Nvidia CEO Jensen Huang downplayed our fears that his company’s chips would help the Chinese military, a few days before his day trying to walk the rope between Washington and Beijing.
exist interview With CNN airing on Sunday, Huang said: “We don’t have to worry about “China’s military using American manufacturing technology because “they can’t rely on it at all.”
“This may be limited at any time; not to mention that China already has a lot of computing power,” Huang said. “Of course, they don’t need Nvidia’s chips, nor do they need the U.S. technology stack to build the army.”
The comments are index years of bipartisan policy in the United States that imposes restrictions on semiconductor companies, forbidding them from selling their state-of-the-art AI chips to Chinese customers.
Huang also repeated past criticisms of policy, believing that the strategy of export control is contrary to the ultimate goal of U.S. technical leadership.
“We want the US technology stack to be the global standard … In order for us to do this, we have to look for all the AI developers in the world,” Huang said.
He added that this means that making the United States a leader in AI, all markets, including China, must use U.S. technology.
Washington’s Latest restrictions Regarding NVIDIA’s sales to China, implemented in April, it is expected to cause billions of dollars in losses for the company. Huang said chip restrictions in May have cut Nvidia’s China Almost half of the market share.
Huang’s CNN interview came just days before he traveled to China for his second trip to the country this year, and Nvidia reportedly is working on another Chip that meets the latest export controls.
Last week, NVIDIA CEO Meet with US President Donald TrumpAnd warned by US lawmakers No meeting Companies related to Chinese military or intelligence agencies or entities designated on the U.S. list of restricted exports.
According to Daniel Newman, CEO of Futurum Group technology consulting firm, Huang’s CNN interview reflects how Huang threaded between Washington and Beijing while maintaining maximum market access.
“He needs to walk a well-known rope to make sure he doesn’t rattle,” Newman said. He also wants to provide a better climate in policies and hopes that China can invest in NVIDIA technology.
But that’s not to say that his focus on Washington is effective. “I think it’s hard to fully accept the idea that China cannot use NVIDIA’s most advanced technology for military purposes.”
He added that he hopes NVIDIA’s technology will be the core of AI training in any country, including for the development of advanced weapons.
US officials Tell Reuters last month China’s large language model launch DeepSeek (which means it uses NVIDIA chips to train its models) is supporting China’s military and intelligence operations.
Huang admitted on Sunday that DeepSeek’s open source R1 reasoning model was trained in China, but there is no evidence that it poses a danger for this reason alone.
Huang praised the R1 reasoning model, calling it “revolutionary” and said its open source nature gives startups, new industries and countries the ability to be able to participate in AI.
“The fact is (China and the United States) are competitors, but we are highly interdependent and it’s nice to respect our competitors to the extent that we can compete and desire to win,” he concluded.
