Generation is a terrible metric.

The 4060 and 4090 are the same generation but drastically different fidelity.

The 3090 is on older generation than the 4060, yet the 3090 wipes the floor with the 4060. Generation doesn't mean what is used too.

Teraflops is also a God awful metric.  The ps5 is 10 tflops and some gpus are over 100 tflops....   the visual difference isn't 10x.

The 4060 *is* a newer generation, just a tiny chip that doesn't deserve its xx60 series name. 

New architectures with a full portfolio will always include these tiny entry-level GPUs. The GT 430 from 2010, for instance, which is just a bit smaller than the 4060, was slower than the 9600 GT (which in turn was already slower than the GTX 8800 from 2006) but still two generations ahead.

I feel there's a bit of bad faith arguing going on here. Can you link to specific quotes made in this thread that say Switch 2 will be equivalent to PS5?

What I can see is people arguing that it'll be capable enough to run many current gen games with compromises, whether those compromises are deemed worthwhile, remains to be seen. The compromises made on Steam Deck seem acceptable to many owners of that system. 

Also, 'The switch 2 will be a ps4 and that is a huge jump for Nintendo.  Day 1 here.  The ps5 stuff is nonsense.', where are you getting this info? DF's best guess has it pegged significantly more capable than PS4, what do you know that they/we don't?

Making a statement like that with that kind of certainty is just silly, we don't even know what node the thing will be manufactured on, which in itself will have a big impact on performance. Who knows, you may end up being correct, but that'll be out of sheer luck as we simply do not have enough info to make that call now.

Something like this?

"I guess you missed the part where they shocased unreal 5 engine demo similar to what what shown for ps5. That cannot run on ps4/ps4 pro specs. It will also have a current gen based cpu, which is much more powerful than last gen, along with a much bettter cpu."

I know semantics are pedantic and eye-rolling... but I think it's important to establish the diference between "equivalent" and "similar".

"Similar" means being almost the same as some other thing, but with some clear differences and particularities. A definition that seems pretty good for Switch 2 ports from PS5/XSX games.

When something is "equivalent" to some other thing it means that both have the same value or qualities, in the sense that you could substitute one thing for the other and there would be no significant difference.

In the context of the quote you provided, the use of the word "similar" should be understood as the Switch 2 version of the Matrix Demo being comparable to the PS5 one (as in, recreating the same scenes, gameplay and overall look), but with some obvious differences. Not equivalent. Similar.

That quote doesn't stablish that the Switch 2 and the PS5 would be "equivalent". In fact, "similar" was used only refering to the Matrix Demo, not the console itself. Another key diference.

Does a system run PS5 class games that aren't able to run on a PS4? Like more than just 1 or 2 of them?

If the answer to that question is "yes", then that hardware in question is not a PS4 class hardware. It can't be by definition.

It's like saying Bo Jackson was only a baseball player ... but he also played NFL football as well. Well, hold on a freaking second there, that's a pretty fucking huge distinction, he's not in that case only a baseball player, he's someone who played both baseball and football, the original statement is misleading.

Steam Deck is not PS4 class hardware, it's a tier above. It's likely Switch 2 is the same deal.

Different graphics settings don't mean you're playing a totally different game either, that's just dumb. Do you go "wow! I'm playing a totally different tier of game now that I've switched from Medium Low to High settings" when playing on a PC? It's still the same damn game. No one freaking says that because that's stupid. If I buy Cyberpunk 2077 from the Steam store and play it on my Steam Deck at 720p low settings and then play it on my home PC at 1440p high settings, that's not two different versions of Cyberpunk 2077, it's the same version of the game. No one seriously thinks they're getting two distinct versions out  of the game because the developer is being generous. 

And that's part of what you get with this PC-ization of home consoles and even portables ... they're just the same game with graphics settings moved up or down and resolution moved up or down. The days of like when the Wii got a completely bespoke, different version of Call of Duty while XBox 360/PS3 had a different version are over, now the hardware lines have become blurred and really all you're going to be seeing is the same game on different platforms, just with the base PC settings moved around a bit and the resolution slider moved around.

And for console generation gate keepers who want to cry about this stuff, go blame Microsoft for it then. The Series S basically brought down the roof for so-called "next generation" game dramatically lower by basically forcing developers to have to make a version of every game that operates only 1/3 of what the next-generation machines are. That's a big advantage for the Switch 2, the Switch 1 didn't have the benefit of a 600 GFLOP PS4 Series S or something to lower the roof for all so-called "next generation" games. 

