Thanks to kopite7kimi, the leaked specifications of NVIDIA's upcoming GeForce RTX 5070 and RTX 5070 Ti graphics cards have been finalized.
Starting with the RTX 5070 Ti, it will come with 8,960 CUDA cores and 16 GB of GDDR7 memory on a 256-bit memory bus, providing a bandwidth of 896 GB/s. The card is said to have a total board power (TBP) of 300 W. The Ti version seems to use the PG147-SKU60 board design with a GB203-300-A1 GPU. On the other hand, the standard RTX 5070 is positioned as a more energy-efficient choice, featuring 6,144 CUDA cores and 12 GB of GDDR7 memory on a 192-bit bus, with a memory bandwidth of 627 GB/s. This model is expected to have a slightly lower TBP of 250 W.
Interestingly, the non-Ti RTX 5070 card will be offered in two board variants, PG146 and PG147, both using the GB205-300-A1 GPU. While the pricing details are still unknown, NVIDIA seems to have made significant differences between its SKUs. The Ti variant not only has an additional four GB of GDDR7 memory but also a substantial 45% increase in CUDA core count, going from 6,144 to 8,960 cores. While we anticipate the initial release of GeForce RTX 50 series cards at CES, the GeForce RTX 5070 and RTX 5070 Ti are expected to launch later, possibly after the RTX 5080 and RTX 5090 GPUs.
