React Compiler reaching 1.0 is a major signal, but the right adoption model is incremental. The compiler can remove a lot of manual memoization pressure, yet it also makes previously hidden Rules of React violations more visible.
The path I would use in a production app starts with linting. Upgrade the hooks lint setup first, fix diagnostics that indicate unsafe render or effect behavior, and only then enable compiler transforms in a measured slice. Compiler output should be treated like a performance optimization layer, not a semantic dependency.
The best candidates are component surfaces with heavy derived values, repeated rendering, and enough test coverage to catch effect timing mistakes. The worst candidates are areas with hidden mutation, broad context churn, or hand-rolled subscription logic nobody wants to touch.
The practical migration checklist:
- Pin compiler versions during rollout.
- Keep existing memoization until measurement says otherwise.
- Add interaction-level tests around the first compiled flows.
- Measure render cost before claiming success.
The compiler is exciting because it rewards ordinary React code. The work is making sure the codebase is ordinary enough for that promise to hold.
Official source: React Compiler v1.0.