The RTX 2060, which debuted in January 2019, is back with twice the memory.
How hard Is it now to obtain a new graphics card? It’s difficult enough that Nvidia is almost literally going backwards with its product range. After weeks of speculation, the official Nvidia spec page revealed a new version of the GeForce RTX 2060, which was initially announced in January 2019. It features twice as much GDDR6 RAM (12GB) and significantly more CUDA cores. There’s no word on a release date or price, but “soon” and “more than you want” appear to be the most likely options.
Aside from the additional RAM, the new RTX 2060 boosts the number of CUDA cores (tiny, very efficient parallel processors) from 1920 to 2160. The basic GPU frequency has also increased from 1365MHz to 1470MHz. The card is essentially identical to the RTX 2060 introduced over three years ago, with the exception of a little reduction in the stock boost speed from 1680MHz to 1650MHz. The RTX 2060 was released at a retail price of $300-350, and Nvidia may follow suit here, but there’s no way of knowing what the extremely competitive market would charge once these cards hit store shelves (and secondary listings).
Reading between the lines, the presence of this new SKU indicates that Nvidia is finding it more simpler to procure the high-speed memory it utilizes in its card designs than the components required to build its RTX 30xx series Ampere CPUs. The de facto introduction of a new 2060 card means that budget-friendly (in relative terms) RTX 3050 and 3050 Ti desktop cards seem unlikely to appear very soon.
Even used RTX 2060 cards are still selling for $500-600 on the secondary market — more than twice their original retail price! — so there’s no knowing how much these new cards will cost once they’ve sold out of store stock. Scalpers and bitcoin miners aren’t reducing their output in any significant way. We can only hope that increased supply would reduce competition and, as a result, the cost of affordable PC gaming gear.