Why one Anthropic update wiped billions off software stocks

Tech workers have been worried for years about the AI tidal wave coming for their jobs, but their bosses are starting to worry now, too. 

Stocks plunged this week as fears escalated that AI advancements will take a bite out of business for many software, data, and professional services companies. The market losses are tied to updates to Anthropic’s AI-powered workplace productivity suite, Claude Cowork, which threatens to replace some software tools ubiquitous in the professional world.

Shares in companies involved in research and legal software, such as Thomson Reuters and LegalZoom, dropped dramatically on the Anthropic news, with a wide swath of software stocks following suit. And shares in Intuit, PayPal, and Equifax all dropped by over 10%, with enterprise software companies like Atlassian and Salesforce deepening their own losses, which started well before the latest AI news. The S&P North American Software Index also slid further this week, worsening a recent losing streak punctuated by a 15% decline in January—the index’s worst month in nearly two decades.

Unlike Claude Code, a coding tool designed for developers, Anthropic built Claude Cowork as a powerful, general-purpose AI agent for non-coders. Available to Anthropic’s $100-per-month premium subscribers, Claude Cowork can knock out easier tasks like searching, collecting, and organizing files. But it’s also capable of taking on much bigger challenges such as making slide decks, producing reports, and pulling and synthesizing information from other business software tools, like Zendesk and Microsoft Teams. 

Claude’s ability to execute complex tasks with dedicated software sub-agents prompted plenty of nervous jokes about humans being replaced by C-suites full of AI. And that was before a new Anthropic update introduced powerful new plug-ins designed to automate tasks across domains like finance, legal, sales, data, marketing, and customer support.

The market is still digesting those new agentic AI capabilities, which could pose an existential threat to the software-as-a-service (SaaS) companies that undergird big chunks of the economy. 

Fears of a zero sum software game grow

Anthropic cofounder and CEO Dario Amodei has made his own ominous predictions about AI displacing human workers. Last year, Amodei predicted that AI could vaporize half of entry-level white-collar roles, sending unemployment as high as 20% within five years. He pointed to losses in industries such as tech, law, consulting, and finance. “We, as the producers of this technology, have a duty and an obligation to be honest about what is coming,” Amodei told Axios. “I don’t think this is on people’s radar.”

Not everyone deeply invested in AI agrees. Nvidia CEO Jensen Huang swatted away worries that AI would eat the traditional software industry after the stock bloodbath that began on Tuesday. “There’s this notion that the tool in the software industry is in decline, and will be replaced by AI,” Huang said, emphasizing that relying on existing software tools makes more sense than reinventing the wheel. “It is the most illogical thing in the world, and time will prove itself.”