NVIDIA looks set to launch a lower-memory version of its GeForce RTX 3050

New launch on the horizon for NVIDIA, which seems fully intent on bringing the third version of RTX 3050 to the marketfollowing the launch in 2022 of a version dedicated to OEMs with some slight differences compared to the regular model.

This new variant, however, seems to be an almost completely different card which will appear on the shelves with the name RTX 3050, but with important differences compared to the 8 GB version and not just because of the “only” 6 GB of memory.

According to information that has appeared online in recent days, it seems that NVIDIA's new proposal has in common with its older sister for the OEM market only the CUDA Core (2304), only to then suffer a cut in the 96-bit memory bus and a series of reductions in terms of frequencies, bandwidth (obviously) and TDP.

Same name but not same card

Doing some quick calculations, the reduction of memory (and related reduction of the bus) combined with the cutting of frequencies and TDP could guarantee a power reduced by approximately 25% compared to its bigger sister: performance which, if confirmed by benchmarks, would make this version of the 3050 a direct competitor of the Radeon RX 6500 XT.

However, there are not only negative sides to proposing a solution with reduced frequencies and TDP compared to full power solutions: with 70W the doors are opened to solutions without a power connector and possibly also to fanless solutions.

This would allow us to offer (more in the OEM channel at our idea) a first price solution that can still guarantee adequate performance for some esports titles but at the same time, with the hardware demands that the latest generation AAA titles are bringing to the table, it seems rather unlikely that this version of the 3050 will be able to guarantee performance in line with the expectations of the most passionate gamers.

A risky move for NVIDIA?

Offer a product with a name already known to consumers with more limited features, can easily displease more than one person: already at the time of the “launch” of the 3050 for the OEM partners several enthusiasts turned up their noses and in that case we found ourselves having to deal “only” with a rather significant cut of CUDA Core and – in part much more laughable – of frequencies.

At the time, what saved the situation was the intended use, that of the OEM channel, which did not give rise to any crime of lèse majesté in the eyes of gamers and which allowed NVIDIA to recover some GA106 chips without too many worries otherwise not salable in the version for the normal market. The same TDP, then, allowed the GPU Boost algorithm to “mask” the cut, making the gap more visible on the data sheet than in everyday use.

We just have to wait for the availability of the card to be able to test its real performances, without forgetting that – even if they may be unattractive for the more “hardcore” enthusiasts – solutions like this hypothetical RTX 3050 6 GB and similar may prove functional in certain segments of the market capable of guaranteeing NVIDIA and its partners important numbers.