Last week, it was reported that NVIDIA is preparing a new “variant” of RTX 3050 that uses AD106 silicon, which is part of the Ada Lovelace architecture (responsible for RTX 40 series GPUs). The “RTX 3050 A” was found in several references, and now the chipmaker has confirmed the chip’s existence, NotebookCheck reports.
RTX 3050 A: 1792 CUDAs, 4GB, 35-50W

While the RTX 3050 A Laptop GPU shares a name similar to Ampere-powered RTX 3050, the specs are anything but – for starters, the CUDA cores are now reduced from 2048 to 1792, and the configurable TGP’s upper limit has been reduced from 80W to just 50W (though this is 5W more than Ampere-based RTX 2050). VRAM is also reduced from 96-bit 6GB GDDR6 to 64-bit 4GB GDDR6.
So far, no laptop has been announced to feature this new GPU variant, though it’s unlikely to land itself on a gaming laptop, as it is expected to be lowest-end chip of the Ada Lovelace family (or the RTX 30 lineup, depending on how you look at it). There is a good chance laptops in the premium segment (think Lenovo Yoga or Dell XPS) may use this chip for specific configurations, though.
Pokdepinion: Still baffles me, that name.