Anthropic states its position on open-weights models: no ban, but chip controls, distillation limits and mandatory safety testing

In a 27 July 2026 post Anthropic set out its position on open-weights models, with Dario Amodei stating the company "has never advocated for a ban on open-weights models." It names two concerns — authoritarian states fielding more capable AI than the US, which it argues is independent of whether weights are open, and misuse for cyber or biological harm, which it says is harder to monitor in open models — and proposes three measures: restricting advanced chip and chipmaking-equipment sales to China, targeted legal action against industrial-scale distillation backed by state actors, and mandatory safety testing for every sufficiently capable model, open or closed.

What it means If mandatory pre-release safety testing becomes the settled ask from a frontier lab, it lands on everyone who publishes weights — including teams shipping a fine-tune of somebody else’s model.