Anthropic’s Long-Term Benefit Trust is an independent group of financially disinterested trustees that holds a special class of company stock giving it the power to elect and remove a growing share of Anthropic’s board of directors — up to a majority. It exists to keep the company’s safety mission from being overridden by ordinary commercial pressure, even as Anthropic raises billions of dollars and races to build ever more capable AI systems.
What it is
Most companies answer to their shareholders alone. Anthropic, incorporated as a Delaware public benefit corporation, already lets its directors weigh a stated public-benefit purpose — “the responsible development and maintenance of advanced AI for the long-term benefit of humanity” — alongside profit. The Long-Term Benefit Trust (LTBT), set up when Anthropic adopted that structure in 2023, adds a second layer on top: a standing panel designed to hold five trustees, chosen for expertise in AI safety, national security, public policy or economics rather than for their business ties to Anthropic. Trustees own no Anthropic equity, receive no share of profits, and are compensated only for their time.
How it works
The Trust’s authority comes from a special class of shares — Class T stock — that Anthropic issued specifically for this purpose. Owning that stock lets the Trust elect a defined, growing number of Anthropic’s directors: starting with one of the company’s five board seats, then expanding to two, and eventually to three — a majority — as set time periods pass or fundraising milestones are hit. The stock carries little economic value; its purpose is voting power, not dividends. The Trust does not run Anthropic day to day or approve individual product launches. Its mandate is narrower and longer-range: appointing directors it trusts to weigh safety seriously, and being consulted on decisions with major implications for AI risk or society.
Trustees are not picked by Anthropic’s executives. The initial group was chosen after a year-long search by Anthropic’s board, and from there the Trust elects its own successors, each serving renewable one-year terms. As of mid-2026 it has four seated trustees: Neil Buddy Shah (chair of the Trust and CEO of the Clinton Health Access Initiative), national-security expert Richard Fontaine, international-affairs scholar Mariano-Florentino Cuéllar, and economist Ben Bernanke, who chaired the U.S. Federal Reserve through the 2008 financial crisis and won the 2022 Nobel Memorial Prize in Economic Sciences.
Why it matters
AI labs have repeatedly promised that safety commitments will hold even under commercial pressure, and then faced skepticism when leadership or funding structures changed. Anthropic’s answer is to make some of that commitment structural rather than just a promise: a body that cannot profit from Anthropic’s success, cannot be replaced at will by executives, and is explicitly charged with protecting a mission rather than a share price. It is not the only such experiment — other AI developers have also tried board-level or nonprofit oversight of a for-profit AI business — but the LTBT is one of the more concrete, publicly documented versions, with a specific stock class and a defined schedule for gaining real board control.
Whether it works as intended is untested at scale: the Trust has never had to remove a director or block a major decision, and its full three-seat majority has not yet arrived. Its value, for now, rests on who sits on it and how seriously Anthropic’s board treats its input — which is why each new appointment, including Bernanke’s, draws attention.
In the news
Anthropic’s newest trustee is Ben Bernanke, added to bring economic expertise to the Trust’s oversight of a technology many expect to reshape labor markets and financial systems — see our report on the appointment.
FAQ
Do the trustees own Anthropic stock or get paid from its profits? No. Trustees hold no equity and receive no profit share; they are compensated only for their time and service, by design, so they have no personal financial stake in Anthropic’s valuation.
Can the Trust fire Anthropic’s CEO or block a product launch? No. It doesn’t manage day-to-day operations. Its power is to elect and remove a defined, growing share of the board of directors, and to be consulted on major decisions — not to run the company directly.
Is this the same as Anthropic’s public benefit corporation status? No, though the two work together. The public-benefit-corporation structure lets Anthropic’s board legally balance profit with its stated mission; the Trust is a separate, independent body layered on top with its own stock class and the power to shape who sits on that board.
Sources: Anthropic: The Long-Term Benefit Trust; Anthropic: Ben Bernanke appointed to the Long-Term Benefit Trust; Harvard Law School Forum on Corporate Governance.