Have the sony engines (decima, nd engine, sucker punch engine, etc.) Any of these features or something comparable?!
We usually don't hear much about existing game engines until new games using those engines reveal what the teams are working on. There's no reason to truck Decima out to GDC, for instance, to show off what it can do when nobody else but Sony teams can use what it can do. (Technically, Sony has allowed some developers like Supermassive or Kojipro to use Decima, but those were special occasions and the games were for Sony platforms anyway.)
Specifically though, no, neither Sony or Microsoft have announced any kind of virtualized micropolygon geometry "infinite polys" tech. Similarly, Unity Engine and CryEngine and Lumberyard and other technologies have not, to my knowledge, announced anything like Nanite either. (If we're just talking Lumen, that's a little bit of a different story since that incorporates a lot of different aspects of lighting, and there are many third-party systems for UE and elsewhere that would be in competition, more or less.) What Epic has introduced here has been swirling in the ether for a while beforehand (
John Carmack was talking about virtualized geometry in concept over 10 years ago and that inspired Epic to invest in it,) and maybe other teams have had something in the works similar in anticipation of these new platforms (or are quickly working to rip it off, maybe?), but so far there's not been any direct announcements in response to Nanite or Lumen from other technology providers that I'm aware of?
...That all said, developers have to
use it, and even though Unreal Engine 5 is staggering technology conceptually, developers are still trying to figure out how to use it in a game; even Epic Games is still working out where and how best to use it.
Lumen is I would assume a no-brainer: if you're working on next-gen only releases, you like this lighting solution over what you're used to or offered elsewhere, and your performance tests are in line with expectations, then go Lumen. If you also want to include past-gen in the plan, I don't know much about the fallback method for Lumen? I know with Nanite, you can build default states of LOD on your polygonal models to have baseline previous support, I assume Lumen is the same that you'd just do build and test for both methods, but if you
really did Lumen right, that seems like it'd be a lot of work to do over again.
Nanite, it has some major drawbacks that you'd have to be aware of before using it. It cannot be deformed (bend or stretched or offset ala grasses,) you cannot set an opacity/transparency, Raytracing has limitations when used on Nanite objects, and you cannot use Forward Rendering or do Vertex Paint on Instances. Also, it's next-gen only. The good news is, it's really easy to designate a model for using Nanite, and developers can play around to figure out where best to use it versus other traditional (or novel) methods. But if you looked at either of those UE5 demos from Epic Games and though, "Ah ha, graphics is solved!!", maybe step it back a bit...
There are cases where amazing technology gets introduced in an engine and shown off in action via demos but then never really lives up to the promise. Remember UE4's own Sparse Voxel Octree Global Illumination lighting system? That got dropped before the technology was finalized (even though other technology partners went on to add versions of SVOGI in their tech.) I don't think
anything like that will happen with Nanite or Lumen, it's already so solid and effective (and has already been deployed on a console, which is important,) and should just be a general good even if it in real life it may or may not make games as amazing as we imagine. Sometimes tech doesn't work out like it looks like it would in demos, though. For example, Unity introduced something called ECS (Entity Component System) as part of its DOTS system, and you may remember three years ago they had that cool-looking MegaCity demo with a stunning number of detailed objects on screen (it's not the same as Nanite, but it is their way of rethinking how game data is used to scale up what gamers see.) Looked great, they even ported the demo to a phone, but then come a few years later, you don't see massively detailed worlds (and behind the scenes, developers have apparently not been aggressively using this tech,) and
in the most recent Unity 2021 builds, ECS is not compatible with where Unity is headed in the future.
So, this was a dead end? Stuff happens.