Anthropic launches AnthroPAC to influence US AI policy battles

Anthropic’s creation of AnthroPAC formalizes what has been obvious for a while: frontier labs are now political actors, not just research outfits. A company that markets itself on AI safety and constitutional training is setting up a compliant, employee‑funded vehicle to back lawmakers it sees as aligned with its vision for AI regulation and defense policy. This is standard practice in other industries, but new territory for a lab whose products could shape global power balances.([axios.com](https://www.axios.com/2026/04/03/anthropic-midterms-pac))

The timing is telling. Anthropic is simultaneously fighting a designation that paints it as a national‑security supply‑chain risk, pushing back on military use‑cases it views as incompatible with its safety commitments, and asking Washington to adopt strong guardrails on model misuse. AnthroPAC gives it a more direct lever over who writes those laws—and who oversees the Pentagon budgets it depends on for cloud and compute contracts.

For the race to AGI, this reinforces that trajectory is no longer set only in labs and data centers. If Anthropic and its peers successfully engineer a friendly regulatory perimeter—favouring high‑safety, high‑capital players—they could slow open‑source competitors and smaller labs, entrenching an oligopoly over the most capable systems while still accelerating frontier research inside that club.