October 28, 2026
Anthropic PBC’s registration statement includes more than 90 pages of risk disclosures for prospective investors ahead of the Claude developer’s planned public offering, including “existential risks to humanity” posed by advanced artificial intelligence and a corporate governance structure the company warns may make it difficult for voting shareholders to replace directors and managers.
The sprawling range of risks laid out in the AI giant’s Monday filing with the U.S. Securities and Exchange Commission acknowledges the company’s increasing use of AI to advance its own research and development may lead to the technology outpacing Anthropic’s ability to keep up with safety and security measures, and the registration statement points to broader worries about oversight of the technology.
. . .
Corporate Governance
Anthropic warned Monday that its governance structure may bring its own risks, calling the setup “a novel capital and corporate governance structure that has no direct public company precedent.”
The company is incorporated as a public benefit corporation under Delaware law, and it has created a long-term benefit trust of “financially disinterested trustees with expertise in AI safety, national security, public policy and social enterprise,” which can elect a majority of the company’s board of directors, per the registration statement. Among other things, Anthropic also plans to issue a special class of stock after its IPO to an LLC consisting of the company’s seven co-founders, entitling that group to 50.1% of the voting power of company shares for matters that come to a shareholder vote.
. . .
Jan Messerschmidt, a securities litigation partner at Cohen Milstein Sellers & Toll LLP, said it isn’t new for tech company founders to have arrangements that allow them to have majority control, including through dual class stock. But conventional structures often see the founders reducing their control over time, he said.
Anthropic’s structure is “very different,” he said, because the seven founders’ majority control is fixed and in perpetuity, so long as they hold 2% of their IPO holdings.
“I’ve never seen that before,” Messerschmidt said.
Anthropic’s registration statement gives the impression that the company’s governance structure is designed to further its public benefit corporation aspect, which requires it to balance its shareholders’ financial interests with the interests of those affected by the company’s conduct and its overall purpose as a public benefit. But Messerschmidt questions what will happen when those interests diverge or when shareholders disagree and aren’t able to use their votes.
He said Anthropic is “basically asking investors to buy stock in a company that they will never have a say in,” and he’s keeping an eye on how the governance structure may be interpreted by courts.
“Protection of shareholders will come from disclosure, liability and fiduciary duty law, not the ballot box,” Messerschmidt said. “One of the risks that the company acknowledges is that the fiduciary duty law here is largely uncharted. So it’s uncharted waters that we’re entering into.”
Read Anthropic’s ‘Existential Risks’ Warning Sets Stage for AI IPOs.