Shading issue with Vray GPU / Normal map creates weird results

[^Lele^]

You mentioned this happened also "with every imported geometry like megascans".
I couldn't repro it with meshes with the right UVs, or none at all.
If you had one such case, we'd love to see it!

[kosso_olli]

Not at all, they are two different things: there is no sign of stretching in that case (as it's got to do with the space the normals are calculated in, for cpu and gpu under max).
In yours, it looks much as the UVs are calculated wrong, for whatever reason: if you could provide a file and the texture for investigation, it'd be good for us.

EDIT: attached, a substance painter starter model, with normal mapping rendered under GPU (beauty + bumpNormals). CUDA or RTX makes no difference to me.

Interestingly, the RGB beauty is flawed as well: The bump map reveals the underlying tesselation. Take a look at the beige leather parts in the left quarter of the image. It closely reminds me of the same issue we had with the CPU renderer many years back.

[fixeighted]

You're right and that happened as I tested, however I did something to make it go away....which I now can't remember what or reproduce

[^Lele^]

The Bumpnormal RE shows the normals are fine, that was the goal of the exercise.
The shading issue was of my own making as the model had not been subdivided, nor were the normals smoothed (any smooth look was due to the fine noises perturbing the geo normals.) .
EDIT: I was wrong: the issue is with high gloss on normal-mapped surfaces (with this substance 3.0-driven bump.). will report, thank you.

[kosso_olli]

EDIT: I was wrong: the issue is with high gloss on normal-mapped surfaces (with this substance 3.0-driven bump.). will report, thank you.

As said, it looks similar to an issue which was present with normal maps many years ago. Tried to find the thread which also had a detailed explanation on why this happens by Vlado, but I can not find it anymore.

[^Lele^]

It could be, i filed a report, the devs will check it out soon enough.