Amodei: No open-weight AI ban, curb China chips
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic CEO Dario Amodei published a position statement saying he has "never" advocated a ban on open-weights AI models, addressing recent industry debate over Chinese open-weight releases
- Amodei listed reasons top chips should not be sold to China, aligning with broader US export-control efforts targeting AI compute
- Amodei called for global model testing as a safety mechanism rather than blanket bans on open-weight distribution
- Coverage across Bloomberg, CNBC, TechCrunch, Politico, The Verge, and The Register emphasized different facets — The Verge framed it as "Why China is giving away its best AI models," while The Register highlighted Nvidia CEO Jensen Huang pushing back on "open-weights fearmongering"
- The statement follows several days of industry discussion about open-weight models, especially Chinese ones, and their implications for US AI competitiveness
Why it matters: Amodei's position splits the difference: he formally rejects open-weight bans, which matters because Anthropic could have lobbied for restrictions, but pairs it with China chip controls that still constrain Chinese AI development through compute scarcity — a middle path that lets Anthropic look reasonable while supporting compute-side containment.
