Shopify is leaving React Native for Swift and Kotlin, and the deciding factor was coding agents

For six years Shopify was the reference customer that made React Native credible for serious apps. On September 10 it said that era is over: the company is rebuilding its mobile apps in native Swift and Kotlin, and the reason it gives is not that React Native let it down. The reason is that coding agents changed the math that made cross-platform worth choosing in the first place.

What Shopify actually said

Shopify’s 2020 bet on React Native rested on a simple trade: write one codebase instead of two, and accept some distance from each platform in exchange for the engineering time you save. The company now argues that agents “can now do enough of the implementation, translation, testing, and review work that it’s no longer the deciding factor it was in 2020.” Build the same feature twice, in Swift and Kotlin, and the duplicate effort that used to be the whole argument against native is largely absorbed by the tooling. The Shop app has already shipped as a native rebuild. The flagship Shopify app is underway now, including more than 300 screens, home and lock-screen widgets, an Apple Watch app, and Siri Shortcuts, with Point of Sale and Inbox on the list.

The numbers come from a real rebuild, not a slide

The Shop migration is the part with evidence attached. Shopify reports the rebuild ran 12 weeks from proof of concept to app-store release, with a core team of six engineers on the native foundations before feature teams joined. The agents were not running loose: Shopify built a migration workflow it calls Pi that inspected the React Native source, documented its behavior, planned the native version, and reviewed parity, plus a debugging tool called Tardis that gave agents live access to app events, logs, and state so they could diff screenshots and payloads against the old app. Shopify’s reported results favor native: startup time down 23 percent on iOS and 50 percent on Android, the Android binary 37 percent smaller, Android scrolling at 120 FPS, and crash-free stability rising from the historical 99.5 percent to 99.95 percent. Those are vendor figures on the vendor’s own hardware, so read them as direction rather than benchmark, but the direction is consistent and large.

Why this is a signal and not just a vendor decision

Shopify is not a cautionary anecdote. It was the company that argued loudest, and with the most production code behind it, that React Native was ready for apps at scale. A reversal from that particular source lands harder than the same decision from a team that never committed. As the independent developer Simon Willison put it, the move happened “for the exact reason you would expect,” and he was careful to credit React Native as a strong platform across the six years Shopify ran on it. The framing matters: this is not a bug report on React Native, it is a claim that the economic case for it has shifted under everyone.

What it means for teams choosing today

The honest reading is narrower than the headline. Shopify did not win this with off-the-shelf agents; it won it with custom infrastructure, Pi and Tardis, built specifically to drive and verify a migration, plus a React Native codebase complete enough to serve as an executable specification. Most teams have neither. So the lesson is not “abandon React Native this quarter.” The lesson is that the core assumption under cross-platform, that one shared codebase is always cheaper than two native ones, is no longer safe to treat as a constant. When agents absorb a meaningful share of the cost of maintaining parity, the things native buys you, direct platform access, fewer layers between your code and the OS, and first-day support for new platform features, stop being luxuries you trade away by default.

For a working mobile team, the concrete action is to re-run a decision most people settled years ago. If you chose React Native or Flutter primarily to avoid writing features twice, that reason is weaker in 2026 than it was when you made the call, and it is worth measuring rather than assuming. The framework question is quietly turning back into a question about platform fit and performance, which is where it started before cross-platform promised to make it go away.

Sources cited

1. Native is now the future of mobile at Shopify (Shopify Engineering) shopify.engineering
2. Migrating the Shop app from React Native to native (Shopify Engineering) shopify.engineering
3. Native is now the future of mobile at Shopify (Simon Willison) simonwillison.net