NVIDIA's RTX 5050 is again the subject of chatter from the rumor mill, with an apparent confirmation that GDDR6 memory will be employed with the desktop version of this GPU (but not the laptop variant).

VideoCardz reported last week that its sources claimed that NVIDIA hasn't given up on the desktop spin of this graphics card, despite it falling off the radar in recent times, while the laptop version was what all the rumors were focused on.
Other sources have jumped on this confirmation bandwagon - obviously add salt with all this - and Benchlife has now said that NVIDIA's graphics card making partners have received updated info on the RTX 5050 desktop.
This confirms GDDR6 is to be used and that both Samsung and SK Hynix will be supplying those memory modules, whereas before it was only the former, we're told. That said, it won't be until later in the production cycle of the RTX 5050 that SK Hynix joins the fold.
With finer details like this starting to filter through, it seems like the RTX 5050 desktop is indeed set to become a reality (after NVIDIA skipped the RTX 4050 variant, which was mobile-only as you doubtless recall).
The specs are reportedly the same as we've heard before: 8GB of GDDR6 VRAM, with the GB207 chip sporting 2560 CUDA Cores. This will, of course, be a budget desktop graphics card, and we can keep our fingers crossed that NVIDIA will pitch the price at a suitably low level.
Note that Benchlife claims the laptop RTX 5050 will use GDDR7 VRAM, and it's just the desktop model that will run with GDDR6, which is unusual. Indeed, if this is correct, out of all the Blackwell GPUs, it's only the desktop RTX 5050 that'll supposedly use the slower type of video RAM - and if true, hopefully that will be reflected in the price.




