Cathedral, a new startup founded by members of the now-defunct Department of Government Efficiency (DOGE), claims it breached sensitive systems at AI giant Anthropic.
According to a draft report from Cathedral reviewed by the Washington Post, the startup notified Anthropic that its AI agents had found ways into the company’s internal systems by exploiting vulnerabilities in third-party business software. Cathedral said its agents took less than 12 hours to find paths to sensitive information, including meeting transcripts, salaries and job offers, and Slack messages. However, Anthropic said the researchers had previously spent several days unsuccessfully trying to breach its software directly.
Anthropic confirmed that the researchers gained access to a “limited set” of its sales, human resources, and IT systems, including customer and partner business contact information and some basic user account records. However, the company said the researchers did not access Claude users’ conversations, source code, or AI model weights. The hack was part of a cybersecurity testing program in which Anthropic rewards researchers who find vulnerabilities, but the company accused Cathedral of violating the program’s rules by failing to disclose the full extent of its activity. Anthropic did not immediately respond to Gizmodo’s request for comment.
Cathedral is an AI-driven cybersecurity startup focused on military applications founded by former DOGE staffers Gavin Kliger, Luke Farritor, Marko Elez and Jack Stein. Kliger was also previously the Pentagon’s chief data officer and was reportedly involved in the Pentagon’s high-profile fallout with Anthropic. Another co-founder, Elez, had to resign from DOGE after he was linked to some ridiculously racist tweets like “normalize Indian hate” and claiming that he was “racist before it was cool,” though Musk promptly rehired him.
The company has been valued at $1.4 billion and raised $160 million from investors including Andreessen Horowitz and Sequoia Capital. Trump allies Marc Andreessen and Shaun Maguire, who are associated with the respective firms, sit on Cathedral’s board.
The hack reportedly raised national security concerns in Washington, according to an unnamed senior Trump administration official, with fears that similar software vulnerabilities can be abused by Chinese hackers.
“The fact that weaknesses in ordinary enterprise software can compromise the infrastructure of even the most advanced organizations should be a wake-up call for every company in America,” Cathedral’s report says, according to the Post.
Massive AI-driven hacks have dominated headlines for months, but Anthropic’s tools have usually been the assistants to hacking rather than the systems being hacked. Recently, a Chinese-speaking hacker reportedly used AI agents powered by DeepSeek and Anthropic’s Claude Code to break into more than 100 business websites over several weeks. The campaign targeted more than 1,200 websites, although the extent of any data theft remains unclear, according to Reuters.
Anthropic is one of the leading AI labs in the United States and is creating some of the most advanced models in the entire world. Many users, including major corporations, governments, and everyday people, rely on Anthropic’s systems, and any software vulnerabilities could have major implications for the private data and operations of these users. Fellow leading AI lab OpenAI was also recently targeted in a so-called ethical hack, which proved software vulnerabilities that exposed several OpenAI employees’ ChatGPT accounts.
But Washington’s fears hover more around AI lab espionage. Government officials fear that if Chinese hackers can get hold of Anthropic’s trade secrets, they could replicate and perhaps surpass the capabilities of Anthropic’s models. AI lab espionage has become a hot topic, especially after a former OpenAI engineer was accused of secretly planting a recording device in a restricted conference room. The Saudi national was allegedly in contact with an unnamed member of the Saudi royal family, and the FBI is now exploring the possibility that the engineer was trying to steal information from the leading AI lab to offer it to the Saudi government.
