Anthropic earmarks about 20% of its shares for public benefit

What the company already does

Much of Anthropic’s work on its mission stays inside the company. That includes AI safety and interpretability research, applied biomedical research, and discounts on Claude for nonprofits and research institutions. Those efforts can serve the public while also carrying commercial value, Anthropic says, so it will keep them within the company.

The new fund is intended to give Anthropic more flexibility to pursue its mission in other ways. It could make large donations to established nonprofits or provide direct support to groups of people.

One possible focus is the effect of AI on workers. Anthropic has already committed $200 million to studying labor-market risks and looking for solutions. The company says the pace of AI development may call for larger interventions, which the fund could support.

A different route for the shares

Anthropic is structuring the fund as a wholly owned LLC, rather than as a 501(c)(3) or another type of organization. The company says this will let it choose among different forms of charitable support. The LLC’s Anthropic shares will not be used for lobbying or political donations, in the United States or elsewhere.

The commitment builds on earlier promises. Anthropic’s co-founders pledged to donate 80% of the shares they held when they started the company. Anthropic also has a program through which employees can donate shares to charity, with the company matching their contributions. About half of employees participate.

What the percentage leaves open

The roughly 20% figure is substantial, but the announcement leaves key details for later. Anthropic has not yet disclosed the valuation milestones that could increase the share, or how the fund will be governed. It says it will provide more information about the fund’s structure and governance in the coming months, alongside a report on its mission work.

I think the most important test will be whether the fund adds capacity for public-benefit work that Anthropic would not otherwise pursue. The company’s existing programs often overlap with its commercial interests; a fund that can give directly or make large outside grants could widen that scope. But until Anthropic explains who will make allocation decisions and what triggers a larger commitment, the 20% figure describes an intention more clearly than an operating plan.

Anthropic has also recently given METR access to its systems comparable to that of employees for safety evaluations, and begun preliminary talks with several other organizations, including Hugging Face. It has separately described steps to strengthen protections ahead of the 2026 U.S. midterm elections, and opened an office in Singapore while signing a memorandum of understanding with the country’s government on safe and responsible AI adoption.

The fund could turn company growth into a larger public-benefit commitment. Whether that commitment becomes more than a shareholding pledge will depend on the governance and milestones Anthropic has yet to publish.