Nvidia discloses more China risks, but CEO Jensen Huang praises Trump

Nvidia for the first time said that restrictions on the use of open-source AI models from China such as DeepSeek and Qwen could hurt its business.

Even as Nvidia reported another blockbuster quarter of 69% sales growth on Wednesday, the maker of artificial intelligence chips warned of more risks to its business emerging in the technology conflict between the U.S. and China.

Tucked into Nvidia's quarterly filing with U.S. securities regulators, Nvidia for the first time said that restrictions on the use of open-source AI models from China such as DeepSeek and Qwen could hurt its business, as could U.S. rules barring connected vehicle technology from China, where Nvidia's long-struggling car chip business has finally flourished.

While Nvidia CEO Jensen Huang on a conference call with analysts praised U.S. President Donald Trump's decision to rescind an export rule put in place by President Joe Biden that would have regulated the flow of Nvidia's chips around the world, the company's quarterly filing noted that no new rule had been issued in its place and that a "replacement rule may impose new restrictions on our products or operations."

On the other hand, Huang criticised new export curbs imposed by the Trump administration in April. The curbs stop the company from selling its H20 chip made for the Chinese market, which Huang called "a springboard to global success."

The export limits cost Nvidia $2.5 billion in sales during its just-ended fiscal first quarter, and it expects another $8 billion sales hit during the current fiscal second quarter. Sales of the H20 in China earned Nvidia $4.6 billion in revenue as customers stockpiled the chips before the curbs set in. The China business accounted for 12.5% of overall revenue.

"The question is not whether China will have AI; it already does. The question is whether one of the world's largest AI markets will run on American platforms," Huang said, later adding that "AI export controls should strengthen U.S. platforms, not drive half of the world's AI talent to rivals."

Huang also argued that keeping Chinese open-source models such as DeepSeek and Qwen running on Nvidia chips provides U.S. firms with valuable insight on where the global AI industry is headed.