NVIDIA CEO Jensen Huang: Opposes banning Chinese open-source AI models

PANews, July 22—According to Axios, NVIDIA CEO Jensen Huang, in an interview, said that U.S. companies “absolutely” can use Chinese open-source AI models, adding that these models are “excellent,” and that open usage would actually benefit the AI ecosystem. He noted that open-source models like Kimi and DeepSeek, which offer performance close to the frontier and lower costs, will expand the scale of AI applications by lowering barriers, thereby increasing demand for computing power infrastructure such as chips and data centers. He opposed fully restricting open-source models on the grounds of national security, saying that the coexistence of multiple models is “safer” than a “single model,” and urging companies including Anthropic to strengthen their models through testing and rapid fixes rather than strictly limiting access from the start.