Jensen Huang, the CEO of the company that makes the chips powering most of the world’s AI infrastructure, just told everyone something they probably didn’t want to hear: China is winning the AI talent war, and it’s not particularly close.
Speaking at a CSIS event, Huang claimed that approximately 50% of the world’s AI researchers are Chinese. Not 50% of researchers in China, mind you. 50% of all AI researchers, globally.
The numbers behind the claim
Huang’s assertion extends beyond just headcount. According to his remarks, China also accounts for roughly 70% of recent AI patent publications. In English: for every ten new AI patents filed around the world, seven of them are coming out of China.
Huang described China’s AI ecosystem as “vibrant, rich, and incredibly innovative.”
In a July 22, 2026 interview with Axios, Huang went further, asserting that US companies should “absolutely” utilize Chinese open-source AI models. That’s a notable position for the head of a company that sits right at the center of US-China tech tensions. Nvidia has had to navigate a minefield of export restrictions, licensing requirements, and political pressure over chip sales to China.
Why talent matters more than chips
Huang’s comments also implicitly highlight a workforce development gap in the US. His advocacy for reskilling American workers and investing in domestic talent pipelines suggests that the current output of US universities and training programs isn’t keeping pace with demand.
What this means for investors
The open-source dimension is particularly interesting. Chinese AI labs have been releasing competitive open-source models, and Huang’s explicit endorsement of using those models could accelerate adoption among US companies that were previously hesitant for political or reputational reasons.
Huang has also pushed for increased dialogue between US and Chinese researchers on safe AI practices. If the two largest AI ecosystems develop incompatible safety frameworks, it creates fragmentation risk for companies operating globally.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.

1 hour ago
21









English (US) ·