Shopify has decided to leave React Native, and much of the industry has treated that as a verdict on React Native. It is a decision by one company, for its own reasons.
What Shopify said
In its engineering post, Shopify says it will migrate all its mobile apps to Swift and Kotlin, using AI throughout. The Shop app has already shipped as a native app, the Shopify app is under way, and the rest will follow. InfoQ reported it as Shopify abandoning React Native to rewrite its flagship apps.
The reasoning is specific. Shopify adopted React Native in 2020 to stop building every feature twice. Its team now writes that agents can do enough of the implementation, translation, testing and review work that building twice is no longer the deciding factor it was in 2020. The Shop app rebuild went from proof of concept to the app stores in 12 weeks, started by a core group of six engineers with feature teams joining midway. By Shopify’s own measurement, iOS startup is 23 per cent faster, Android startup is 50 per cent faster, and crashing sessions are ten times fewer.
I take those results at face value. The reaction to them is where I disagree.
A company that is not your company
Look at what the 12 weeks needed. Shopify built its own workflow on top of a coding agent, with specialised subagents for review, planning and parity checking. It built a debugging tool, Tardis, that gives agents live access to the app’s events and logs so they can compare the old app with the new one. Its post on the Shop app also says that generated code can meet a feature’s requirements while adding duplication, architectural drift or performance problems, and that native expertise remained essential.
That is a large company with native expertise, a big existing codebase to compare against, and the budget to build tooling around the migration. The results are its own, measured against its own earlier build, on the first of its apps to move. The Shopify app has more than 300 screens and is still in progress.
Your version of building twice also costs people. Two codebases means hiring and keeping engineers who know Swift and Kotlin, and coordinating two release trains. A company that has no such team cannot buy them with agent credits.
Twenty months
In January 2025 Shopify published Five years of React Native at Shopify and wrote: the future of React Native is bright and we think Meta have been great stewards of the project. In the new post its author recalls having written then that Shopify planned to keep investing in it.
I don’t hold that against them. It is a candid change of mind, and it happened because a tool got better. But the reasoning is conditional on how much agents can do at a given moment, and that moment moves. Shopify may change position again in a few months. If you rewrote your product on the strength of the last one, you would be rewriting again.
What you share, and who shares it
Most businesses gain from sharing more between platforms than they duplicate, and the question is what to share. React Native shares almost everything, and its biggest advantage is people: a team that knows React can work across web and mobile with mostly the same skills. Shopify’s own reasons for choosing it in 2020 included letting developers work across the stack. Flutter asks for Dart, and Kotlin Multiplatform shares logic in Kotlin, so a company with a large web team gets less overlap from either.
None of that stopped being true on 10 September. Agents cut the cost of writing the second codebase, and they do not turn a Swift engineer into a React engineer. What changed for Shopify is that one cost of the trade-off went down, for a company able to use the change.
A team I work with answered the same question differently. It recently ported an app from native to Kotlin Multiplatform in about four weeks. Two things decided it: its mobile experts could work across the whole product instead of half of it, and it could share business logic while keeping a native experience. Quality of experience and cost per unit of value both moved significantly in the right direction. Those reasons are not the ones that favour React Native, and they were right for that team. Each business is its own world.
Architecture by headline
This is a pattern, and it is wider than mobile. I call it architecture by headline: someone prominent publishes a decision or a claim, and everyone else treats it as evidence about their own situation.
The last few months have been full of it. There was the argument over whether AI can be controlled at all, followed by a wave of chatter about Jevons paradox as a comfort or a threat, depending on the day. There was the rise and fall of OpenClaw, which peaked in March and collapsed in April. Each one sent some teams redrawing their plans.
I made the same point about a five-year migration that took two weeks: cheap code changes what a migration costs, and it does nothing about whether you should do it.
Three questions to ask first
- What problem did they have, and do I have it? Shopify’s was the cost of keeping two platforms in step.
- Do I have what made their answer work: native engineers, tooling, a budget for both?
- What does it cost me if the answer changes again?
Shopify’s answer is Shopify’s. If it changes its mind in another twenty months, will your roadmap survive that too?
Leave a Reply