Quick verdict
Cross-platform (React Native + Expo) is the 2026 default for most new mobile apps. Pick native for deep OS features, games, AR/VR, or apps where native-only distinguishes you. Pick PWA for content-first tools where App Store discoverability isn’t important.
What each path actually is
- Native: Swift for iOS, Kotlin for Android. Two separate codebases, two teams (or a very hardworking solo dev). Max performance and deep OS access. Highest per-feature cost.
- Cross-platform: React Native or Flutter. One codebase, both platforms. “Native-enough” for typical product apps, with AI generation on top making it unusually fast to ship in 2026.
- Web (PWA): a website that can install to the home screen and work offline. Lowest cost, no App Store review needed, but limited device features and zero discoverability inside the store.
Head-to-head comparison
| Dimension | Native | Cross-Platform | Web / PWA |
|---|---|---|---|
| Performance | Best | Near-native | Variable |
| Two codebases? | Yes (iOS + Android) | No | No |
| Device features | Full | Most | Limited |
| App Store presence | Yes | Yes | No |
| Discoverability | Store search | Store search | Web only |
| Hiring pool | Small | Huge | Huge |
| AI tooling quality | Moderate | Excellent (RN) | Excellent |
| Time to MVP | 8–16 weeks | 1–4 weeks (with AI) | 1–2 weeks |
| Typical cost | High | Medium | Low |
When to pick native (Swift/Kotlin)
- Games (especially 3D / GPU-heavy).
- AR/VR apps.
- Apps where differentiation is native-only — Apple Watch-first, CarPlay-first, Live Activities, complex widgets.
- Apps with heavy custom animations or gestures that Reanimated/Skia can’t handle.
- Large teams with existing native expertise — the maintenance cost is familiar.
- Apps where you’re shipping one platform (e.g., iOS-only pro tools). The cross-platform tax isn’t worth paying.
When to pick cross-platform (React Native + Expo)
- Consumer or prosumer apps that need to ship on both iOS and Android.
- MVPs and products where velocity matters more than the last 5% of native polish.
- Teams with web / JavaScript talent who don’t want to become native experts.
- Apps where AI code generation is part of your workflow — React Native has the best 2026 tooling.
- Products that might eventually want a web version — Expo + React Native Web sharing 60–70% of code.
- Solo founders and small teams. Most apps in this category, full stop.
When to pick web / PWA
- Content-first products (blogs, news, reference tools).
- Internal tools where users install via a link, not the App Store.
- B2B SaaS companions where the real product is on desktop web.
- Prototypes where App Store review time is a blocker.
- Geographies where App Store access is restricted and web is the main channel.
The 2026 AI-builder shift
The old rule was: “pick native if you can afford it, cross-platform if you can’t.” In 2026, AI flipped this. Three reasons:
- AI tooling quality. LLMs produce shippable React Native code at far higher quality than Swift or Kotlin. The velocity gap is now 3–5x in favor of cross-platform.
- Expo maturity. EAS Build, EAS Submit, EAS Update, Expo Router have turned the ecosystem from “a compromise” into “a better DX than native.”
- Native’s remaining advantages shrank. Live Activities, widgets, App Clips — all accessible via Expo config plugins. You pay complexity tax when you need them, not always.
For most 2026 product apps, cross-platform is not the budget choice anymore — it’s the correct choice.
Decision framework (3 questions)
- Do I need App Store distribution? Yes → native or cross-platform. No → PWA is viable.
- Do I need deep native features (AR/VR, complex widgets, games)? Yes → native. No → cross-platform.
- Does my team live in JavaScript/TypeScript? Yes → React Native + Expo. Already in Dart / Flutter? Stay there. Already in native? Stay there unless rebuilding.
Common mistakes
- Picking native because “it’s more professional.” The App Store does not care.
- Picking PWA when App Store discoverability is part of your GTM.
- Building in Flutter because a tutorial recommended it, without considering hiring and AI tooling.
- Treating the decision as permanent. You can migrate later — start with the fastest path.
- Ignoring AI generation in the decision. In 2026, it’s one of the biggest velocity levers.
The simplest next step
For the vast majority of 2026 founders: pick React Native + Expo, generate the first cut with ShipNative, extend in Cursor when needed. Migrate to native only if a feature forces you. In practice, almost no one ever needs to.