Who controls powerful AI? Anthropic has a plan for long after its leaders depart

Anthropic’s planned public-company structure is designed to preserve its mission as it develops increasingly powerful AI. The company expects a valuation of about $1.5 trillion and is organized as a Delaware public benefit corporation, which must weigh investor interests against broader social goals. Its IPO filing describes safeguards intended to limit pressure from shareholders: seven co-founders hold 50.1% of the votes on certain matters, including shareholder proposals and elections for some director seats. A separate long-term benefit trust holds special shares that give it the power to elect a majority of the board.

The filing also sets out how control could change as founders leave and their supervoting shares disappear. The board would then be split into three groups, with directors serving staggered three-year terms - a structure less responsive to shareholders than the annual elections now common at large U.S. companies, and one historically used to deter hostile takeovers. The trust’s arrangements are meant to preserve flexibility while keeping directors bound to Anthropic’s public-benefit purpose. Because the trust controls the board, Anthropic qualifies as a “controlled company” under Nasdaq rules and can use exemptions from some governance requirements, including placing President Daniela Amodei on a committee despite her relationship to CEO Dario Amodei. A board safety and risk committee includes Novartis CEO Vas Narasimhan and former White House official Chris Liddell.