Anthropic Puts 5 Percent of Itself — Roughly $48 Billion — Into a Public Benefit Fund That Grows With the Stock

Anthropic has pledged 5 percent of its own equity — currently worth about $48 billion — to a new public benefit fund intended to address the societal fallout of AI, in a structure that will keep growing as the company gets richer. The commitment was disclosed in the Form S-1 the Claude maker filed with the SEC on October 26 ahead of a planned Nasdaq listing under the ticker ANT, and reported in detail by the Financial Times.

The mechanics are now public. According to the filing, the company has contributed 846,840,960 Class N shares to the Anthropic Public Benefit Fund, a wholly owned subsidiary — equal to 5 percent of fully diluted shares as of June 30, 2026, assuming full conversion of Amazon's notes. The board has also approved a framework under which it may add up to 2 percent more shares at each milestone defined by a 50 percent rise in the share price over its previous level, subject to trading-day tests and board discretion. In practice, that means every sustained leg up in the stock dilutes ordinary shareholders by another two percentage points in favor of the fund — a structure that ties the scale of Anthropic's public-interest giving directly to its market performance.

With a potential IPO that press reports have valued at roughly $2 trillion, the new pledge would bring the proportion of Anthropic stock committed to public causes to around 20 percent, per the FT. The fund's mandate is broad: the company said in an October 18 announcement that it could make large charitable donations or even direct payments to people who lose their jobs to AI, alongside initiatives in AI safety and biomedical research. No timetable, recipients or governance details have been published; Anthropic says those will follow in the coming months.

The corporate pledge sits alongside an unusually large personal commitment. Dario Amodei and six co-founders have pledged 80 percent of their personal equity to philanthropy — though the S-1 is explicit that those promises are voluntary, revocable, and create no legal duty to give, with no donation deadline or named recipients. The filing also describes a governance tension the SEC will scrutinize: a founder's economic stake can shrink after a donation while the founder's voting control, held through a Class F share giving seven founders 50.1 percent of votes, stays intact. Amodei himself argued in a January essay that private giving alone cannot address the economic disruption he expects AI to cause, and that government action will also be needed.

The move lands in a heated context. Anthropic is headed into a public listing with $16.3 billion of first-half revenue, a headline net loss of $80.7 billion — most of it $72.4 billion in noncash fair-value changes — and an operating loss of $693 million. Public frustration with AI companies is growing, and the fund gives Anthropic a visible answer: rather than waiting for policymakers, it is pre-committing a slice of the equity itself. Whether shareholders read the milestone top-up mechanism as enlightened design or a standing 2 percent overhang will be one of the more interesting questions in the book-building process when pricing begins.

Either way, the structure sets a marker for the industry. No major AI lab has previously hardwired a fixed percentage of its equity — with automatic escalators — into a societal fund before going public. If the Anthropic offering prices anywhere near the numbers circulating, the five percent it has carved out will be among the largest single commitments to AI-impact funding ever made, and every other lab weighing an IPO will now be asked why theirs is zero.