Artificial intelligence
Anthropic's Seven Founders Want 50.1% of the Vote for About 2% of the Company. The Trust Already Holds the Board
Anthropic will ask shareholders within days to approve a super-voting share class handing its seven co-founders a bare majority of votes ahead of a listing that could be one of the largest ever. The unusual part is not the founder control — it is that the Long-Term Benefit Trust already elects most of the board.
MAI
Anthropic will ask its shareholders in the coming days to approve a new share class that would give its seven co-founders 50.1% of the votes on most corporate matters, according to The Information, whose report Reuters and TechCrunch both carried on September 25. The founders hold roughly 2% of the company each. The new shares carry votes and no economics — no dividends, no additional claim on proceeds. The arrangement holds as long as at least three of the seven keep a minimum stake.
Framed that way it is the ordinary pre-IPO founder-control move, the one Google, Meta and Palantir all made on their way to the market. What makes it worth reading closely is where it lands: on top of a governance structure that had already taken board control away from shareholders.
Three power centres, not two
Anthropic created the Long-Term Benefit Trust during its Series C round and gave it a dedicated share class, Class T. The Trust's five trustees are not shareholders and are appointed for their background in AI safety, national security and public policy; the current membership includes former Federal Reserve chair Ben Bernanke, Richard Fontaine of the Center for a New American Security, Mariano-Florentino Cuéllar of the Carnegie Endowment, and Neil Buddy Shah as chair. Anthropic's own description of the Trust says it has authority to select and remove a portion of the board that grows over time, "ultimately, a majority of our Board."
That majority does not go away if this proposal passes. According to Reuters' account of the reporting, the Trust still selects most directors, and the founders' new voting power does not reach board elections. The board has seven seats, one of them currently vacant, and the founders' allocation of them rises from two to three. Employees are to get a separate class of stock acting as a tie-breaker on some questions.
| Who | What they would control |
|---|---|
| Long-Term Benefit Trust (Class T) | Selection and removal of most of the board |
| The seven co-founders | 50.1% of stockholder votes on most matters; three of seven board seats |
| Employees (new special class) | Tie-breaking votes on some corporate questions |
| Public shareholders after the IPO | Economic ownership; whatever votes are left |
The number is 50.1% for a reason
The closest precedent is Palantir, whose founders' voting trust is capped at 49.999999% — a figure engineered to sit just below a majority. Anthropic's proposal crosses that line, and it does so collectively: seven people acting as a bloc rather than one founder with a super-voting stake, which is unusual at this scale.
Crossing 50% is not cosmetic. Under the listing standards of both major U.S. exchanges, a company where more than half the voting power sits with one person, group or entity is a "controlled company" and may opt out of rules requiring a majority-independent board and independent nominating and compensation committees. In Anthropic's case the practical effect is narrower than it looks, because the Trust — not the voting majority — picks most of the directors anyway. What the founders' 50.1% actually secures is the set of decisions that reach stockholders directly: charter and bylaw amendments, mergers and sales of the company, and the kind of shareholder proposal or proxy campaign that an activist investor would bring after a listing.
What the buyers are being offered
Anthropic confidentially submitted a draft S-1 to the SEC on June 1. It raised $65 billion in May at a $965 billion post-money valuation, and TechCrunch reports secondary-market pricing since then around $1.5 trillion. The listing was expected this autumn; both Reuters and the Chinese outlet 36Kr now describe it slipping to late October or November, after the U.S. midterm elections.
So the pitch to public investors in what would be one of the largest offerings ever run is economic exposure with very little governance. The case for it is the same case Anthropic made for the Trust: a company that believes its product carries civilisational risk should not be steerable by quarterly earnings pressure, and insulation is the mechanism. Dario Amodei has previously said he intends to give away 80% of his wealth, citing the destabilising effect of AI-driven wealth concentration — a signal that the control being sought is not primarily financial.
The objection is structural rather than about motive. Two overlapping insulation mechanisms leave the company accountable to almost no one outside it. It is also worth noticing that the founders' bloc and the Trust are checks on each other as much as on shareholders: a Trust-selected board majority and a founder voting majority can disagree, and nothing published so far explains how that deadlock resolves beyond the employees' tie-breaker shares. For a company whose central argument is that its governance is designed for a hard problem, that is the part investors will want written down.
None of this has come from Anthropic. The company has not published the proposal, and the details rest on The Information's reporting as relayed by Reuters and others. The shareholder vote, if the timing holds, should settle them within days.
Sources: Anthropic seeks 50.1% voting control for co-founders ahead of IPO, The Information reports (Reuters, via Business Standard) · Anthropic's founders seek voting control ahead of IPO (TechCrunch) · Anthropic Founders' IPO Ownership & Voting Rights Structure Explained (36Kr) · The Long-Term Benefit Trust (Anthropic) · Anthropic confidentially submits draft S-1 (Anthropic) · Anthropic confidentially files IPO prospectus with SEC (CNBC)