Anthropic CEO Dario Amodei published a statement Tuesday rebutting what he called a recent uptick in inaccurate claims about the company’s policy stances, answering White House AI and crypto czar David Sacks’ charge that Anthropic runs a “sophisticated regulatory capture strategy based on fear-mongering.”

Amodei stressed areas of agreement with the Trump administration, including a two-year, $200 million agreement awarded in July by the Department of War to prototype frontier AI capabilities for national security, and partnerships that put Claude in classified networks. He reiterated Anthropic’s longstanding preference for a uniform federal AI standard over a patchwork of state laws, noting the company respectfully disagreed with the 10-year state-law moratorium proposed in the One Big Beautiful Bill—a provision that failed 99-1 in the Senate.

He also cited progress on model bias, pointing to a January Manhattan Institute study and May Stanford data showing Anthropic’s models were not outliers, and to recent system cards. While advocating for a federal standard, Anthropic supported California’s SB 53, which requires large developers to publish safety protocols and exempts companies under $500 million in revenue.

The exchange follows a week of public swipes between Sacks and Anthropic, with Reid Hoffman defending the lab on X and Sacks firing back.