Anthropic is scrambling to contain fallout from a source code leak, an unwelcome setback as the company eyes an initial public offering as early as this year. The incident has made the company particularly sensitive given its ongoing IPO race with rival OpenAI.
GitHub, the developer code-sharing platform, said on July 1 that Anthropic had filed a mass takedown request targeting 8,100 repositories that had cloned and stored source code from "Claude Code," the company's core development tool. The move is seen as Anthropic's effort to halt the further spread of proprietary information across the internet. The leak put the company at risk of exposing key trade secrets and know-how to competitors.
According to Chaofan Shou, chief technology officer of U.S. security firm Puzlland, and others, the Claude Code source code had been exposed through NPM, a package repository used by developers worldwide. The leaked code amounted to more than 512,000 lines across 1,900 files. The exposed code quickly spread to code-sharing platforms including GitHub.
Anthropic has stressed that the leak was not the result of hacking or a security breach. The company issued a statement saying, "No sensitive customer data or credentials were leaked or exposed." It added, "This was a product release packaging issue caused by human error, not a security breach, and we are implementing measures to prevent a recurrence."
This marks the second leak of key Anthropic assets within the past week. On June 25, a configuration error in the company's content management system temporarily exposed specifications related to "Claude Mythos," a next-generation AI model that has not yet been officially released.
Anthropic maintains that sensitive customer data and authentication credentials were not affected by the incidents. However, industry observers have raised concerns that the leaks could affect the company's IPO planned for the fourth quarter of this year. Particular attention is being paid to whether the incidents will influence the competitive dynamic with rival OpenAI, which is also expected to go public this year. The market estimates Anthropic's valuation at $380 billion (approximately 567.46 trillion won) and OpenAI's at $852 billion (approximately 1,292.65 trillion won).