Why Airbnb’s React Native Experiment Ended With a Complete Rewrite | by Yash Batra | JavaScript in Plain English
Article Summary
Airbnb bet big on React Native to unify their mobile codebase. Two years later, they scrapped it all and went back to native.
This deep dive examines why Airbnb's high-profile React Native experiment failed despite the framework's promise of code sharing and faster development. The article breaks down the technical, organizational, and performance issues that accumulated at scale.
Key Takeaways
- Bridge serialization overhead killed performance for scrolling and animations
- Platform quirks required native rewrites anyway, eroding code sharing benefits
- New bug classes (bridge deadlocks, stale props) increased debugging complexity
- Developer expertise split across JS and native created organizational friction
- Hidden costs of runtime complexity outweighed theoretical maintenance savings
The theoretical 30-40% code sharing gains were consumed by bridge overhead, platform-specific workarounds, and a new category of cross-platform bugs that native code never had.
About This Article
Airbnb's React Native bridge created async boundaries that led to memory leaks and lifecycle differences between iOS and Android. As the codebase grew, these issues became harder to debug and more expensive to maintain.
The team rewrote the app in native code and removed the JS-to-native bridge. UI logic now runs directly in platform-specific languages, giving them better control over performance.
Without the bridge layer, Airbnb fixed an entire category of bugs and got predictable performance back for scrolling and animations. The tradeoff was having to maintain separate iOS and Android codebases.