Anthropic’s chief financial officer Krishna Rao says artificial intelligence is already doing most of the coding work inside the company, signalling how rapidly white-collar jobs may evolve from execution-heavy roles into AI-supervision roles.

Must read: Explained: Why Anthropic’s new AI tool is spooking India’s IT services industry

Speaking on Patrick O’Shaughnessy’s “Invest Like the Best” podcast, Rao said more than 90% of Anthropic’s code is now written using Claude Code, the company’s AI-powered coding assistant. 

“90 plus percent of our code is actually written by Claude Code,” Rao said during the podcast. 

Rao described a workplace where AI systems increasingly handle large parts of software engineering, financial reporting and operational workflows, while human employees focus more on oversight, judgment and strategic decision-making.

“We’ve hired a lot more people because of that,” Rao said, referring to AI-driven productivity gains. “That actually means that we can get a lot more done.” 

The comments come at a time when companies globally are debating whether AI will replace white-collar workers or significantly augment them. Anthropic’s internal use case offers one of the clearest indications yet of how frontier AI labs themselves are restructuring work around AI agents.

Must read: BT explainer: Why Anthropic’s Claude Mythos could become a major security risk?

Rao said Anthropic’s finance team has also adopted Claude extensively. According to him, the company now uses AI systems to generate financial statements, while its monthly review process is already “90 to 95% ready” before human intervention. 

“What used to take hours to produce now comes down to 30 minutes,” Rao said, referring to internal reporting workflows. 

He added that AI is shifting employees away from manually gathering information and toward interpreting and acting on insights generated by AI systems.

“Everyone kind of becomes a manager,” Rao said. “The implications of that and the productivity gain that can come from that, we’re very, very early in that, but the potential for it is incredible.” 

Anthropic, one of the world’s leading AI startups and the creator of the Claude family of models, is currently preparing for a potential public listing. The company has already reached an estimated valuation of around $90 billion ahead of any IPO plans. 

During the podcast, Rao also highlighted how central computing infrastructure has become to Anthropic’s operations, calling compute resources the “lifeblood” of the business. 

“The compute that we procure is the lifeblood of our business. It is the most important thing in the company,” Rao said. “It is the canvas on which everything else gets built.” 

He revealed that Anthropic has signed large-scale infrastructure deals with Google and Broadcom for TPU deployments beginning in 2027, alongside a separate agreement with Amazon for Trainium chips. Rao described the broader compute commitment as exceeding $100 billion. 

Rao said he currently spends nearly 30% to 40% of his time on compute-related decisions, underscoring how AI infrastructure has become a strategic priority for frontier model companies. 

While optimistic about AI’s long-term potential in areas such as healthcare and productivity, Rao also cautioned against ignoring risks associated with rapid technological change.

“I don’t think we should just tell everyone everything is going to be great because there are likely to be bumps on the road,” Rao said.