Cat Wu, the product lead who helped steer Claude Code from an internal prototype to an industry-defining tool, has left Anthropic. Her departure, confirmed by the company and announced on X on Tuesday afternoon, comes at a precarious moment as Anthropic prepares for a major initial public offering later this month. Wu served as one of two primary leaders of the product, working alongside original creator Boris Cherny to turn Claude Code into a critical revenue driver for the AI lab.

The Rise of Claude Code

Wu joined Anthropic in summer 2024 after testing an early version of the coding agent, noting in her farewell post, "It worked, I was hooked, and I joined the team." Under her and Cherny's leadership, the tool launched in research preview in February 2025 and subsequently exploded in popularity among software engineers. The product’s success allowed Anthropic to capture significant market share against rivals like OpenAI, driving up compute costs and revenue through heavy developer usage. The team’s trajectory was not without drama; in the summer of 2025, Wu and Cherny briefly departed for AI coding startup Cursor before returning to Anthropic just weeks later.

Strategic Ambiguity and Departure

The timing of Wu’s exit is notable given Anthropic’s aggressive push toward an IPO. She did not specify the reasons for leaving and did not immediately respond to requests for comment. Her role extended beyond coding; she also oversaw Cowork, Anthropic’s tool for non-coding knowledge work. In an April interview on Lenny's Podcast, Wu described her mission as guiding the product toward the "AGI-pilled" vision Cherny held for the future, bridging the gap between technical development and commercial viability. Her departure follows that of Thariq Shihipar, another key contributor to Claude Code, who left last month, raising questions about stability within the core product team during this critical financial phase.

Key Takeaways

  • Cat Wu led product for Claude Code alongside creator Boris Cherny, driving its adoption as a viral AI tool.
  • The exit occurs weeks before Anthropic’s expected IPO, creating potential uncertainty for investors.
  • Wu previously tested the prototype in 2024 and joined after a brief stint at Cursor in 2025.
  • Her departure follows that of colleague Thariq Shihipar, signaling potential churn in the core team.
  • Wu also managed Cowork, Anthropic’s knowledge-work tool, indicating her influence spanned multiple product lines.

The Bottom Line

Anthropic's IPO narrative relies heavily on the momentum of Claude Code, and losing its primary product evangelist alongside other key staff introduces unnecessary volatility into the investor story. While Wu's exit may not halt the product's technical trajectory, it removes a crucial bridge between the engineering vision and commercial execution at the most sensitive moment in the company's history.