Maxing everything means enabling path tracing which is a massive performance hit. Basically to do that you need a 4090 or a cheaper 40xx card and use frame generation.
The “PS5/Series X equivalent” settings are a mixture of low/medium without any hardware ray tracing (I think it still does some software ray tracing similar to UE5 Lumen) using FSR2 with post processing effects done before the upscaling (low resolution). On PC if you use high post processing effects they happen after the upscaling and thus at much higher resolution (this is the “heaviest” option outside of the ray tracing stuff)
Remedy hasn’t made a game that targets “native” resolution since Max Payne 2. Basically starting with the first Alan Wake game they have used lower internal resolutions and used the budget for other graphical effects. (On PC you still always had the option to use native resolution)
The feature sets involving AI in the newest cards are really the key, I think, and why my last gen premium card is lagging behind. DLSS 2.0 on the 4080 and 4090 especially are just revolutionary in how they work. The last gen cards just can’t quite keep up. It’s like magic, making frames out of seemingly thin air.
In my opinion this is a good problem to have! I’d rather have games pushing the state of the art than sticking to what the PS5 and Xbox whatever it’s called these days can do.
Max settings also means enabled path tracing, which current-gen Nvidia's 4000 series have better hardware acceleration for. So you can't really compare performance to previous hardware which don't have that hardware acceleration. No amount of optimization is going to fix that, and it should probably be turned off if you're not on a 4000 series or newer card.
In that same way, AMD's path tracing hardware acceleration falls behind Nvidia's. A current-gen high-end AMD RX 7900 XTX does worse than Nvidia's last-gen not as high-end 3070 in Cyberpunk 2077's path tracing update. https://youtu.be/cSq2WoARtyM?t=503