Anthropic Says It Will Give 5% of Its Shares to the Public Good
Anthropic said on October 18, 2026 that it will commit 5% of its shares to a new public benefit fund. Its board may add another 2% later, if it chooses to. The company's news post, titled "Investing in the public benefit," is the main statement of the plan.
CNBC reported that the fund is meant to address problems such as AI-driven job loss. Al Jazeera reported on October 19 that the company pledged 5% and plans to increase that amount. On October 20, Reuters reported that the fund plans a record $9 billion payout in 2026. About 70% of its budget is aimed at its first two goals. These are plans, not finished results.
Anthropic is a Public Benefit Corporation. That is a type of company that must weigh what its owners earn against what the wider public gains. Its Public Benefit Charter says the company exists to develop and maintain advanced AI responsibly, for the long-term good of humanity.
The charter lists five goals. They are: AI safety problems get solved; AI safety becomes a universal practice that the law backs up; AI strengthens democracy and global stability; AI helps people become more prosperous and flourish; and everyone shares in AI's benefits. Co-Founder and President Daniela Amodei is quoted on the page, and a 2026 Public Benefit Report is linked from it. Anthropic's Claude's Constitution page describes the company in similar terms.
A separate group, called the Long-Term Benefit Trust, sits above the company's board. Anthropic describes it as five independent trustees who can choose and remove board members. Their job is to balance the interests of shareholders and the public. You can read the company's original description of the trust. Anthropic also published a post titled "Ben Bernanke joins the Long-Term Benefit Trust," dated July 9, 2026. Earlier in the year, it published a post introducing The Anthropic Institute, dated March 11, 2026.
The big numbers describe two different kinds of promise. A share commitment is a promise of ownership, and what it is worth depends on when and how those shares are sold. The $9 billion is a reported spending plan for 2026. The 2% extra depends on a board decision that has not happened yet.
In my view, the best test of this fund is how it works in practice, not the percentage. The focus on AI-driven job loss is specific enough to check. Do the first two goals have measurable targets? How are grants or programs picked? Do the payouts arrive on time? Those answers will say more than the pledge itself. Those are open questions.
A more hopeful reading is also available. A major AI company has written a public benefit goal into its legal structure, set up a trust with power over its board, and set aside shares for the effects of its own technology on workers. Others can now study that setup, copy parts of it, or criticize it. If the fund's results can be measured against the five goals, this could become a model for how AI companies handle the costs their products place on workers. If the money is hard to trace, skeptics will have a stronger case. The next report, including Anthropic's 2026 Public Benefit Report, will show which path it is on.
