China’s Foreign Ministry rejected Anthropic CEO Dario Amodei’s argument that the United States should preserve advanced-chip controls to prevent a Chinese AI lead. The exchange exposed competing security claims while both sides continued to describe AI governance as a global concern. Beijing called the warning fearmongering, while Amodei framed export limits as part of a broader proposal to slow frontier development.
Amodei published an essay urging the United States to maintain restrictions on advanced AI chips and chipmaking equipment sold to China. He argued that a Chinese lead in frontier AI could create security risks for the United States and other countries. A Chinese Foreign Ministry spokesperson rejected that framing as fearmongering and confrontation. The spokesperson said cooperative global AI governance would better serve all countries.
China’s state-security chief separately warned that hostile actors could use AI against political and institutional security. The exchange occurred alongside calls from some U.S. technology leaders for a broader slowdown in frontier-model development. Export controls already restrict access to some of the most advanced chips and manufacturing equipment. Neither statement established how quickly capabilities would change under a particular control regime.
Advanced-model training depends on chips, energy, data, engineering talent and access to manufacturing supply chains. Export controls can slow access while also encouraging domestic substitutes and alternative supply networks. Corporate executives and government spokespeople are interested participants in the policy debate. Claims about future strategic dominance are forecasts and should be separated from observed technical capabilities.
The record remains incomplete in defined ways. The public exchange did not provide independently testable capability forecasts. The next observable steps are Changes to U.S. export-control rules. Any bilateral or multilateral agreement on frontier-model testing or chip transfers.