Shopify Leaves React Native: What It Means for Your App

Shopify is moving its mobile apps off React Native and back to Swift and Kotlin, and it says the reason is AI coding agents. The company had gone all-in on React Native in 2020. On September 10, 2026, Shopify Engineering published the reversal. If you run a cross-platform app, the news deserves a calm read, not a panic rewrite.

What Shopify actually announced

The original case for React Native was simple: stop building every feature twice, let engineers work across the stack, stop chasing feature parity between iOS and Android. Shopify’s post says one assumption behind that decision changed: “LLMs changed one of the core assumptions behind our 2020 decision.”

Its position now is that coding agents handle enough of the implementation, translation, testing and review that building twice is no longer the deciding factor. Native also keeps the team closer to platform capabilities and first-party tooling, with fewer framework layers in between.

The timing had a second driver. A large refactor to adopt React Native’s New Architecture was coming. If you’ve been following how Fabric and JSI change the runtime, you know that’s not a small migration for a big app.

The numbers

Shopify chose greenfield rebuilds over a gradual migration.

App	Status	Detail
Shop	Shipped native	One-engineer proof of concept in a week, then six engineers to production in 12 weeks
Shopify (main app)	Migrating	300+ screens, widgets, Apple Watch app, complications, Siri Shortcuts. Ships later in 2026
Other mobile apps	Queued	”Soon”

Twelve weeks for a flagship consumer app is fast. It’s also worth being clear about what made it possible. Shopify built an internal system called Helix that forces agent output through checkpoints: test verification, a visual review against the running app, two adversarial code reviews, and human approval before anything lands. It also split business logic from UI behind a CLI, so agents can iterate on core logic in milliseconds instead of waiting on a simulator.

That is real engineering investment. The agents didn’t do it alone.

The open source fallout

If you use these libraries, read this part twice. From Shopify’s post:

• React Native Skia stays maintained through 2026. William Candillon is forking it and the original repo is archived.
• FlashList, at around 2 million weekly downloads, gets critical fixes only. Shopify is looking for stewardship partners.
• Restyle is archived, maintained through the end of 2026.

FlashList is the one I’d watch. A lot of production lists in React Native apps run on it. Critical-fixes-only is survivable for a while, but “while” is the word. If your app leans on it, put a look at alternatives on next quarter’s roadmap, not next year’s.

What this does not mean

It’s tempting to read this as a verdict on cross-platform. It isn’t.

Shopify has hundreds of mobile engineers, a huge native surface (watch apps, widgets, shortcuts), and the infrastructure to keep agent output honest. A five-person team shipping one app on Expo has none of those. For that team, two codebases still means two sets of bugs, two release trains and two of everything to review. Agents write code faster, but somebody still owns it at 2 a.m.

I’d also note what Shopify optimized for. It wants deep platform integration: widgets, Apple Watch, Siri. Those are exactly the places cross-platform frameworks are weakest. If your app is a forms-and-lists product, you may never touch that wall. Our comparison of React Native, Flutter and PWAs still holds for that case.

When the question is worth reopening

The useful thing Shopify did was reopen a decision everyone had stopped questioning. So reopen yours, with numbers:

1. Count what cross-platform costs you today. Engineer hours spent on bridge issues, framework upgrades, and platform features you had to build natively anyway.
2. Price the native alternative honestly. Two codebases, including the review and QA load, with realistic agent assistance, not a best-case demo.
3. Look for a wall, not a preference. A concrete blocker (a platform API you can’t reach, a perf ceiling, an upgrade like New Architecture you can’t afford) justifies a rewrite. “Native feels better” doesn’t.
4. Pilot on one screen or one app. Shopify started with a one-week proof of concept. That is the part to copy.

If you’re weighing the staffing side, we’ve written about hiring a React Native developer and hiring a Kotlin developer, and the cost breakdown for mobile app development gives a baseline for the two-codebase math.

The short version

Shopify’s real message is narrow: the cost of building twice fell, and for a company with its scale and tooling, that flipped the answer. Check whether the same assumption is still true for you before you copy the conclusion.