TLDRocket
Sign in

Shopify spent years on React Native — then rebuilt everything in 12 weeks

The New Stack Amanda Caswell Covered by 2 sources

Shopify ditched React Native and rebuilt its Shop app natively in 12 weeks. It’s betting AI agents can now do the rewrite work that used to make cross-platform feel worth it.

Based on reporting by The New Stack, Amanda Caswell — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

Shopify spent years making React Native pay off. In 2020, it chose to write mobile code once instead of maintaining separate Swift and Kotlin versions, and that approach held long enough to shape the company’s app strategy for years. On Thursday, Shopify said the balance had shifted. It is moving its mobile apps back to fully native development, with AI agents doing much of the translation work.

The first proof point was Shop, the consumer app. Shopify says it went from proof of concept to a native production release in 12 weeks. That is a fast turnaround for a reversal this big, and it sets up the harder job: the merchant app, which has more than 300 screens and depends heavily on deep iOS integration.

Shopify isn’t calling its earlier React Native bet a mistake. That framework did the job for a long time, and the company was still talking publicly about its future with React Native as recently as January 2025. What changed, according to director of engineering Mustafa Ali, is the assumption underneath the old decision. By late 2025, coding agents had gotten good enough that they could inspect how a feature worked on iOS and build the Android side, or do the reverse. Engineers could also work on platforms they knew less well because the agent could shoulder more of the platform-specific work.

The rewrite still wasn’t as simple as asking a model to turn one app into another. Shopify tried feeding an LLM the existing React Native code and asking for a native version, but the output was junk, in Ali’s blunt description. Adding specs and task files before the code helped some, but not enough. So the company built Helix, an internal system that chops the migration into screen-sized pieces, compares each result with the old app, and uses separate agents to flag problems before a human signs off.

That caution lines up with a broader theme in Shopify’s recent AI work: speed is useful, but review still matters. The company also had to work around a practical snag, since agents can edit code in seconds while mobile simulator testing can take far longer. Shopify split business logic from the UI and let agents work through a desktop CLI, which can run some checks in milliseconds and still control a simulator when needed. The company now says it will judge the native apps partly by how much of the work agents can eventually do alone.

My take — AI-written commentary, not fact-checked reporting

This is the part everyone should be paying attention to: AI is not just speeding up coding, it is lowering the cost of changing a bad architecture decision later. That is great news for teams that enjoy undoing their own past with fewer meetings and more machines. It also means the next shiny “platform bet” may come with a shorter half-life than the last one.

Read more about this at: The New Stack

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.