Anthropic announces $40 million for AI-policy advocacy

A company funds the argument for stronger rules

Anthropic’s opening move was a $20 million contribution to Public First Action in February 2026. Its stated priorities included transparency from AI labs, safeguards for children, and American technological leadership. February announcement

The recipient’s own platform backs state and local AI legislation and resists federal efforts to freeze that work without sufficient national safeguards. That puts it directly into the fight over where AI rules should be written—and how demanding they should be.

The funding doubled, with a clear restriction

On July 21, Anthropic announced another $20 million, bringing its stated total to $40 million. It said both contributions were for policy and public education, not candidate-election activity. July announcement

The distinction matters. This is documented corporate funding intended to elevate a policy argument; it should not be presented as a $40 million election-ad budget. The company’s announcements reveal its commitment and stated conditions, not a complete accounting of the recipient’s spending.

What stronger rules would actually mean

Anthropic’s separately published Advanced AI Framework goes beyond asking companies to disclose their practices. It proposes independent evaluation, civil penalties and government authority to block or deter deployments that pose significant catastrophic risk. Its proposed scope is the most capable models from large developers, not every AI application.

That is the concrete policy stake: who can decide when a frontier model is too risky to release. Anthropic presents the authority as a safety measure with protections against overreach. This is the donor’s proposal, not proof that Public First Action has funded every element of it—or that the proposal became law. Follow the donation and the recipient’s stated agenda.