That can’t be the reason. The WiFi is much, much faster

I asked about this in the Unity visionOS discussion, and one of their engineers confirmed that it is in fact the obstacle. They simply can't implement dynamic/gazed foveated rendering without the eye tracking data. It's not even on the roadmap for Unity because there isn't anything they can do until Apple makes the eye tracking data available to Unity's rendering pipeline. The way to solve this would be for visionOS to ask the user if they want to make their eye tracking data available in VR (this is how it currently is with hand tracking data). Then Unity can use dynamic foveated rendering (which is ideal, and already works in AR), or fall back to fixed foveated rendering if the user wants to keep eye tracking private. That will make it so people can see everything clearly in high resolution in VR -- and VR developers won't need to choose between heavy upscaling and fixed foveated rendering (which lowers resolution around the center of the view -- and requires people to turn their heads instead of their eyes to see clearly). It would make the VR rendering really amazing.

I only mentioned Unity, but every other game engine and rendering system will be needing dynamic foveated rendering for VR also. It’s a key optimization for VR, and it’s not really optional for such high resolution displays. Vision Pro already has everything needed to get it working completely, and making the real-time eye tracking available to VR rendering (with the user’s permission) is the only thing missing.