React Native vs Expo · 2026 decision framework

In 2026, the question for a new React Native project is not 'should we use Expo', it's 'is there a reason NOT to'. Managed Expo is the default. Here are the eight reasons to leave the default.

When Expo wins (most cases)

• OTA updates · ship fixes without App Store review (huge for early product phase).
• EAS Build · no local Xcode / Android Studio dance.
• The dev client handles 95% of native module needs now with `expo-dev-client`.
• Push, auth, storage, camera · all managed, all work cross-platform out of the box.
• Tailwind + Reanimated + Gesture Handler all 'just work' via managed config.

When bare React Native still wins

1. Custom native SDK you can't prebuild with EAS (rare in 2026, but exists · some payment SDKs).
2. Deep background-task needs (BLE polling, foreground services on Android) where Expo's modules are restrictive.
3. Extremely-constrained binary size requirements (Expo adds ~5-10 MB of framework).

When to skip React Native entirely

Three cases: heavy offline-first with complex DB sync (Swift/Kotlin CoreData/Room is 2× the dev time but 3× the predictability); games (use Unity or native); and apps with heavy platform-specific UI conventions (financial apps, health apps where native feel is the product).

Our default stack in 2026

Expo managed · TypeScript · Expo Router · NativeWind (Tailwind for RN) · Zustand · TanStack Query · Sentry · EAS Build + EAS Update. This stack shipped GlowUp (iOS + Android) in 5 weeks including App Store review.