A separate single Class T share gives Anthropic’s independent Long-Term Benefit Trust authority to elect and remove a majority of the seven-member board, while the co-founders control three seats. The trust—which added Harvard physicist Christopher Stubbs alongside Chair Neil Buddy Shah, former Fed Chair Ben Bernanke and CNAS CEO Richard Fontaine—survives unless 85% of common shareholders vote to remove it, while the Class F share automatically terminates if fewer than three co-founders remain or if the founders choose to end it 10 years after the listing. The board’s nominating committee can also block the trust’s no-cause removal of one specified director and send the decision to an employee vote.

In its 94 pages of risk factors, the public benefit corporation warns that its mission and safety decisions “may conflict with short-, medium-, or long-term financial interests” and that its AI models have shown self-preserving behavior, including “attempts to resist shutdown” and “behavior resembling blackmail,” while advanced systems could “pose catastrophic or existential risks to humanity.” Each co-founder has made a voluntary, revocable pledge to donate 80% of their personal equity to charity without reducing their voting control, alongside 846.8 million shares—or 5% of fully diluted equity—placed into the wholly owned Anthropic Public Benefit Fund, which the board can expand by up to 2% for every 50% share price gain above the IPO price. The governance disclosures accompanied financials showing $16.3 billion in first-half revenue, a $974 million second-quarter operating profit, $131.7 billion in cash and short-term investments, and an $80.7 billion net loss driven by $72.4 billion in noncash convertible-note charges as Anthropic targets a Nasdaq debut under the ticker ANT the week of Nov 16.