Anthropic Loses Claude Code Product Lead Cat Wu Weeks Before IPO

By Olivia Hart

Anthropic has confirmed that Cat Wu, product lead for Claude Code, is leaving the company just weeks before a planned IPO announced for this month. She announced the move herself on X on Tuesday afternoon, thanking the Claude Code team and its users and pointing to the progress the tool has made. The company confirmed her departure.

What she did not do is explain it. Wu gave no reason for her decision and did not immediately respond to media inquiries. That silence matters: she was, alongside Boris Cherny, one of the two public faces of Claude Code, and her exit lands directly before the most consequential event in Anthropic's corporate calendar.

Who is leaving, exactly?

Wu joined Anthropic in the summer of 2024 after testing an early prototype of Claude Code, which she described as compelling in her Tuesday message. By the time of her departure she carried a broad mandate:

• Product lead for Claude Code, the tool that positions Anthropic as one of the most significant recent players in the AI industry;
• Responsibility for Cowork, Anthropic's product for knowledge work unrelated to coding, built on the same AI models;
• Cross-functional work spanning marketing, sales and finance, plus recruiting several product managers;
• A regular presence answering user questions about Claude Code on social media.

In April, she laid out her roadmap in her own words: an "AGI-pilled" version of Claude Code, a concept she attributed to Cherny. That ambition now passes to someone else, whom Anthropic has not named. The takeaway: Wu was not a peripheral engineer but the product's chief evangelist, and her portfolio has no announced successor.

Is this an isolated exit?

It is not. The pattern around Claude Code runs deeper than one departure.

Wu and Cherny themselves left Anthropic for the startup Cursor in the summer of 2025, returning only a few weeks later. Last month, Thariq Shihipar, also involved in Claude Code's development, announced his own departure. Anthropic has not publicly addressed what is driving the turnover.

After their return from Cursor, Wu and Cherny oversaw a growth phase: Claude Code gained momentum at the end of 2025. Reports also indicate that AI bills of major tech companies using Anthropic have surged as their consumption of Claude tokens climbed — a signal of real, paid adoption rather than hype alone. The plain reading: repeated departures at the top of a flagship product, weeks before an IPO, invite questions the company has so far declined to answer.

How did Claude Code get here?

The tool's trajectory was fast by any standard. It launched in a research version in February 2025, showcased in a short demonstration video featuring Wu and Cherny. Claude Code initially established itself as an automatic coder for software engineers, capable of executing entire coding tasks quickly.

It is presented as one of the most significant products in the recent AI industry, and it is said to have given Anthropic a viral product that boosted revenue and helped secure additional funding for computing capacity. Those claims — "said to have," "presented as" — come without disclosed figures. Neither Anthropic's IPO valuation nor Claude Code's standalone revenue has been made public in the reporting available.

What does this mean for the IPO?

The timing is the story. A product lead exiting weeks before a massive IPO, without a stated reason and without a named successor, is the kind of detail investors scrutinize — especially when the product in question is described as central to the company's revenue growth and fundraising story. Anthropic says nothing beyond confirming the departure, and Wu says nothing beyond thanks. Until either side fills that gap, the market will price the silence itself: the tool's momentum is documented, but the stability of the team behind it is now, on the evidence of the past year, an open question.