You can buy the high VRAM version or you can lower settings.
5060 Ti 8GB delivers a monstrous amount of performance at a relatively cheap price thanks to 8GB. Around 15-20% Faster than PS5 Pro, But with a PS6 level feature set. An absolute steal at 379$.
If you want must have 12GB, go buy 12GB B580. But 95% of people don't want to do that. Do you want to know why?
The major misunderstanding of PC here that people have is not understanding that consumers don't want high VRAM GPUs.
They want high VRAM Nvidia GPUs.
AMD and their influencers (MLID, Kepler, AMDUB, etc) have complained in the past about the phenomenon of buyers who view Radeon as a tool for getting cheaper Nvidia.
It's not a deflection. The major issues of 8GB games are concentrated in bad ports that eventually get patched. That doesn't mean that games can't take advantage of >8GB GPUs. But that practically all games can be made and are made to provide a great experience on 8GB GPUs.
The major contention here is that PS5 has slightly more VRAM than 8GB in many titles. Thus exact PS5 settings would push you over the VRAM buffer. Which requires creating a system of memory management. If the system is implemented well we can see 8GB PC surpass PS5.
TLOU P1 had worse textures on launch PC than on 8GB than the 256MB VRAM PS3. That's not a PC issue, it's an issue of a comically broken streaming system they tried to copy paste from PS5 onto PC. There is also that windows itself is terrible and uses hundreds of MBs of VRAM it doesn't need to use.
The actual solution here is 9GB 96b GPUs. Nvidia wants to do this but they know that they will get burned at the stakes by AMDUB, GN, etc for it. So they can't do it.
Thus the VRAM whining complex is just making things worse. Imagine someone that reviews every iPhone and his entire argument is that the iPhone 18 Pro Max should be 499$. Citing some Chinese Pico phone as proof. Is that a serious reviewer?
Again, this is just trying to price out people from gaming with extra steps.
VRAM now is 10$/GB, and once you Apply Nvidia's margin, that's 18$/GB. Apply AIB/Distribtor/Reatiler margin? We are are 20$/GB on the MSRP.
5050 now is 289$ (subsidized by patriotic 5090 buyers).
5050 12GB? 289+4*20 = 369$
5070 18GB? 629+6*20 = 749$
Now would AMDUB Shell 369$ RTX 5050 12GB? 749$ RTX 5070? Off course not. He'd burn it at the stake.
He complained 4060 Ti 8GB should be 299$. So Nvidia made RTX 5060 which is exactly what he asked for at 299$. Except now it needs to be 150$ or less at 8GB or have 12GB at 299$. Does it matter that GDDR7 3GB was still sampling and not in mass production when 5060 launched? Off course not.
All these takes are variations of either:
1. Nvidia has to have no margin. Why do they have margin?
2. Not understanding that Memory's moore's law is facing similar challenges to the logic one.
AMD people always assume that 93% of the market now buying Nvidia are behaving irrationally.
It can't possibly be because RDNA2/3 Windows driver is 12% slower than the open source Mesa Linux driver. It cannot possibly be because AMD has defacto put those cards in maintaininance mode long ago. It cannot possibly be because Turing and Ampere buyers loved having the lifetime of their GPU extended by getting DLSS SR and multiple software updates that retroactively upgraded their cards (4.0, 4.5).
It cannot possibly be because Nvidia architectures are forward looking and thus they don't have an upscaler to use that can reach the image quality of Nvidia running 4.5 in performance mode.
And when they get FSR4 INT8 unofficially, they get stuck having to spend 12.2ms on N22 doing 1080p => 2160p upscale from 1080 to 4K

. Because AMD didn't think Gamers didn't need tensor cores until Nvidia proved that they do.
No, it's because they are dumbasses.
The reality is that consumers are hyper rational. They understand at this point that the high VRAM on AMD GPUs is bait. It's why they demand an AMD GPU be 15-30% cheaper than the raster equivalent at Nvidia. It's also why they don't care when AMD offers lots of VRAM.