US Treasury's Bessent, China's He to launch talks on AI, trade, critical minerals

On Sunday, US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng will convene in Manhattan to discuss potential agreements on AI, tariffs, and critical minerals ahead of a high-stakes summit between US President Donald Trump and Chinese President Xi Jinping. The talks, held at JP Morgan Chase's headquarters, will also involve US Trade Representative Jamie Pressley Greer and will cover the status of a US-China trade truce expiring on November 10, the flow of Chinese rare-earth magnets and critical minerals, and potential guardrails for artificial intelligence following security breaches.

Analysts anticipate that Washington and Beijing will reach small steps to prevent escalating tensions, but a breakthrough seems unlikely. Anna Ashton, a China trade analyst, expects a status quo outcome as both sides aim to avoid further escalation in their delicate trade relationship. The meeting follows a series of similar discussions between Bessent, He, and Greer over the past 16 months, focusing on reducing tariffs, increasing Chinese purchases of US agricultural goods, and purchasing more than 200 Boeing aircraft.

The Bessent-He-Greer meeting comes after the November 2025 truce in Busan, South Korea, which capped US tariffs on Chinese goods at about 20 percent after multiple rounds of tit-for-tat escalation. The US Supreme Court later invalidated the Trump tariffs, and the administration has been restoring them with new authorities. The truce saw China pledge to restore critical minerals flows to US and global users, but China's performance has been subpar, according to a senior US official.

The AI discussions are particularly significant as the US and China are the two major forces driving AI development and global adoption. Bessent expects the talks to cover both open- and closed-weight AI models, with the US remaining the leader in AI and open to discussions on avoiding shared risks and preventing the bifurcation of AI systems. The discussions will also address potential guardrails for AI to keep powerful models out of the hands of malicious non-state actors.