Amodei: Anthropic never advocated open-weight AI ban

Anthropic CEO Dario Amodei on Monday pushed back against accusations that his lab wants to ban open-weight AI models, saying in a blog post that his company has “never advocated” for such a ban and calling models without dangerous capabilities “a public good.” The statement comes as Silicon Valley is deeply divided over how to handle the rise of powerful open-weight models, especially from China.

Industry pressure mounts

The debate intensified after the debut of Kimi K3, a Chinese open-weight model that approaches U.S. frontier performance at a fraction of the cost, rattling the industry. In response, Nvidia, Microsoft, Meta, Google, OpenAI and dozens of other companies signed a letter urging Washington not to restrict open-weight AI. Anthropic was notably absent from that letter, fueling accusations that the company was using safety concerns to shield its closed-model business from competition. White House AI adviser David Sacks was among those critical of Anthropic’s position.

Amodei's narrow stance

Amodei acknowledged that a ban on Chinese open models would shield U.S. labs like Anthropic from competition but insisted “that has never been my goal.” He argued that such a ban would do little to stop bad actors, who “are unlikely to be legitimate US businesses.” Instead, he proposed three narrower policies: tighter controls on advanced chips and chipmaking equipment flowing to authoritarian governments, a crackdown on industrial-scale distillation (training models on outputs of more advanced systems), and mandatory safety testing for sufficiently capable models, whether open or closed. He said these measures address the chokepoints that determine how powerful AI models can become.

Criticism and competition

Anthropic faced criticism after declining to join the open-weight letter. Google and OpenAI, its two biggest closed-model rivals, signed on over the weekend after originally abstaining, leaving Anthropic as the only major holdout. Amodei’s blog post was seen as an attempt to clarify his position and draw a line between opposing a blanket ban and supporting targeted controls on the most powerful AI systems. He agreed that open models can expand access and competition but rejected the claim that they necessarily make AI safer or give defenders an advantage.

The bottom line

Amodei does not want Washington to ban open models. He wants Washington to regulate the chokepoints—chips, distillation, and safety testing—that govern how powerful those models can become. The Trump administration has weighed actions including sanctions against Chinese labs accused of stealing U.S. AI research through distillation. Amodei’s proposals offer a middle path that could shape the regulatory debate in the months ahead.