Nvidia’s RTX 5070 Ti is reportedly set to include 16GB of GDDR7 memory

In contrast to the 5080, which is said to utilize 32 Gbps GDDR7 memory, the 5070 Ti is rumored to incorporate slower 28 Gbps GDDR7 chips, which results in lower memory bandwidth. Nevertheless, this provides the GeForce RTX 5070 Ti with a 33% increase in memory bandwidth compared to the 4070 Ti SUPER that it is designed to succeed.

For those tracking developments, this isn’t the first mention of the RTX 5070 Ti. Anticipations are that the RTX 5070 Ti will feature 8,960 CUDA cores along with a significant power rating of 300W. However, according to Wccftech’s sources, those figures have undergone some revisions. They claim that the RTX 5070 Ti will utilize a scaled-down variant of the RTX 5080, relying on the GB203-400 SKU. Additionally, the power consumption has been increased to an astounding 350 watts, which is rather surprising for an XX70-class GPU. It’s plausible that this is merely one of the configurations Nvidia is currently testing, though that is yet to be determined.

Kopite7kimi has also weighed in, suggesting that their latest information points toward a more moderate total board power of 285 watts, consistent with its predecessor.

Nvidia is scheduled to announce its RTX 50 series of graphics cards at CES 2025. The CES keynote is set for January 6th, with CEO Jensen Huang presenting the session. During this event, we anticipate Nvidia will unveil their premium RTX 50 series GPU lineup. At present, the specific advantages for gaming that Nvidia has integrated into the RTX 50 series (Blackwell) graphics architecture remain unclear. We expect Nvidia to enhance AI capabilities, as well as make further strides in ray tracing technology. However, it is still uncertain whether Nvidia intends to introduce a new iteration of DLSS alongside their RTX 50 series.

And as always, while waiting for more exciting news, remember to use our comparison tool to find the best deal for your next favorite game.