years, namely that NVIDIA is apparently two generations ahead of Radeon in technology. Furthermore, the improvement seen by the RTX 40 would place the RTX 40 well above the RX 7000 in rasterization performance.

Bear in mind that RX 7000 performance predictions are just rumors for now, but the blow to the table that the NVIDIA GeForce RTX 40 meant is intended to make it difficult for AMD, as even Intel is, at least at this moment, higher in terms of technology. In other words a Since Intel is not able to properly cover all ranges, it is currently a rival with more potential than AMD to counter the strong dominance exercised by NVIDIA..

In short, it looks like the RX 7000 will be a significant improvement compared to the previous generation of Radeon graphics, but its evolution might not be enough to compete with the guarantees with the RTX 40. This would force AMD to play easy with the resource, the price, because otherwise its future dedicated graphics will hardly be attractive to consumers.