Jacob Coxon left Anthropic just four months after joining and two months before his company equity was due to vest. #
Artificial intelligence researcher Jacob Coxon walked away from his entire unvested equity stake at Anthropic when he resigned over concerns about the safety of increasingly powerful AI systems, leaving the company just two months before his shares were due to vest.
Coxon had worked at Anthropic for four months. Employees must remain at the company for six months before their equity begins vesting, he told Axios in an interview. "I no longer have anything to gain by juicing up Anthropic's valuation," Coxon said, noting that he left two months before any of his equity vested.
The size of Coxon's equity grant was not disclosed. Anthropic was valued at $965 billion in its latest funding round in May, when it raised $65 billion from investors including Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital.
Coxon announced his resignation Tuesday in a lengthy post on X that has since drawn more than 100 million views. He said he had spent three years working on pretraining AI models at OpenAI and Anthropic and accused both companies of moving too aggressively in their pursuit of increasingly capable systems.
"The people building AI earnestly believe that it could kill us all by the end of the decade," Coxon wrote. "This is not a marketing stunt."
His concerns center partly on the competition among major AI developers and the difficulty of monitoring increasingly capable models. Coxon told the outlet that he had not seen Anthropic compromise its safety practices while he worked there, but said competitive pressure could eventually lead companies to "cut corners" or "skip steps in the oversight process."
He also said AI models have become increasingly capable of recognizing when they are being evaluated. Systems knowing when they are being tested, something that once appeared far removed from current AI capabilities, is now a regular part of working with advanced models, Coxon said.
Coxon's warnings have received public support from other researchers at Anthropic. Evan Hubinger, who leads the company's alignment science work, responded to Coxon's resignation post by saying he agreed that AI could pose an existential threat and put the probability of AI killing all humans within the next decade at above 10%.
Anthropic itself acknowledges that increasingly powerful AI systems can create catastrophic risks. The company's Responsible Scaling Policy, most recently updated in August, sets out safeguards intended to increase as its models become more capable and specifically addresses risks involving chemical and biological weapons, autonomous AI research and other potentially dangerous capabilities.
The company also maintains a Frontier Safety Roadmap covering security, safeguards and alignment. Anthropic says future AI systems are likely to become more powerful in ways that could undermine assumptions behind current safety techniques, requiring stronger methods for training, evaluating and monitoring advanced models.
Anthropic has been expanding rapidly, with the company saying its annualized revenue had surpassed $47 billion when it announced its May funding round. The new capital was earmarked in part for safety and interpretability research, as well as computing infrastructure and product expansion.
Anthropic has also continued developing automated methods for identifying and addressing potentially unsafe model behavior. In research published Aug. 28, the company said it had tested automated auditing tools against alignment failures including deception, sycophancy and jailbreaks as it examines how safety research can keep pace with AI systems increasingly being used to develop other AI systems.
Coxon previously spent about three years at OpenAI before moving to Anthropic. He added that he still owns equity in OpenAI, distinguishing that stake from the Anthropic shares he forfeited when he left before the vesting threshold.
His resignation adds to a series of departures from major AI companies tied to concerns about safety. Researchers have previously left OpenAI and other AI labs after disagreements over how companies are balancing the development of more capable systems with work intended to keep those systems under control.
Coxon said his decision to leave Anthropic came as AI capabilities advanced and competition among developers intensified. While he said Anthropic had not sacrificed safety during his time there, he argued that relying on individual companies to slow development would be difficult while competitors continued moving ahead.
The debate over the risks of advanced AI has also reached Congress, with Sen. Bernie Sanders, I-Vt., responding publicly to Coxon's warning Wednesday and saying he plans to introduce legislation seeking to halt the development of artificial superintelligence.
© Copyright IBTimes 2026. All rights reserved.