Claude Code now runs Bun written in Rust

In mid-July 2026, it became clear that Claude Code had quietly adopted a major architectural change: Bun, the JavaScript runtime used by the tool, had been rewritten in Rust. Jarred Sumner, the creator of Bun, announced this milestone in a post titled "Rewriting Bun in Rust," noting that Claude Code v2.1.181 (released June 17th) and all subsequent versions now run the Rust port. The change yielded a 10% startup speed improvement on Linux, though Sumner noted that "barely anyone noticed"—a testament to how seamlessly the transition occurred.

To verify the claim, an independent examination of the Claude Code binary revealed telling evidence. Searching the binary for version strings turned up "Bun v1.4.0 (macOS arm64)," a preview version that has not yet been publicly released on GitHub; the most recent official Bun release available at the time was v1.3.14 from May 12th. A second search for Rust source file references in the binary returned a list of 563 filenames, confirming that Rust code—including modules for the development server, production build system, and bundle system (src/runtime/bake/dev_server/mod.rs, src/runtime/bake/production.rs, src/bundler/bundle_v2.rs, and hundreds more)—was indeed compiled into the production binary.

The implications are significant: Bun written in Rust is now running in production across millions of different devices through Claude Code installations worldwide. The Rust rewrite represents a substantial engineering effort, involving hundreds of source files and a complete reimplementation of Bun's core functionality. The fact that the deployment was nearly invisible to users—"Boring is good," as Sumner put it—reflects successful infrastructure engineering: the change delivered measurable performance gains while maintaining compatibility and stability.

The migration of Bun to Rust represents a significant infrastructure upgrade within Claude Code, yet it unfolded almost invisibly to end users. By the time v2.1.181 shipped on June 17th, the Rust port was stable enough to ship in production across millions of devices simultaneously. The fact that the binary contains 563 Rust source files underscores the scale of the rewrite; this was not a minor refactoring but a substantial reimplementation of core runtime functionality. The 10% startup speed improvement on Linux provides a tangible performance win, and the deployment of a preview Bun version (v1.4.0) before its public release suggests confidence from the development team in the stability and readiness of the Rust implementation. The author's observation that "boring is good" reflects a shift in how infrastructure success is measured—not by visibility or hype, but by silent, reliable performance.

No comments yet. Be the first to share your thoughts!