Anthropic IPO Governance: Founders and AI Safety Trust Keep Control
Anthropic’s IPO structure gives founders and an AI safety trust board control, leaving public investors little power until strict sunset conditions are met.
Summary
On October 27, 2026, Anthropic, founded in 2021 to make AI controllable and beneficial over the long term, is approaching an IPO that will give public investors little authority. Its public benefit corporation status lets directors balance profit with AI safety and protects mission-based decisions from shareholder lawsuits. The board, with one vacancy, is chaired by president Daniela Amodei and includes her brother, CEO Dario Amodei, Netflix cofounder Reed Hastings, former Microsoft and GM CFO Christopher Liddell, Novartis CEO Vas Narasimhan and Spark Capital partner Yasmin Razavi.
An independent Long-Term Benefit Trust owns Class T stock empowering it to elect and remove four directors, a board majority. Chair Neil Shah, formerly GiveWell managing director, serves with former Federal Reserve chair Ben Bernanke, national security expert Richard Fontaine and Harvard physicist Christopher Stubbs. Existing members will fill its one vacancy. Anthropic’s seven founders jointly hold stock giving a founder majority 50.1% of shareholder votes and three directors. Without founder consensus, those shares vote proportionally with common stock. For one specified LTBT director, the board’s nominating committee may reject a nomination or removal without cause and send the dispute to employees, formalizing power resembling staff influence after OpenAI’s board fired, then reinstated, CEO Sam Altman.
Founder divisions could amplify common shareholders, while employees could override the LTBT by siding with the board. Founders may surrender control 10 years after the IPO, and it ends automatically when fewer than three remain. The LTBT then survives until 85% of common shareholders abolish it, or it steps down with board approval. Common shareholders gain control only after both special classes disappear.
Positives
- Public benefit corporation status lets Anthropic’s board prioritize AI safety alongside profit without inviting shareholder lawsuits over mission-based decisions.
- Four Class T directors give the independent Long-Term Benefit Trust a board majority dedicated to Anthropic’s mission.
- Employee voting can check disputes between the board and LTBT over one specified trust-controlled seat.
- Founder control may end 10 years after the IPO and automatically ends when fewer than three founders remain.
- Common shareholders eventually gain full control once both founder and LTBT special stock classes disappear.
Risks & concerns
- Public market investors will have little influence over major Anthropic decisions, including most board appointments.
- Seven founders retain 50.1% of shareholder voting power when a majority agrees and can elect three directors.
- Existing LTBT members will select their own replacement for the trust’s current vacancy.
- An 85% common shareholder vote is required to abolish the LTBT after founder control ends.
- OpenAI’s Sam Altman episode shows how employee loyalty can overturn intended governance checks, making Anthropic’s similarly complex structure unpredictable.