Anthropic announced on July 21 that it will donate another $20 million to Public First Action, a political group pushing for government-imposed AI safeguards, doubling its total contributions to the group to $40 million as the 2026 midterms approach. Public First Action bills itself as bipartisan, focused on public education about AI and campaigning for transparency requirements, accountability measures and whistleblower protections for people who flag AI-related risks.
The donation is consistent with Anthropic's public regulatory posture: the company says it has "long argued that frontier AI companies should be transparent about what their models can do and how they're managing the risks," and has already supported several newly passed state laws requiring greater AI-developer transparency. Doubling down with real midterm-cycle money, rather than just supportive statements, signals Anthropic sees 2026 as a genuine inflection point for U.S. AI policy.
“Doubling down with real midterm-cycle money, rather than just supportive statements, signals Anthropic sees 2026 as a genuine inflection point for U.S.”
The spending puts Anthropic in direct political conflict with a rival effort: Public First Action has clashed with a super PAC backed by OpenAI president Greg Brockman, meaning two of the industry's most prominent labs are now funding opposing sides of the regulation debate with real campaign money rather than just competing lobbying memos. That's a meaningfully different phase of the AI policy fight than the informal Washington influence campaigns that have dominated the last two years.
The timing is notable on its own. The same week Anthropic is doubling its regulatory-advocacy spending, the White House is separately pursuing IP-theft sanctions against Moonshot AI over allegations it improperly distilled Anthropic's own Fable model -- meaning Anthropic is simultaneously playing defense on a stolen-model claim and offense on domestic AI policy in the same news cycle, a dual-front position most of its peers aren't navigating right now.
For founders and operators, the practical signal is that AI regulation is no longer a background risk to model in five-year plans -- it's an active midterm-cycle political fight with real money behind both sides, and outcomes could shape transparency and safety-testing requirements well before 2028. LPs should watch whether other frontier labs follow Anthropic and OpenAI into direct political spending, and whether Public First Action-backed candidates outperform in races where AI policy becomes a visible wedge issue this fall.
