China Rebukes Anthropic CEO’s Call to Restrict AI Chip Sales

China’s Foreign Ministry has rebuked Anthropic CEO Dario Amodei over his call to keep advanced AI chips and chipmaking equipment out of China, arguing that confrontation will undermine global AI governance. The response puts a government-level objection to a policy Amodei presents as necessary to preserve the U.S. lead in advanced AI.

Asked about Amodei’s position at a regular press conference, Foreign Ministry spokesperson Guo Jiakun said all parties should work together on AI. His remarks were a direct answer to Amodei’s argument that sales restrictions on cutting-edge AI chips and the equipment used to make them should continue.

The hardware dispute behind the diplomatic clash

Amodei’s position is not a call to ban Chinese open-weights models, which are AI systems whose model parameters can be accessed by others. In a July statement, he said Anthropic had never advocated such a ban and instead argued for limits on powerful chips and chipmaking equipment supplied to China.

His case rests on a competitive and security judgment. Amodei warned that a Chinese lead in AI would pose a grave danger to the United States and the world. He has also called for action against chip smuggling and other workarounds that provide access to restricted hardware.

Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance which serves no one’s interest.

Guo Jiakun, spokesperson for China’s Ministry of Foreign Affairs

Two incompatible descriptions of the same policy

For Amodei, chip controls are a targeted way to prevent an authoritarian government from building AI models more powerful than those developed in the United States. His July statement argued that China’s domestic production capacity is limited and that access to U.S. chips would therefore be pivotal to training more capable systems. That is Anthropic’s assessment, not an independently established outcome in this exchange.

China’s response casts the same approach as an obstacle to collaboration. State-run Global Times went further, calling Amodei’s essay a Cold War-style effort to contain China’s AI sector. The gulf is not over whether AI needs governance in the abstract; it is over whether restricting a rival’s access to advanced computing can sit alongside a cooperative governance agenda.

What Amodei has proposed instead

• Do not sell powerful chips or chipmaking equipment to China.
• Crack down on smuggling and workarounds used to obtain those chips.
• Require safety testing for sufficiently capable models, whether they are open or closed.

Cooperation remains the unresolved claim

Amodei’s own position contains a qualification that makes Beijing’s rebuttal especially consequential. He wrote that effective safety testing would need to be global, meaning China would need to participate. He has argued that limited cooperation may be possible where preventing biological weapons is in China’s interest, even while advocating restrictions intended to constrain China’s frontier AI development.

That leaves a practical policy question with no answer in the exchange: what cooperation on AI safety could look like when one side describes chip restrictions as a necessary security measure and the other describes confrontation as a threat to governance. China’s Foreign Ministry has supplied its answer in principle—cooperation rather than rivalry—without identifying a shared framework or terms that would bridge the hardware dispute.