Anthropic CEO Attacks NVIDIA: Exporting AI Chips to China is a Grave Mistake

SadaNews - The CEO of Anthropic, Dario Amodei, sparked widespread controversy during his participation in the World Economic Forum in Davos, sharply criticizing the U.S. administration's decision to allow the export of advanced artificial intelligence chips from NVIDIA and AMD to approved Chinese clients.

Last week, the United States granted approval to sell Nvidia H200 chips, along with a production line from AMD, to China, reversing a previous ban.

Although these chips are not the latest or most powerful in the companies' arsenal, they are classified as high-performance processors used in artificial intelligence applications, which has made the decision a subject of extensive debate, according to a report published by TechCrunch that was reviewed by Al Arabiya Business.

During a panel session in Davos, Amodei directed direct criticisms at the U.S. administration and chip manufacturers, warning of the implications of the decision on U.S. national security.

He stated that the executives at these companies justify the move by claiming that export restrictions on chips hinder their business, but he described this reasoning as "short-sighted."

Amodei added: "The United States is years ahead of China in chip manufacturing, so exporting these processors would be a huge mistake.”

In an interview with the editor-in-chief of Bloomberg, he confirmed that AI models have profound implications for national security, describing them as representing "real intelligence or human-like understanding."

Amodei compared the future of artificial intelligence to the existence of a nation of geniuses within a data center, urging people to imagine "100 million minds smarter than any Nobel laureate," all under the control of one country.

However, the most provocative statement came when Amodei declared that the decision to export chips was "insane," likening the move to "selling nuclear weapons to North Korea while boasting that an American company made their covers."

These remarks carry increased sensitivity, given that NVIDIA is a key partner and major investor in Anthropic, as the company relies entirely on the graphics processing units produced by NVIDIA to run its AI models via major cloud computing platforms like Amazon, Microsoft, and Google.

The two companies had just announced a deep technical partnership two months ago, along with an investment from NVIDIA in Anthropic that could reach $10 billion, with mutual promises to improve and integrate technologies between the two parties.

However, the remarks in Davos put this relationship under scrutiny, as Amodei appeared to position his partner in the role of an arms dealer.

Observers believe that Amodei's statements may stem from a genuine concern over the rise of Chinese AI labs and a desire to prompt Washington to tighten its policies.

Conversely, this tone reflects the extent of the shift in how leaders of AI companies view global competition, which is now described as an existential struggle.

Despite the severity of the statements, Amodei does not seem to be worried about their implications for his company's business, given the strong financial position that Anthropic enjoys, having raised billions of dollars and estimated to be worth hundreds of billions, while its software assistant Claude is gaining increasing popularity among developers.

Analysts conclude that what happened in Davos reflects a new phase in the AI race, where diplomatic considerations and partnership relations are overshadowed by concerns over national security and technological superiority.