
Bessent emphasized that the United States and China must fulfill their leadership responsibilities in competition. The two sides established a super-smart dialogue mechanism but did not commit to restricting technological development.
AI-generated summary
The United States and China are actively competing for global AI dominance. Washington and Beijing have established a new dialogue mechanism for "superintelligence" in the hope of reducing major risks through communication. The Trump administration is unwilling to expand restrictions on AI development, fearing that excessive regulation may cause the United States to lose its technological advantage over China. China has also resisted calls to slow down the development of cutting-edge AI models.
U.S. Treasury Secretary Bessent said that even if the United States and China establish a cooperation mechanism in the field of artificial intelligence (AI) to jointly prevent technological risks, they should not have any illusions about their differences. Both sides must still fulfill their respective leadership responsibilities in competition. As the United States and China actively compete for global AI dominance, Washington and Beijing have established a new dialogue mechanism for "super intelligence" (SI), hoping to reduce major risks through communication, but have not committed to restricting the development of related technologies.
CNBC reported that Bessant pointed out in a prepared speech to the National Committee on U.S.-China Relations in New York on Tuesday that this new mechanism is designed to prevent risks faced by both parties and avoid misunderstandings between the two countries' different systems. The two parties will exchange best practices for testing advanced AI models before they are released to ensure that human judgment can still control this technology, and hope to establish a mutual reporting mechanism for major events involving super-intelligent models.
However, this does not mean that the United States and China will give up their competition. The Trump administration has been reluctant to expand restrictions on AI development, fearing that excessive regulation may cause the United States to lose its technological advantage over China. In a recent interview with CBS, U.S. President Trump said that he does not object to setting up safety guardrails for AI, but he does not want relevant regulations to make the United States fall far behind China.
China has also resisted calls to slow down the development of cutting-edge AI models. China’s Ministry of Foreign Affairs stated in September this year that spreading panic and confrontation will only disrupt global AI governance. Wang Huiyao, founder of the China and Globalization think tank who is closely aligned with Beijing's position, also said last month that the US-China talks should jointly face and manage AI risks.
The idea of establishing an AI dialogue mechanism between the United States and China originated from Chinese President Xi Jinping’s visit to the United States in September. The two sides agreed to launch relevant dialogues and establish communication channels to assist in handling major incidents. The leaders of the two countries are expected to hold their third meeting during the Asia-Pacific Economic Cooperation (APEC) Leaders' Informal Meeting on November 18-19 next week, and Bessant will also accompany him.
In his speech, Bessant emphasized that as science and technology continue to advance, the United States and China must show how to strike a balance between competitive needs and leadership responsibilities. He admitted frankly that the interests of the two countries are not always completely consistent, and the methods they adopt when facing various challenges are not necessarily the same.
It is worth noting that this event was held at a time when NVIDIA CEO Jen-Hsun Huang had an extramarital affair with a Chinese businesswoman suspected of Chinese espionage, and the related controversy continued to attract attention. Huang Jen-Hsun visited Beijing with the Trump delegation in May this year, and was also invited to attend a state banquet in Washington in September. He admitted to having an extramarital affair with the Chinese businesswoman, and U.S. intelligence officials suspected that the woman was involved in espionage activities, further raising concerns about corporate governance and national security.
It is still unclear whether Huang Renxun will attend the US-China summit next week. The South China Morning Post reported that business leaders such as Tesla and SpaceX CEO Musk, as well as Mastercard CEO Milbach, may attend relevant events with the Trump delegation.
AI outlook — possibilities, not facts
The leaders of the United States and China will hold their third meeting during the APEC Leaders’ Informal Meeting on November 18-19 to discuss AI cooperation and competition.
Very likely · Within days
The United States and China will use a new dialogue mechanism to exchange best practices for testing advanced AI models before they are released, and establish a mutual reporting mechanism for major events involving super-intelligent models.
Likely · Within weeks

Huida CEO Huang Renxun has lobbied U.S. President Trump to relax export controls on AI chips to China, especially H200 chips. He argued that it would be more beneficial to the United States to make China dependent on U.S. technology systems, and believed that related computing technology was not a national security issue. He believed that he had successfully persuaded Trump to accept his argument.

China's Northwestern Polytechnical University, which has been included in the export control blacklist by the United States, still purchased two Huida H200 high-efficiency AI chips last month for 874,700 yuan for use by the School of Management. They obtained them through Shaanxi Boyuan Network Technology Company, highlighting the loopholes in Washington's measures to restrict China's access to advanced technology.

Reuters reported that Chinese universities such as Northwestern Polytechnical University used domestic suppliers to purchase AI chips such as Huida H200 that are subject to U.S. export restrictions, highlighting Washington's loopholes in curbing China's acquisition of advanced technology. Huida CEO Huang Renxun will go to the U.S. Congress for a hearing on November 18 and will face cross-party pressure to strengthen control.

The extramarital affair scandal between Huida CEO Huang Renxun and Chinese businesswoman Huang Le may have major political consequences in Washington. U.S. China hawks are taking this opportunity to push hard lines shelved by President Trump, including restricting the export of advanced AI chips to China. Congress has asked Huang to attend a closed-door briefing, and Huida's influence in Washington may be significantly weakened.

Biographer Witte pointed out in the New York Times technology podcast Hard Fork that the data center in Johor, Malaysia involved in the case may have used "fake servers" to transport Huida chips to China through the back door. Huida CEO Jen-Hsun Huang admitted to having an extramarital affair with Megaspeed senior executive Huang Le, who was accused by the U.S. government of being a Chinese spy. The Department of Commerce investigated whether Megaspeed smuggled Huida chips into China, and the entire case was handed over to the Department of Justice late last year. Witte believes that the Malaysian data center is essentially a transfer station for chip smuggling. The industry speculates that there are two methods of tearing off labels and pasting fake labels and fake servers. The U.S. Department of Commerce has difficulties in enforcing laws overseas, and the U.S. government lacks substantial supervision capabilities after chips enter the computer room. Witte emphasized that Huida has significantly tightened control this year, requiring data centers to submit customer certificates to ensure that end users are legal and not in China.

The Seventh Manned Spaceflight Academic Conference revealed that China’s manned lunar landing mission will adopt the “two launches and docking around the moon” route. The article compares the differences in the moon landing plans of the United States, Russia, Europe and China, analyzes engineering trade-offs, technical cooperation and "safe zone" rule disputes, and points out that the moon landing is a long-term technical endurance race.