Apple Vision Pro

Apple Vision Pro is Apple's first spatial computing headset, announced at WWDC in June 2023 and released in the United States on February 2, 2024 at $3,499. It is a video see-through mixed reality device — virtual content composited over a real-time passthrough feed — that operates entirely without controllers, using eye gaze, hand gestures, and voice as its only input modalities.

Hardware

Displays: Two micro-OLED panels deliver a combined 23 million pixels — approximately 4K resolution per eye — at a pixel density high enough that individual pixels are not resolvable at normal viewing distances. The displays are covered by a single continuous curved glass front and driven by a custom display engine that calibrates for each user's prescription via swappable ZEISS optical inserts.¹

Processing: Two chips operate in tandem. The M2 chip (Apple's laptop-grade processor) handles application compute, graphics rendering, and the operating system. The purpose-built R1 chip handles all real-time sensor processing: fusing data from 12 cameras, 5 sensors, and 6 microphones to compute head position, hand pose, eye gaze, and passthrough video with a 12ms motion-to-photons latency — below the threshold at which the brain perceives a mismatch between physical head movement and visual feedback.³

Sensors: The sensor suite includes outward-facing cameras for world tracking and hand detection, downward-facing cameras for hand pose refinement, infrared flood illuminators and cameras for eye tracking, a LiDAR scanner for depth, and a TrueDepth camera for face tracking and Persona capture.

Audio: A spatial audio system with head-related transfer functions plays sound that appears to emanate from specific 3D locations in the environment, updating in real time as the user moves.

Power: A woven cable connects to an external battery pack worn on the body, providing approximately 2 hours of untethered operation. The external battery arrangement keeps the headset weight at approximately 600–650g.

Input

Apple Vision Pro has no physical controllers. Input is entirely through:⁴

Eye tracking: The user looks at an element to target it. The system knows which element the user is looking at from the four NIR eye tracking cameras inside the visor. The gaze signal is never exposed directly to applications — only the resulting interaction event (a tap on the looked-at element) is delivered, preserving user privacy.

Hand gestures: A pinch gesture (thumb and index finger touching) confirms a selection, equivalent to a click. Spread fingers, swipe gestures, and two-handed interactions control scrolling, zooming, and window manipulation. No wrist bands, gloves, or controllers are required — the cameras observe the hands resting naturally on the user's lap or on a surface.

Voice: Siri and system-level dictation provide voice input throughout visionOS, accepting both commands and free-form dictation for text entry.

visionOS

visionOS is Apple's operating system for Vision Pro, built on foundations of iOS and macOS but with a spatial layout model: applications run as floating windows and volumes in the user's physical space rather than on a fixed screen.² Applications can be windowed (flat panels at any position), volumetric (3D objects in space), or fully immersive (replacing the real world entirely via Full Space).

Standard iOS and iPadOS applications run unmodified as 2D windows in visionOS, giving Vision Pro immediate access to a large software library from launch.

EyeSight

A distinctive outward-facing display on the front of the headset — named EyeSight — shows a representation of the wearer's eyes to people nearby. The intent is to preserve social awareness: bystanders can see whether the user is engaged with virtual content (eyes obscured, iris display shown instead) or is looking at the physical space and available for interaction (eyes visible and correctly positioned). EyeSight dims when the user enters a fully immersive environment.³

Optic ID

Optic ID is Vision Pro's biometric authentication system, using the infrared eye tracking cameras to capture the unique pattern of each user's iris. Authentication happens in under a second during initial look-and-pinch interactions, without any explicit unlock step.

Market Context

At launch, Vision Pro was positioned as a developer and early-adopter device — a platform introduction analogous to the first iPhone in 2007. The $3,499 price and external battery limited its mass-market appeal, but it established visionOS as a platform and demonstrated a quality ceiling for passthrough fidelity and interaction comfort that competitors would need to address.⁵