Name the product, then the constraint that defines it. A modern headset — Apple's Vision Pro, Meta's Quest, Sony's PlayStation VR2 — has to fill your entire field of view with pixels at a high enough refresh rate that motion feels real. The number of pixels times the refresh rate times two eyes is a brutal compute and power bill, and a device strapped to your face cannot dissipate much heat or carry a big battery. Every spatial-computing headset hits this same wall.
The shared way around it is a trick borrowed from human biology, and it is right there in the patents. Your eye only resolves fine detail at the fovea — the tiny central region of the retina. A few degrees off-center, your acuity collapses. So if the headset knows where your eye is pointed, it can render that small region in full resolution and everything else coarsely, and you will not notice. That is foveated rendering, and it requires eye tracking to work.
Meta's grant US11669160B2, "Predictive eye tracking systems and methods for foveated rendering for electronic displays" (issued June 6, 2023), names the dependency in its own title: predictive eye tracking for foveated rendering. "Predictive" matters — the system has to guess where your eye will be a frame or two ahead, because rendering is not instantaneous, and a late high-resolution patch defeats the purpose. The CPC class G06F 3/013 (eye-tracking input) sits right alongside the rendering classes.
Sony got to the same place from the console side. Its grant US10720128B2, "Real-time user adaptive foveated rendering" (issued July 21, 2020, with an earlier sibling US10192528B2 from 2019), claims adapting the foveation to the individual user in real time. Sony's framing leans on power savings — the CPC classifications include G06F 1/3265, a power-management class — which is exactly the constraint a battery-or-thermally-limited headset cares about.
Three records, one story. Meta and Sony are different companies with different products and different business models, but their patents describe the same mechanism because they are solving the same physics. That convergence is the signal: when independent assignees file on near-identical techniques, you are looking at table-stakes engineering for the category, not a proprietary edge. Eye tracking in a headset is sold as a luxury input method; it is really a compute-budget necessity.
The boundary worth stating: these grants cover specific eye-tracking and rendering methods, and a given headset may use its own variant or license around them. But the strategic read is clear. The reason your headset watches your eyes is not mainly to let you click with a glance — it is so it can avoid rendering millions of pixels you would never have seen sharply anyway. Follow the constraint, not the demo, and foveated rendering is the whole reason high-resolution spatial computing fits on your head at all.