In a surprising move, NVIDIA has introduced new GPUs just before CES 2022. The new GPU called the NVIDIA GeForce RTX 2050 follows the naming scheme of old NVIDIA chips — the new chips start with "3" — but the name isn't all. The new RTX 2050 comes with the same Ampere architecture found in the new RTX 3050 series.

This GPU is targeted towards affordable gaming laptops and NVIDIA says the first laptops with this GPU will be available in the spring of 2022. As mentioned before, the new NVIDIA RTX 2050 isn't based on the Turing design. Instead, it features the Ampere GA107 architecture. The RTX 2050 comes with 2,048 CUDA cores and 4GB of GDDR6 memory. What's different (and a slight downgrade) from the 3050 series is that the RTX 2050 features only a 64-bit memory bus. The 3050 features a 128-bit memory bus. It also draws a little less power (45W) than the higher-priced GPUs (80W).

As The Verge notes, it seems that NVIDIA has introduced these new chips to provide a middle ground between the RTX 3050 and the RTX 1650. In addition to the RTX 2050, NVIDIA has also announced two new entry-level chips — the MX550 and MX570. Although NVIDIA doesn't dwell on the number of CUDA cores and other details, it says that they have much faster memory and greater power efficiency than the MX450. It does, however, reveal that the MX550 is based on the old Turing architecture while the MX570 is based on the new Ampere architecture.

Source NVIDIA | Via Engadget