Nvidia’s upcoming RTX 2050 laptop GPUs don’t sound “new” at all, considering that the company already released the RTX 3050 and 3050 Ti GPUs for laptops earlier this year (via VideoCardz). But it’s clear that Nvidia isn’t ready to retire its 20 series GPUs just yet — Nvidia says that the RTX 2050 will be available in laptops starting in the spring of 2022.
The RTX 2050 is based on Ampere GA107 architecture, not Turing
The RTX 2050 isn’t just a piece of old hardware brought back from the dead. Unlike other 20 series chips, the RTX 2050 is based on Ampere GA107 architecture, not Turing. Ampere is the same architecture that the RTX 3050 (and all other 30 series cards) is powered by, and packaging it within a 20-series card could potentially offer some of the same ray-tracing improvements at a cheaper price. The RTX 2050 comes with the same 2,048 CUDA cores and 4GB of GDDR6 memory that the RTX 3050 has. However, it offers a 64-bit memory bus when compared to the RTX 3050’s 128-bit, potentially leading to a bit of a downgrade in performance.
While it’s possible that Nvidia is releasing a GPU that doesn’t belong to the latest and greatest 30 series due to the ongoing chip shortage, the move may also come as a way to provide a middle ground between the RTX 3050 and the RTX 1650. Nvidia might want a mid-range GPU that offers a boost in performance over the RTX 1650, but is still more affordable when compared to devices that carry the RTX 3050.
In addition to the RTX 2050, Nvidia also announced the entry-level MX550 and MX570 laptop chips, but the company hasn’t provided many details about their specs. Nvidia vaguely mentions “more CUDA Cores” and “faster memory speeds” with no solid numbers to back it up just yet. All this comes about two weeks after Nvidia released a new version of the RTX 2060 —which was originally launched in 2019 — that offers an upgrade of 12GB of video RAM when compared to the 6GB of VRAM that the standard model comes with.