Meta

Meta's Cuttlefish on Mac: Running Android’s Virtual Device Framework Natively on Apple Silicon

Original post Cuttlefish on Mac: Running Android’s Virtual Device Framework Natively on Apple Silicon

Article Summary

Meta's Reality Labs just solved a problem that's been fragmenting Android development for years: running high-fidelity Cuttlefish emulators natively on Apple Silicon Macs. No more choosing between convenience and accuracy.

Meta's Reality Labs team needed to test AI glasses and other Android devices, but faced a painful choice: use lightweight Goldfish emulators everywhere with poor fidelity, or use accurate Cuttlefish emulators only on Linux. With most developers on Macs, they ported the entire Cuttlefish stack to macOS by reviving QEMU support and solving three critical platform gaps.

Key Takeaways

Critical Insight

Meta consolidated from two emulator platforms to one high-fidelity Cuttlefish setup running natively on Apple Silicon, cutting maintenance overhead while giving all developers and external partners the same accurate virtual devices used in CI.

Continue reading

What This Article Covers

Problem

QEMU's vhost-user support and HVF memory-region tracking only worked on Linux. The build and signing paths were tied to the same platform, so Meta's Reality Labs couldn't run Cuttlefish on macOS.

Solution

Meta brought QEMU support back into Cuttlefish. The team loosened the Linux-only build gates on vhost-user frontends and libvhost-user, then added Darwin build support with JIT entitlements and notarization.

Impact

The unified stack now handles Bluetooth, Wi-Fi, and cellular connectivity, plus 3D graphics, OTA update testing, fastboot, and snapshotting. It's shared identically with external developers.

The article reveals how this work enables AI agents to run validation loops alongside human developers in ways that weren't possible before.

This page summarizes “Cuttlefish on Mac: Running Android’s Virtual Device Framework Natively on Apple Silicon” by Meta, published Nov 13, 2026. All credit for the original work goes to the authors.

Read the original on engineering.fb.com (opens in new tab)

AI tools for Android

Browse all AI tools →