Apple Vision Pro: A Technical Deep Dive

Apple doesn’t call the Vision Pro a headset. They call it a spatial computer — and once you understand what’s inside it, that distinction starts to make sense.

The Hardware

At the heart of the Vision Pro sit two chips working in tandem: the M5 (in the updated 2025 model) and the R1.

The M5 handles general computing — running apps, rendering environments, processing input. It’s the same chip found in Apple’s latest MacBooks, with a 10-core CPU, 10-core GPU, and 16GB of unified memory. Having desktop-class compute inside a headset is genuinely unusual.

The R1 is the more interesting chip. Its sole job is processing sensor data and pushing pixels to the displays with a 12-millisecond photon-to-photon latency. That number matters: 12ms is fast enough that your brain doesn’t perceive a lag between moving your head and the image updating. Go above ~20ms and motion sickness sets in. The R1 is Apple’s solution to that problem, processing input from 12 cameras, 5 sensors, and 6 microphones simultaneously.

The Displays

The Vision Pro uses micro-OLED displays — not the standard OLED panels you’d find in a phone or TV. Micro-OLED packs the pixel emitters directly onto a silicon wafer, allowing for an extraordinarily dense 7.5-micron pixel pitch. The result is 23 million pixels total across both eyes, at resolutions of around 3660×3200 per eye.

For context, a 4K TV has about 8 million pixels. The Vision Pro has nearly three times that, split across two displays the size of postage stamps.

The M5 model supports refresh rates up to 120Hz, with specific rates (90Hz, 96Hz, 100Hz) to support judder-free playback of 24fps and 30fps video content.

Input: No Controllers

One of the Vision Pro’s most technically ambitious decisions is ditching handheld controllers entirely. Input is handled through three modalities:

Eye tracking — four dedicated cameras track where your eyes are looking with enough precision to use gaze as a cursor. You look at something to select it.

Hand tracking — cameras map your hand position and finger gestures. A pinch of your thumb and index finger acts as a click. You don’t need to hold your hands up; resting them in your lap works fine.

Voice — Siri handles text input and commands throughout the OS.

This system works surprisingly well for browsing and casual use, though developers have noted it has limitations for tasks requiring precise or sustained input.

Sensors

The sensor array is extensive: two high-resolution main cameras, six world-facing tracking cameras, four eye-tracking cameras, a TrueDepth camera, a LiDAR scanner, four IMUs, a flicker sensor, and an ambient light sensor. The LiDAR in particular enables accurate room mapping, letting virtual objects sit convincingly in physical space.

EyeSight and Personas

Two features address the social awkwardness of wearing a headset. EyeSight is a curved lenticular OLED display on the front of the device that shows a rendering of the wearer’s eyes — dimming when they’re immersed in content, brightening when they can see the room. It’s an attempt to signal presence to people nearby.

Personas are photorealistic avatars generated by scanning your face, used during FaceTime calls so others aren’t just staring at the outside of a headset.

The R1 and visionOS Architecture

visionOS is built on the same foundation as iPadOS but redesigned around spatial interaction. Apps exist as floating windows in 3D space rather than a 2D grid. The OS uses the R1 to ensure that the passthrough camera feed — your view of the real world — never lags or stutters, which is critical for a device you’re wearing on your face.

One notable feature: Mac Virtual Display, which lets you wirelessly use the Vision Pro as a massive private monitor for your Mac, with wide and ultra-wide layout options.

Battery and Comfort

The battery is external — a hockey-puck sized pack that connects via cable and lasts around two hours. This was a deliberate tradeoff to keep weight off the face. The M5 model weighs 750–800g excluding the battery, which is heavier than the original M2 version.

Comfort remains one of the most common criticisms. The weight distribution and fit vary significantly between users, and long sessions are tiring for many people.

Where Things Stand

The Vision Pro is currently in an interesting position. Apple refreshed it with the M5 chip in October 2025 but made no major design changes. Reports suggest a second-generation model is on pause, and Apple’s longer-term focus appears to be shifting toward AR glasses — starting with a display-less AI glasses product similar to Meta’s Ray-Bans, with a display-equipped version to follow.

For now, the Vision Pro remains a first-generation product with first-generation tradeoffs: extraordinary technology, a steep price ($3,499), limited battery life, and an app ecosystem still finding its footing. But as a technical achievement — cramming desktop compute, micro-OLED displays, and a real-time sensor fusion chip into a wearable device — it’s hard to argue it isn’t impressive.