Sources within the Chinese supply chain assure that NVIDIA has opted for a very different memory configuration on your RTX 30s. As our regular readers will know, the green giant has maintained a conservative bet in this regard, we just have to see that the RTX 2070, RTX 2080 and RTX 2080 Ti, as well as their Super variants, have the same amount of graphics memory as the GTX. 1070, GTX 1080 and GTX 1080 Ti, 8 GB the first two and 11 GB the last two.

A few days ago we saw that AMD was working to say goodbye, finally, to high-end graphics cards with 8 GB of graphics memory, a move that, frankly, seems very successful, since it is likely that we will end up seeing a considerable increase in the consumption of graphics memory during the middle of the life cycle of next-gen consoles, and that consumption increases even more if Sony and Microsoft decide to launch a PS5 Pro and an improved Xbox Series X. You already know how this is, in the end, video game developments start from consoles and their base at the hardware level.

Returning to the RTX 30 series, the source of the news indicates that NVIDIA plan to use the following memory configurations on your upcoming high-end graphics cards:

TITAN RTX: 24 GB of 21 GHz GDDR6X memory on a 384-bit bus. It will have 5,376 shaders.
RTX 3080 Ti: 20 GB of 21 GHz GDDR6X memory on a 384-bit bus. This model is rumored to have 5,248 shaders.
RTX 3080: 10GB of 18GHz GDDR6X memory on a 320-bit bus. It could have 4,352 shaders.
RTX 3070 Ti: 8GB of 18GHz GDDR6X memory on a 256-bit bus. It would have 3,072 shaders.
RTX 3070: 8 GB of 16 GHz GDDR6 memory on a 256-bit bus. It would have 2,944 shaders.

If these settings are confirmed we will see a significant increase in the amount of available memory of the RTX 3080 Ti compared to the RTX 2080 Ti, nothing more and nothing less than 9 GB more. However, on the RTX 3080 the increase would be much more modest, just 2 GB more compared to the RTX 2080. In the case of the RTX 3070 Ti and RTX 3070 there would be no increase in graphics memory.

I remind you that the RTX 30 will bring important improvements at the architectural level thanks to the adoption of Ampere, although it is said that the most interesting news will be centered around ray tracing (they could multiply performance up to four times of Turing), and in the support of the DLSS 3.0 technology, which promises to take intelligent rescaling to an unimaginable level, thanks to the use of innovative features such as intelligent compression of textures and other graphic elements to reduce memory consumption.

If all goes according to plan, the presentation of the RTX 30 will take place next September 1.