We’ve been building on React Native since long before it was the obvious choice. Our first chat SDK shipped in React and React Native, and our open-source mobile sample apps have tracked the framework’s evolution closely. This post is a 2025 check-in on where React Native, React.js, and cross-platform mobile development stand — and where we think the lines are drawn between “use cross-platform” and “go native.”
React Native in 2025
The New Architecture (Fabric, TurboModules, Hermes, Bridgeless mode) has been the dominant story since 2022. By 2025 it has matured to the point where it’s the default rather than the opt-in. The practical wins:
- Synchronous native module calls when you need them.
- Significantly lower memory footprint and better startup time vs the old bridge-based architecture.
- Concurrent React features (Suspense, transitions) actually work in mobile contexts.
- A unified rendering pipeline that closes most of the perceived “feel” gap with native.
Expo has continued its trajectory toward being the default way most teams ship React Native in 2025. EAS (Expo Application Services) handles the build, submit, and update pipelines that used to be the most painful part of a React Native project. For most B2B mobile apps, the question is no longer “Expo or bare React Native” — it’s “which Expo workflow.”
React.js in 2025
On the web side, the React story has consolidated around:
- Server Components. What started as Next.js-specific in 2023 has become the default mental model for a new React app in 2025. Most teams are now shipping a mix of server and client components.
- Next.js, Remix, TanStack Start, Astro. The framework layer is more diverse than it was in 2023 — Next still has the largest share, but Remix and TanStack Start have meaningful followings, and Astro is the standard pick for content-heavy sites.
- Bun and the runtime side. Bun has gone from interesting to credible for production. Node.js remains the default but no longer the only mainstream choice.
- Tailwind, shadcn/ui, and the design-system shift. The “copy this component, don’t install it” pattern that shadcn popularised has become the dominant way teams assemble UIs.
Where cross-platform still wins
For most B2B mobile apps in 2025 — internal tools, vertical software, chat-driven applications, AI-agent front-ends — cross-platform via React Native is the right default. The reasons haven’t fundamentally changed:
- One engineering team, two stores, plus web if you want it.
- Faster iteration from prototype to production.
- Code reuse with your web app, especially if both are React-based.
- A mature ecosystem of libraries for the integrations most B2B apps care about — auth, chat, push notifications, payments, analytics.
Where it falls short
- Heavy graphics workloads — games, AR/VR, advanced video processing.
- Apps where the native platform UX conventions are part of the value proposition (deep Apple Watch integration, iOS-specific accessibility features, etc.).
- Apps that need very low-latency access to specific hardware sensors.
For everything else — and “everything else” still covers the bulk of B2B mobile work — cross-platform is the boring, right answer.
What’s changed for chat and AI in cross-platform apps
The biggest 2025 shift specific to chat and AI agent integration in React Native apps: streaming UIs are now table stakes. Users expect token-by-token streaming for AI responses, typing indicators that mirror chat patterns, and the ability to interrupt or branch a conversation mid-stream. Building that well requires the chat SDK and the AI agent layer to be designed as a single system rather than bolted together at the app layer.
Some of our earlier React Native posts from the 2021–2024 archive still hold up — why React Native is the best option for blockchain-based app development and open-source React Native app templates and UI kits remain useful starting points if you’re spinning up a new project.
Related from Ethora: Ethora Chat SDK — the React Native chat SDK with self-hosted and dedicated-server options.