Full Article
Does Airbnb still use React Native? The full story (2016–2025)
Everyone in mobile development has heard that Airbnb dropped React Native. The 2018 blog posts get referenced constantly. But the story before and after those posts is more interesting than the posts themselves.
2016: The adoption
Airbnb's mobile teams were small compared to the web team. React Native looked like the fix: web engineers could write JavaScript and ship features to iOS and Android without learning Swift or Kotlin.
It worked, initially. The team built production features including parts of the Experiences product. Web engineers were shipping to mobile for the first time.
2017: The problems compound
React Native versions 0.43 through 0.49 were rough. The team ran into issues that kept getting worse:
- Screen readers didn't work reliably across platforms, creating accessibility gaps they couldn't ignore
- Navigating between native screens and RN screens felt janky to users
- Every RN version upgrade risked breaking things that already worked
- The JavaScript bridge bottlenecked complex interactions
React Native never grew past 15–20% of the mobile codebase. The rest was still native. Maintaining both was harder than committing to either one.
June 2018: The sunset
Gabriel Peal published a 5-part series on Medium titled "Sunsetting React Native." It became the single most-referenced piece of writing in the React Native community.
The technical reasons were specific:
- All JS-to-native communication went through an async bridge, which capped performance
- RN screens initialized slower than native screens
- No synchronous access to native views, which ruled out certain UI patterns
- "Cross-platform" still required plenty of platform-specific code
- The hybrid codebase was harder to maintain than pure native would have been
2018–2019: Going native
With React Native gone, Airbnb poured resources into native tooling:
- MvRx (now Mavericks): A Kotlin-first reactive framework for Android that Airbnb open-sourced. It got serious traction in the Android community.
- Epoxy: Declarative RecyclerView layouts. Airbnb's answer to the composability problem on Android.
- Lottie: The animation library that renders After Effects files natively. Lottie actually predated the RN sunset and is still actively developed.
On iOS, the team migrated to Swift and built their own architecture patterns, later adopting SwiftUI. The broader engineering effort also tackled page performance measurement and messaging infrastructure as part of the native rebuild.
2021: Ghost Platform
This is the part most people miss. Ghost Platform is a server-driven UI system, and it solved the actual problem React Native was supposed to solve.
How it works: feature logic lives on the server, not in the app binary. The server defines UI components, and each platform renders them natively. New features ship without app store updates. One backend change updates iOS and Android at the same time.
Ghost Platform gave Airbnb cross-platform shipping speed with full native performance. It's a fundamentally different approach than React Native, but it hit the same goal.
2023–2024: Trio
Airbnb built Trio, a Jetpack Compose architecture layered on top of Mavericks. Declarative UI, reactive state management, type-safe navigation. They're still investing in making native Android development better rather than replacing it.
2025: The everything app
Airbnb's 2025 redesign expanded the app beyond accommodation booking into experiences, services, and AI-powered trip planning. All of it runs on the native stack with Ghost Platform handling cross-platform delivery.
The bottom line
Airbnb does not use React Native. They replaced it with Swift and Kotlin, then built Ghost Platform for the cross-platform velocity they originally wanted. The native tooling they've open-sourced since then (Mavericks, Epoxy, Lottie) has shaped how other companies build mobile apps too.
What makes the Airbnb story worth revisiting isn't the React Native failure. It's that the replacement actually worked, and worked differently than most people expected.
