U.S. Treasury Secretary Scott Besant warned that if the United States loses the artificial intelligence (AI) competition with China, it could have serious consequences for national security and the economy. His comments come amid growing concerns in the U.S. over the rapid emergence of China's open-weight AI models.
According to a Bloomberg report on September 9, Besant stated at a Breitbart News event in Washington the previous day, "If China wins the AI competition, there will be no tomorrow," adding, "If China gets ahead of us in AI, nothing else will matter." He also noted that even with significant defense spending, the U.S. would fail to protect itself if it falls behind China in the AI race.
Besant's remarks appear to reflect recent opposition in the U.S. regarding the construction of AI data centers. Former President Donald Trump has also warned that domestic backlash against data center construction could threaten the economy and ultimately risk ceding AI leadership to China.
This concern is contrasted with China's rapid development of AI data centers in sparsely populated areas in the north and northwest, far from major cities.
The growth of Chinese open-weight AI models is further fueling U.S. anxiety. Unlike the closed models led by U.S. companies such as OpenAI and Anthropic, open-weight models allow users and developers to download and modify the model's weights. Chinese firms are increasingly releasing competitively priced models that approach the performance of U.S. models.
According to the AI-focused newsletter Interconnects, downloads of Chinese open-weight AI models exceeded 328 million last month, more than double that of U.S. models. This surge is attributed to the release of powerful new products like Moonshot AI's 'Kimi K3' and DeepSeek's updated 'V4 Pro.'
The U.S. government has also raised suspicions that Chinese AI companies are engaging in large-scale distillation activities to acquire technology from leading U.S. AI models. Distillation refers to a technique that enhances the performance of one AI model by utilizing the output results of a more advanced model.
On September 8, the National Security Agency (NSA), the Federal Bureau of Investigation (FBI), and the Cybersecurity and Infrastructure Security Agency (CISA) jointly released a cybersecurity document claiming that Chinese AI companies have been conducting industrial-scale distillation activities targeting advanced models from U.S. firms like Anthropic, OpenAI, Google, and SpaceX. The document also suggested that these activities likely occurred with the knowledge of the Chinese government.
In response, Chinese Foreign Ministry spokesperson Mao Ning urged the U.S. on September 9 to stop its baseless accusations and slander against China, emphasizing the need for both countries to strengthen cooperation as AI powers.
Meanwhile, ahead of the scheduled U.S.-China summit on September 24, both countries are preparing to discuss AI safety issues, as reported by Reuters. This will mark the first time since the start of the Trump administration's second term that the two nations will engage in official bilateral talks focused solely on AI.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.