Is it Blazer? Or Zero999?

And similar is very subjective.  Some might argue Hogwarts and Witcher 3 on the switch are similar to the ps4 versions and are effectively the same game.  Some others would argue the two above mentioned games are not similar at all.  

Most people are being realistic, it is just a few outliers that are funny.

"The switch 2 will wipe the floor with the ps4, so it will also wipe the floor with the ps4 pro, though with a much closer gpu gap vs that one."

I suppose you could argue semantics once again but wiping the floor of the ps4 pro sounds like a ps5.    And it would be easy to discredited the post as trolling, except one person openly agreed.  

At the end of the day, at least for me, the Switch 2 is day 1 for me.  Nintendo software is amazing.  I fully anticipate buying third party on the ps5/pc for the very noticeable fidelity and performance gains.  

Something else to add to the discussion is that "power levels" get ambiguous when you compare the mid-gen refreshes of the 8th gen to the Series S. 

Depending on metric of choice you'll get different answers: Image quality? The last gen refreshes tend to have higher native resolutions in cross-gen games than the Series S. Performance? Due to their slow CPU's, the Series S is able to outperform the last gen refreshes when it comes to average framerates and stability. Graphics quality? The Series S is capable of more modern features than the mid-gen refreshes. 

This is important, because like the Series S, the Switch 2 (given the current speculation) is probably going to outperform the mid-gen refreshes in certain areas, but not others (definitely not image quality.) 

I think with the Switch 2 we'll see soft images, slightly softer than what we see with the Series S, but in terms of performance and graphics features it'll outclass the mid-gen refreshes when docked. 

So if we were to separate out these three "scores", it is likely going to be something like: 

1. feature/graphics quality: Steam Deck < Mid-gen refreshes < ROG Ally < Switch 2* <~ Series S (Switch 2* might see more ray-tracing support than Series S, despite being a half GPU-tier worse in rasterization, due to the Nvidia chip)

2. image quality: Steam Deck < Rog Ally ~ Switch 2* handheld (although the image will look better on small screen) < PS4 < Switch 2 docked  < Series S < 8th gen refreshes 

3. performance (defined mostly by stability at target framerate): 8th gen refreshes < Steam Deck ~ Rog Ally ~ Switch 2 (docked and handheld) < Series S. 

Then you have the PS5 and XBS X in a tier above all of these low-end platforms with mid-tier (~ RTX 3060; ~ RX 5700xt) gaming PC's. 

Then above them you have two more tiers of gaming PC's which we could call: high-end (RTX 3070 and better; RX 6700 and better) and "enthusiast" (RTX 3090/4070ti and better; RX 6900 and better.) 

*Note: when I say Switch 2 this is with the assumption it will have a T239 and would be roughly comparable or slightly better than an RTX 2050 25W. 

Correct. 1440P doesn't need to be the target for a VRAM bottleneck.

That wasn't my argument.

As for the 3050 4GB vs 6GB.

Keep in mind the 4GB actually has a 128bit memory bus (4x 32bit) and the 6GB model has a 96bit memory bus (3x 32bit) and in very bandwidth-demanding scenarios, the 4GB card can actually be faster depending on the demands of the software.

https://technical.city/en/video/GeForce-RTX-3050-4GB-mobile-vs-GeForce-RTX-3050-6GB-mobile

However, if everything is not kept equal...
3050 6GB - 2560 Cuda Cores @ 1237Mhz
3050 4GB - 2048 Cuda Cores @ 1237Mhz

Then the 6GB part is the obvious winner by sheer increase in functional processing units.

The 3050Ti will often show a sizable advantage over the 6GB variant due to it's higher levels of bandwidth, especially when a ton of alpha effects are being used.

A high probability that 1080P is not the target.

The original Switch came with 4GB of Ram.
1GB of Ram was used for the OS/Background tasks leaving 3GB for games.

Yet... Hogwarts legacy is running in that 3GB of Ram pool.

I'll let you draw the conclusion for that one.

The 3050 6GB vs 3050Ti.

It also showcases that the 3050 Ti 4GB thanks to it's higher levels of bandwidth, can outperform the 6GB card.

Which brings me back to my original point about developers building their software for the hardware environment... And also brings me back to my original point that more than 4GB of VRAM is not super important in the grand scheme of things in these low-end GPU's.

The fact is... A 2050 4GB is going to be roughly as powerful as the Tegra in the Switch 2.0.

You aren't going to get much more than that as it's not financially responsible.