Anthropic CFO Says Claude Writes Most Company Code, Yet Hiring Keeps Growing

AI Is Taking Over the Execution Layer

Anthropic CFO Krishna Rao says AI now writes more than 90% of the company's code, but the result has not been a smaller workforce. Instead, Anthropic has continued hiring because Claude is letting teams move faster and take on more work.

Speaking on Patrick O'Shaughnessy's Invest Like the Best podcast, Rao described Claude as a productivity accelerant inside Anthropic. Rather than replacing employees outright, the company's AI systems are handling more of the execution layer across software development and internal business workflows.

That shift changes what human teams spend time on. Engineers and operators increasingly focus on judgment, review, strategy, and directing AI systems instead of manually producing every first draft themselves.

Claude Code Handles Most Programming Work

According to Rao, more than 90% of Anthropic's code is now written by Claude Code. The finance team is also using AI heavily: financial statement drafts can be prepared to roughly 90% to 95% completion before human review.

Work that previously took hours can now be compressed into about 30 minutes in some finance workflows. The key point is not that humans disappear from the process, but that they review, refine, and supervise increasingly capable AI-generated output.

Why More Automation Can Mean More Hiring

Rao's comments challenge the simple idea that AI productivity automatically leads to fewer jobs. His argument is that once AI makes each employee more capable, the company can pursue more projects, expand faster, and justify hiring additional talent.

He framed Anthropic's approach around "talent density" — smaller, highly skilled teams equipped with powerful AI tools. In that model, employees become managers of AI agents, coordinating fleets of systems that can work across multiple tasks in parallel.

This is a very different story from companies using automation mainly as a cost-cutting justification. Anthropic's view is that there is still plenty of work to do, and AI increases the amount of useful work teams can realistically attempt.

The Bigger Signal for White-Collar Work

Anthropic's internal use of Claude offers a preview of how AI may reshape white-collar jobs more broadly. Routine drafting, coding, analysis, and reporting are becoming increasingly automated, while human value shifts toward oversight, decision-making, quality control, and domain expertise.

That does not remove the disruption risk. Many roles will likely change fast, and companies will need to rethink workflows, skills, and management structures. But Rao's comments suggest one possible path: AI does not only reduce headcount; it can also increase ambition.

The Bottom Line

Anthropic is one of the clearest examples of a company building AI and aggressively using it internally. If Claude Code can already produce most of the company's code while hiring continues, the future of work may be less about humans versus AI and more about which organizations learn to manage AI-native teams first.

---

Source: The Times of India