Anthropic Loses Claude Code Product Chief Cat Wu Ahead of IPO

Cat Wu, head of product for Claude Code and one of the two public faces of Anthropic's breakout tool, left the company on Friday — weeks before Anthropic's initial public offering, which Business Insider reports is expected later this month.

Wu announced the exit Tuesday afternoon on X, thanking the Claude Code team and its users and reflecting on how far the tool has come. She did not say what prompted the departure and did not immediately respond to requests for comment. Anthropic confirmed she has left.

The timing is awkward, to put it mildly.

Why Wu's departure matters

As head of product for Claude Code, Wu served as one of the tool's two top leaders and evangelists alongside its original creator, Boris Cherny. Under that pair, Claude Code became what Business Insider calls one of the most important products in the AI industry's history.

The numbers logic is straightforward. Claude Code gave Anthropic the viral product it needed to lift revenues and unlock more funding for compute. It also turned Anthropic into a bigger rival for OpenAI. Tech giants' AI bills with Anthropic soared as developers burned through Claude tokens on entire coding tasks.

She's the second Claude Code veteran to walk recently: Thariq Shihipar, who also worked on the tool, announced his own departure last month.

What's the backstory on Wu and Claude Code?

Wu joined Anthropic in summer 2024 and tested an early prototype of Claude Code. In Tuesday's post, she recalled that moment plainly: "It worked, I was hooked, and I joined the team."

Anthropic released Claude Code in research preview in February 2025, announcing it in a brief demo video that featured Wu and Cherny. That summer, the two decamped to AI coding startup Cursor — then returned to Anthropic a couple of weeks later, The Information reported.

The climb accelerated at the end of 2025. The tool first took off as an automatic coder for software engineers, who suddenly had a product that could sprint through entire coding tasks.

Wu also oversaw Cowork, Anthropic's tool for non-coding work, which runs on the same AI models and handles knowledge work for non-software engineers. In an April interview on Lenny's Podcast, she described her job as charting the path to the "AGI-pilled" version of Claude Code that Cherny envisioned — a nod to the industry term "artificial general intelligence."

She said she worked with marketing, sales, finance and other teams, hired some of Anthropic's product managers, and regularly answered user requests about Claude Code on social media.

What happens now?

Wu hasn't said where she's headed. When one commenter on her X post asked, "Are you gonna take that 6 months vacations now?", she answered with two words: "post agi!"

Cherny remains, and Claude Code's momentum doesn't hinge on one departure. But losing a product leader who doubled as the tool's public evangelist — right before an IPO where Anthropic will pitch investors on exactly that revenue engine — is a timing problem Anthropic didn't need.