Please note that this post is tagged as a rumor .
NVIDIA to reintroduce GeForce RTX 2060 with 12GB memory?
We have new information on planned update to GeForce RTX 20 series.
Back in January, it was reported that NVIDIA is bringing back a large stock of RTX 2060 graphics cards. This would be a direct response to the market situation at the time which was just the beginning of price and stock issues of discrete GPUs in 2021. This was also before NVIDIA launched its RTX 3060 graphics, so it was understandable that a good mid-range solution is needed.
Fast forward to September and NVIDIA already launched its RTX 3060 non-Ti model in March offering competitive performance and a large memory buffer of 12GB. The card was later updated to the LHR variant after NVIDIA made a mistake by unlocking full cryoto hash rate performance through a developer driver. However there is a problem with RTX 3060 graphics card and that’s availability at MSRP. The situation has not changed much since the card had released, which forces NVIDIA to seek other options.
According to our sources, NVIDIA is now planning an updated RTX 2060 GPU based on Turing architecture. A model known as PG161 would feature 12GB GDDR6 memory and the same TU106-300 GPU, except wth a KX postfix. Thus, the card would only see a change in doubled memory capacity compared to RTX 2060 6GB which launched in January 2019. NVIDIA has allegedly told board partners to anticipate this card by the end of this year, possibly in January, depending on how quickly they can be manufactured.
An introduction of RTX 2060 12GB paints a grim outlook for mid-range GPU availability in 2022. NVIDIA is clearly expecting Ampere shortages to continue, while already planning alternatives to customers who simply cannot wait.
| NVIDIA GeForce RTX x60 Series | |||||
|---|---|---|---|---|---|
| VideoCardz.com | GeForce RTX 2060 SUPER | GeForce RTX 2060 12GB | GeForce RTX 2060 6GB | GeForce RTX 3060 Ti | GeForce RTX 3060 |
| Picture |  |  | |  | |
| Architecture | NVIDIA Turing | NVIDIA Turing | NVIDIA Turing | NVIDIA Ampere | NVIDIA Ampere |
| Board | PG150 | PG161 | PG160 | PG190/PG142 | PG190 |
| GPU | 12nm TU106-410 | 12nm TU106-300-KX | 12nm TU106-300 | 8nm GA104-200 | 8nm GA106-300 |
| Base Clock | 1470 MHz | TBC | 1365 MHz | 1410 MHz | 1320 MHz |
| Boost Clock | 1650 MHz | TBC | 1680 MHz | 1665 MHz | 1777 MHz |
| CUDA : RT : Tensor | 2176 : 64 : 136 | 1920 : 48 : 120 | 1920 : 48 : 120 | 4864 : 38 : 152 | 3584 : 28 : 112 |
| Memory | 8GB GDDR6 | 12GB GDDR6 | 6GB GDDR6 | 8GB GDDR6 | 12GB GDDR6 |
| Memory Clock | 14 Gbps | TBC | 14 Gbps | 14 Gbps | 15 Gbps |
| Memory Bus | 256-bit | 192-bit | 192-bit | 256-bit | 192-bit |
| Memory Bandwidth | 448 GB/s | TBC | 336 GB/s | 448 GB/s | 360 GB/s |
| TGP/TBP | 175W | TBC | 160W | 200W | 170W |
| MSRP | 399 USD | TBC | 349 -> 299 USD | 399 USD | 329 USD |
| Release Date | July 9th, 2019 | Dec ’21/Jan ’22 | January 14, 2019 | December 2nd, 2020 | Late February 2021 |