Ray tracing and DLSS 2/3 most probably yes. They are behind in ray tracing simply because they use compute shaders for ray tracing. Only they get full blown ray tracing cores I think they will perform similar.

Same with DLSS 2/3. There is a precedent for this. XESS was able to get pretty close to DLSS within an year. All AMD has to do is use machine learning for upscaling.

The biggest issue for AMD is not these features. But the fact that NVIDIA keeps introducing new features. So everytime AMD could catch up with one Nvidia Would have probably introduced 2 more.

Can they? Yes, will they? Nvidia hopes not, so they will do whatever they can so it doesn’t happen. Which is a good thing.

The turning point will be when path tracing will become the default and soon thereafter the only mode in new games. This will likely become a thing in some future console generation, likely the second or later from now. This will very likely be tied to AMD IP combined with Sony and Microsoft own research. At that point AMD IP must be good enough to make it happen without much if any disadvantages to Intel and Nvidia offerings.

AMD is behind but it’s working to some extent. They don’t have a long R&D history like Nvidia and they aren’t rushing it either, leveraging partners as well. They have to be the best or very competitive in like 10-15 years.

AMD has to catch up to Intel first. ARC has better hardware support for this stuff

AI tasks can usually be broken up into training which requires a bunch of high precision floating point power versus inference which can get by with less precision.

Last generation AMD essentially abandoned the AI data center world to NVidia and focused on double precision vector performance to get HPC contracts. This generation Lisa seems to be saying they’ll once again avoid competing head to head in the data center but this time emphasizing inference performance at the expense of training, which they’ll once again mostly cede to NVidia. Since these are the same sort of resources you need to do DLSS I think there’s no reason to think they can’t compete on a hardware level.

For the datacenter side, once you’ve got your model moving it from one architecutre to another isn’t as complicated as training it in the first place and I’m optimistic about AMD’s strategy there. But keeping up on the consumer software support side is a much tougher row to hoe. As a Linux gamer AMD graphics cards are the best choice for me but I understand that software stability on their closed source drivers is… maybe a bit lacking. So on a consumer level I’m really not that optimistic about AMD translating hardware parity into a good experience for most gamers.

They already have AI cores for other applications.just like this cobsole generation.AND thinks rtx wont became the norm until the next console generation.this is why they are not trying to make rtx their gimmick yet to try to catch up to early adopter nvidia for the last 2 generations.pepperidge farm remembers how awful BFV ran on 2080s. And their FSR3 “application”(keyword here because biggest turnoff for upscalers are how badly implemented their libraries are) is much better application to FSR1 and 2. Management of AMD doesnt think investing to RTX and upscalers when their CPU and APUs making a lot of money. And radeon generations last 2 releases have been pretty usable and price competitive seeing nvidia squeeze more on consumer lineup

No because Nvidia has software companies on their side and that proprietary software is simply more developed than that for AMD.

I think it highly depends on what patents NVIDIA holds. Patents hold technology back, and one of the major obstacles for companies is finding ways around the patents other companies hold.

Can they? Yes!
Will they? Probably not. Nvidia has more software developers, more industry support, an already established technology, more money… I don’t see it coming.