NVIDIA continues to work on Blackwell, the architecture that will give life to the GeForce RTX 50, a new generation of high-performance graphics cards for the general consumer market, which will be launched during fourth quarter of 2024. The launch will be staggered, meaning the most powerful models will arrive first and the lower ranges will arrive in later months.

With that in mind, it’s very likely that NVIDIA will follow the strategy we saw with the GeForce RTX 40, which is to launch the GeForce RTX 5090 and GeForce RTX 5080 together, and both will be available between months October and November. The GeForce RTX 5070 Ti and GeForce RTX 5070 should arrive in the first quarter of the year, and the GeForce RTX 5060 Ti and GeForce RTX 5060 should debut in the second quarter.

Although the introduction of the GeForce RTX 50 will take place much earlier, NVIDIA will use the CES 2025 scenario to talk about the most important keys of this new generation at the architectural level.

It will be one of the most important advances that the green giant is going to make with this new generation of graphics cards efficiency, which will be much better than what we saw with Ada Lovelace, and frankly, those are big words because all the models that make up this generation simply a fantastic ratio of power to consumed watt.

In terms of performance, it is said that the generational leap that the GeForce RTX 50 will see over the GeForce RTX 40 it won’t be that big like the one we saw between these and the GeForce RTX 30, at least in rasterization. This leads me to believe that NVIDIA could focus its efforts primarily on improving ray tracing performance and may also surprise us with some important new developments in its DLSS ecosystem.

It makes sense, especially given NVIDIA’s commitment to full ray tracing, also known as path tracing. However, DLSS is also a must to improve performance and counteract the impact that ray tracing has when used in games, so I think it’s very likely we’ll see developments on both fronts.

We still don’t have really reliable information about the possible specifications of the GeForce RTX 50, although everything indicates that this new generation of graphics cards it will use TSMC’s 3nm node and have GDDR7 memory in its most powerful models. Top entry-level and mid-range like the GeForce RTX 5070 and below could continue to use GDDR6X memories.