NVIDIA further complicates its product line by announcing the RTX 2060 (12 GB) with 2,176 CUDA cores, 14 Gbps VRAM and a 185 W TGP
NVIDIA has lifted the lid on the GeForce RTX 2060 (12 GB), which it will be releasing as a Founders Edition card. Incidentally, NVIDIA has informed The Verge that add-in board partners (AIBs) will also announce custom versions of the Turing-based card.
As expected, the GeForce RTX 2060 (12 GB) offers a blend of specifications from the RTX 2060 (6 GB) and RTX 2060 SUPER. On the one hand, the RTX 2060 (12 GB) has 2,176 CUDA, 64 RT and 136 Tensor cores, matching the RTX 2060 SUPER. It has a 1,650 MHz boost clock like the RTX 2060 SUPER too, 30 MHz lower than the RTX 2060 (6 GB).
On the other hand, the RTX 2060 (12 GB) retains its predecessor's 14 Gbps memory clock speed, along with its 192-bit memory bus. As a result, the RTX 2060 (12 GB) has a 336 GB/s memory bandwidth like the RTX 2060 (6 GB). However, the RTX 2060 (12 GB) has a 185 W TGP, meaning it has more power at its disposal than the RTX 2060 (6 GB) and RTX 2060 SUPER.
The RTX 2060 (12 GB) will be available from December 7, although perhaps only in the US. Currently, NVIDIA's UK website does not reflect the announcement of the RTX 2060 (12 GB), for example. NVIDIA is yet to announce pricing for its Founders Edition card, nor an MSRP for AIBs. However, it has indicated that it will cost more than US$349, the RTX 2060's launch price. According to The Verge, an NVIDIA spokesperson stated
It [RTX 2060 (12 GB)] is a premium version of the RTX 2060 6GB and we expect the price to reflect that.