We’ve been hearing rumors about Blackwell, the architecture that NVIDIA will use in the GeForce RTX 50, for quite some time, and it’s curious because the green giant hasn’t finished launching the GeForce RTX 40 series yet. I understand that you may question the credibility of these rumors, but we have to keep in mind , that company never stops working on developing new architecturesand so it’s only natural that it’s immersed in Blackwell while it completes the launch of its current line of graphics cards.

Blackwell will be the architecture that replaces Ada Lovelace, as we said, will be the basis of the GeForce RTX 50. This means that it is still at least a year and a half away, so there could be changes in the core configuration developed by NVIDIA. However, where we don’t expect major changes is the foundation or hard core of this architectureand that’s where they show the latest information.

In theory, Blackwell will use the draft monolithic core and that it will retain the same foundation we saw in Ada Lovelace. This means that the GeForce RTX 50 will have a classic division into shaders, RT cores for ray tracing and tensor cores for AI acceleration. We can also expect a new generation “Optical Flow Accelerator”.

GeForce RTX 50 will be built on TSMC’s 3nm nodeand due to this significant reduction of the process vs. Ada Lovelace, which is built on TSMC’s 5nm node, we can expect a noticeable improvement in performance with lower consumptionand also a major increase to the maximum number of shaders in the most powerful models of this new graphics generation.

It’s all good news, but we have to give you some bad news, and that’s it TSMC’s 3nm node is 25% more expensive than the 5nm nodewhich means that the price of next-generation graphics cards could go up again, although in the end everything will depend on the market situation and the competition that then prevails between NVIDIA, AMD and Intel, who will continue to fight for a niche in the sector.

Going a little deeper into the Blackwell architecture, it is said that while NVIDIA will maintain this division into shaders, RT cores, and tensor cores, it will also introduce major “structural” changes in SM unitswhich will strengthen its commitment to ray tracing and improve noise reduction technology, likely to give a increased support for patch tracking (layout of communications).

For the rest, we can expect to use the standard PCIe Gen5 and jump into a new one GDDR7 memory, which will significantly improve bandwidth thanks to its frequencies of up to 36 GHz. There’s still a lot up in the air, but I’m sure these first brushstrokes will give you an idea of ​​what we can expect from the GeForce RTX 50, a generation that could break the 3GHz barrier and double the performance of the current generation.