Nvidia and Anthropic clash over US chip export rules due to take effect 15 May

Nvidia and Anthropic sit on opposite sides of US artificial intelligence policy. Chip export restrictions are due to take effect on 15 May, CNBC reported on 1 May. The two companies disagree over how tightly the United States should control the sale of advanced AI chips abroad.

Anthropic, the AI startup backed by billions from Amazon, argued for tighter controls and enforcement in a blog post on Wednesday. The company said Chinese smuggling tactics involved chips hidden in "prosthetic baby bumps" and "packed alongside live lobsters." It cited the 2022 arrest of a woman carrying chips into China and the 2023 seizure of "computer display cards" with a lobster shipment in Hong Kong. Nvidia's response was blunt.

"American firms should focus on innovation and rise to the challenge, rather than tell tall tales that large, heavy, and sensitive electronics are somehow smuggled in 'baby bumps' or 'alongside live lobsters,'" an Nvidia spokesperson said.

The restrictions in question date from former President Joe Biden's term and are called the "AI Diffusion Rule." The rule puts global export controls on advanced AI chips and model weights. The aim is to stop rival nations like China from gaining ground in an escalating AI arms race. President Donald Trump is reportedly working on updating these restrictions, which adds another layer of uncertainty to an already contentious policy.

Anthropic relies heavily on Nvidia hardware to train its models, yet it is calling for tighter restrictions that could limit Nvidia's overseas business and revenue from chip sales. The company argued that compute access is the key strategic chokepoint in the race to build frontier AI. It proposed lowering the export threshold for Tier 2 countries, tightening the rules to reduce smuggling risks, and increasing funding for enforcement.

"Maintaining America's compute advantage through export controls is essential for national security and economic prosperity," Anthropic wrote.

Nvidia pushed back on the use of policy to limit competitiveness. "China, with half of the world's AI researchers, has highly capable AI experts at every layer of the AI stack. America cannot manipulate regulators to capture victory in AI," a spokesperson said. Anthropic stood by its position. "Anthropic stands by its recently filed public submission in support of strong and balanced export controls that help secure America's lead in infrastructure development and ensure that the values of freedom and democracy shape the future of AI," a company spokesperson said in a statement.

Nvidia CEO Jensen Huang met with Chinese trade officials in mid-April. On Wednesday in Washington, D.C., he said China is "not behind" the U.S. in AI and praised Huawei as a top global tech company.

"They're incredible in computing and network technology, all these essential capabilities to advance AI," Huang said. "They have made enormous progress in the last several years."

Huang has since gone further. At the Computex tech forum in Taipei on Wednesday, he called US chip export controls a "failure" because they spur Chinese development, The Guardian reported on 21 May. "The local companies are very, very talented and very determined, and the export control gave them the spirit, the energy and the government support to accelerate their development," Huang said.

He said Nvidia had written off "billions of dollars" in sales, and its share of China's AI chip market had fallen from almost 95% at the start of the Biden administration to 50%. "China has a vibrant technology ecosystem, and it's very important to realise that China has 50% of the world's AI researchers, and China is incredibly good at software," he said.

The controls have also pushed Chinese firms toward their own silicon. DeepSeek, the Chinese startup developing large language models, is planning to enter the silicon business, according to Reuters, which cited three people familiar with the matter. Ars Technica reported on 7 July that DeepSeek has been working on the move for about a year, meeting potential partners and hiring engineers. The focus is on data center chips for inference, not training, and the goal is likely to reduce reliance on both Huawei and Nvidia.