Virtual realities are the media in which spatial intelligence becomes constitutive rather than optional: an extended-reality (XR) device cannot draw a single pixel until it knows where it is, what surrounds it, and where the user is looking and reaching. This makes XR the most demanding real-time deployment target for the perception, reasoning, and generation capabilities of the previous part, specialized to run egocentrically, on-device, and within a hard latency budget.
Beneath everything in XR sits the platform itself, the runtime that localizes the device, maps and registers the surrounding world, senses the user’s body, and composites the result in real time; layered on top is the application that has driven the medium hardest, entertainment, spanning live AR and VR experiences and virtual production as well as the offline, high-fidelity reconstruction and performance capture that supply them with authored content.