Nvidia 3080 driver support questions

Hey, got a question similar towards cybapunk here → The madlad himself …and, I was hoping you could give a more up to date answer as you did theirs. I have a 3080 sadly, and I was hoping I could get your feedback on how I can implement my own 3080 Nvidia GPU so it can function within qubes. If I also got to ask in the development page, lmk!

In my use case of it, I would purely use it for Fedora for the graphical side, Whonix for tor, and Arch for everything else I haven’t decided on.

The Nouveau project may be good if there isn’t any other way for it to be used, but with projects like PopOS, or Nobara (Which is based on Fedora), which use their own implementations, would there be a way for a user to implement what they’ve done and fuse it within Qubes since dom0 is based in fedora as well? I’d have to ask around how to do that, so if you know how to implement it, I’d love your feedback on it.

I understand that with the hypervisor and virtual machines if I were to somehow implement that purely with Fedora, templates like Arch or whonix may not receive the same capabilities since they would not hold the same code. If my idea for how I’d implement graphics on Fedora would work, would it mess up somehow if I installed Nouveau on Whonix or Arch instances? (Unless there’s a better way to install drivers.)

Once again apologize for my lack of understanding, but this is why I ask. I heard that with virtual machines, you need two GPU’s, or at least a gpu and a cpu with implemented graphics because for Nvidia, it’s hard for their gpu’s to divide between two distinct instances of hypervisors? Would love to clear that misunderstanding since I only have one GPU and a AMD 4090 CPU.

Would love your feedback and thank you for sharing your time and answer back if you do reply.