Shopify Ditches React Native, Rebuilds Apps in Swift and Kotlin

Shopify just reversed one of the most-cited case studies in mobile engineering. In a post published today, the company said it is moving its flagship apps back to native Swift and Kotlin, six years after it went all-in on React Native to build iOS and Android from one shared codebase. The reason isn't that React Native failed. It's that AI coding agents changed the math that made a shared codebase worth the tradeoffs in the first place.

• Shopify is rebuilding its mobile apps in native Swift (iOS) and Kotlin (Android), reversing the React Native strategy it adopted for its largest app back in 2020.
• The company says AI coding agents have closed the cost gap that made cross-platform frameworks attractive: building a feature twice, once per platform, no longer costs twice the engineering time.
• Its rebuilt Shop app took 12 weeks with AI assistance; the main Shopify app spans 300+ screens plus widgets and an Apple Watch companion.
• FlashList, the list component Shopify open-sourced for React Native, still pulls around 2 million downloads a week, so the framework isn't being abandoned as a project, just as Shopify's own foundation.

What did Shopify actually announce?

Shopify's engineering blog said the company is moving mobile development off React Native and onto native Swift and Kotlin going forward. This isn't a small side project changing direction. Shopify Mobile, the merchant-facing commerce app used by millions of stores, is one of the apps this affects, and it's a sprawling piece of software: 300-plus screens, home-screen widgets, and an Apple Watch app layered on top. The company already used AI-assisted development to rebuild its consumer Shop app in 12 weeks, and it's applying the same approach to the bigger migration.

RelatedDeno 2.9 Ships Desktop Apps and Post-Quantum Crypto

The framing matters here. Shopify isn't saying React Native broke, or that cross-platform frameworks are a mistake in general. It's saying the specific calculation that justified React Native in 2020 no longer holds for them in 2026.

Why did AI coding agents change the equation?

The core assumption behind cross-platform frameworks has always been simple: writing a feature once and shipping it to two platforms is cheaper than writing it twice. React Native, Flutter, and Kotlin Multiplatform all exist because that assumption is usually true. Shopify's engineering post makes the case that generative coding agents erode it. When an agent can look at an existing iOS implementation and port it to Android, or the reverse, the labor of a second native build shrinks toward the labor of a first one.

Shopify described its own agents implementing a feature on Android using the iOS version as a reference, and vice versa. That's a meaningfully different workflow from a human engineer manually translating UIKit patterns into Jetpack Compose. The company has used LLMs in its development process since 2021, so this isn't a sudden pivot. It's a threshold the tooling apparently crossed this year, and Shopify is willing to bet a company-wide mobile rewrite on it.

There's a second argument in the post that has nothing to do with AI: native code keeps Shopify closer to first-party platform tooling, with fewer framework and dependency layers between the app and the OS. That's a long-standing native-vs-cross-platform argument that predates any of this. AI coding agents didn't create it. They just removed the main cost objection to acting on it.

Shopify joins a short list of large consumer apps that tried React Native at scale and walked back, but its stated reason is new. Airbnb and Udacity left because the framework itself was the friction. Shopify is leaving because the alternative to the framework got cheaper, which is a very different kind of argument and a harder one for React Native's maintainers to counter with a framework update.

Who actually built and maintains React Native, and what happens to it now?

React Native is Meta's project, built to let Facebook and Instagram share code between iOS and Android, and it now powers apps well beyond Meta's own products. Shopify wasn't just a user of React Native, it was one of the framework's most visible outside contributors. FlashList, Shopify's high-performance replacement for React Native's built-in list component, is used across the ecosystem and still sees about 2 million downloads a week according to its own numbers. Shopify shipped a ground-up FlashList v2 rewrite for React Native's New Architecture as recently as last year.

RelatedGitHub Moved Copilot Out of the Editor, and That Tells You a Lot

That history is why this reversal carries more weight than a typical framework-of-the-month story. Losing Shopify as a production flagship doesn't threaten React Native's survival. Meta's own apps and a long list of other companies, including Discord and Coinbase, keep building on it. But losing one of your most active outside maintainers as a headline reference customer is a real reputational hit, and other engineering leaders weighing the same build-twice-with-AI-versus-share-a-framework question now have a concrete data point to point to.

What does this mean for React Native and the wider cross-platform market?

Shopify is publicly traded (NYSE: SHOP), but this is an engineering decision, not a financial signal, and nothing here should be read as investment guidance. The more useful read is for the tooling market: Meta continues to fund React Native regardless of what any single company does, so this isn't a solvency risk for the framework. What it does signal is a possible shift in how well-resourced engineering orgs weigh the build-once-cheaply-in-React-Native tradeoff against build-twice-quickly-with-agents, now that the second option has a credible cost story behind it for the first time. Watch whether other large apps that adopted React Native for cost reasons, rather than for reasons like web-team code sharing, start running the same math.

1. 2020Shopify adopts React Native. Chosen to unify iOS and Android development for its largest merchant app.
2. 2021Shopify begins using LLMs in its development process. Predates the current wave of agentic coding tools by several years.
3. 2025FlashList v2 ships for React Native's New Architecture. A ground-up rewrite of Shopify's own list component, still actively maintained at the time.
4. Sep 10, 2026Shopify announces the return to native. Shop app already rebuilt in 12 weeks with AI assistance; the larger Shopify app migration follows.
5. 2026 onwardShopify Mobile's 300+ screens migrate to Swift and Kotlin. The larger, higher-stakes rewrite Shopify is now committing engineering time to.

Our take

The interesting part of this story isn't that Shopify picked native over React Native. Companies flip that coin in both directions every year. It's the reason they gave. Airbnb and Udacity walked away from React Native because the framework itself caused friction, bugs, or slow releases. Shopify is walking away because AI coding agents made the thing React Native was invented to avoid, writing a feature twice, cheap enough that avoiding it stopped being worth the tradeoffs. If that argument holds up under the strain of an actual 300-screen migration, it's a preview of a broader rebalancing: cross-platform frameworks earned their place by saving engineering hours, and agentic coding is the first real threat to that specific value proposition in a decade. Worth remembering that Shopify still ships FlashList to the rest of the React Native world even as it leaves. That's not a company burning a bridge. It's one making a bet, in public, that it's willing to be wrong about.

Original analysis by GenZTech.