Anthropic’s Planned $2 Trillion IPO Tests External Trustee Control
Anthropic's planned IPO could value it at $2 trillion, testing whether its external trust can balance AI safety, investor returns and control of its board.
Summary
As of September 4, 2026, Anthropic plans to retain its Long Term Benefit Trust after an IPO that could value the Claude maker at $2 trillion. The trust owns no equity but can appoint or dismiss a board majority and chose 4 of 7 directors, including Netflix cofounder Reed Hastings and Novartis CEO Vas Narasimhan. Its 3 members from a possible 5 are chair Neil Buddy Shah, CEO of the Clinton Health Access Initiative, former Federal Reserve chair Ben Bernanke, and Center for a New American Security CEO Richard Fontaine. Former California Supreme Court justice Mariano Florentino “Tino” Cuéllar left after several months to become Anthropic’s chief global affairs officer.
Trustees receive advance notice of major actions, including model launches, meet weekly and with leadership as often as every other week, attend board meetings and consult founders. They urged a limited Mythos cybersecurity model rollout through the Glasswing Project and discussed Anthropic’s automated weapons dispute with the US government, but have remained advisers without forcing tradeoffs between profit and purpose. Anthropic wants the trust to become an AI governance blueprint resembling GAAP standards; Bernanke’s July appointment supports that institutional ambition.
Harvard professor Jesse Fried calls the structure a built in conflict between profit seeking investors and mission guardians, while University of Pennsylvania professor Elizabeth Pollman says geopolitical and corporate AI competition makes competing interests difficult to contract for. The lossmaking company has not stress tested trustee power while pursuing sustainable profits. OpenAI’s nonprofit rooted model showed the danger in November 2023, when its board failed to oust Sam Altman, lost investor and employee confidence, and was mostly replaced as broader restructuring followed. Anthropic trustees can be removed by 85 percent of shareholder voting power, though that threshold could change after listing. Private backers understood the structure and some cited safety, while one venture investor expected commercial necessity to prevail because compute and leading models require capital.
Positives
- 4 of Anthropic’s 7 directors were selected by trustees charged with protecting the company’s long term public benefit mission.
- Trustees receive advance notice of major actions, meet weekly and can consult Anthropic’s leadership as often as every other week.
- The Glasswing Project limited the rollout of Anthropic’s Mythos cybersecurity model after trustee involvement.
- Private investors backed Anthropic with full knowledge of its structure, and several cited its safety emphasis as an investment rationale.
- An 85 percent shareholder vote can remove the trustees, providing an ultimate safeguard against unaccountable control.
Risks & concerns
- Anthropic’s trust has imposed no red lines or major profit versus purpose tradeoffs, leaving its controlling power untested.
- The lossmaking company must satisfy a broader public investor base while funding costly compute and competition for leading AI models.
- Jesse Fried identifies a built in conflict between profit seeking shareholders and mission guardians with little financial exposure.
- OpenAI’s failed November 2023 attempt to remove Sam Altman shows how unconventional AI governance can collapse under investor and employee pressure.
- The 85 percent threshold for removing Anthropic’s trustees could change when the company enters public markets.