Analysis
Representative Ro Khanna, the top Democrat on the House Select Committee on the Chinese Communist Party, is asking leading Chinese AI labs to cooperate on pacing frontier AI development, according to The Information and a companion Washington Post AI & Tech Brief. Khanna has contacted DeepSeek and Moonshot directly, and is separately asking Alibaba, requesting their commitment to join American labs in a binding international agreement to pace the technology's development.
An Unusual Posture From The CCP Committee
Khanna's role on the House Select Committee on the Chinese Communist Party has generally centered on export controls, chip restrictions and broader technology decoupling from China. A direct request for cooperation on pacing frontier AI development is a different kind of overture -- an invitation to coordinate rather than a restriction to comply with -- and it reflects a growing view among some U.S. policymakers that unilateral U.S. slowdowns mean little if Chinese labs simply continue at full speed and close the capability gap regardless.
“Fortune has reported that Chinese labs, facing tighter compute access under U.S.”
Extending A Domestic Pledge Across Borders
The request follows directly from a wave of U.S. lab leadership publicly endorsing some form of capability pacing this month: Anthropic CEO Dario Amodei's essay arguing frontier labs should slow the pace of capability gains drew public agreement from both Sam Altman and Elon Musk. Khanna's ask essentially extends that domestic pledge into a cross-border proposal -- if U.S. labs pace themselves but Chinese labs don't, the argument goes, the pledge accomplishes little beyond ceding relative ground. Fortune has reported that Chinese labs, facing tighter compute access under U.S. export controls, have leaned into efficiency gains as a competitive strategy -- a posture more consistent with racing to close the gap than with voluntarily slowing down.
Why Chinese Labs Have Limited Incentive To Agree
DeepSeek, Moonshot and Alibaba's Qwen team have all built their competitive position substantially on demonstrating they can match or approach U.S. frontier capability despite compute constraints -- a binding pacing agreement would ask them to accept externally-verified limits at precisely the moment their relative position has been improving. No Chinese lab had publicly responded to Khanna's request as of this reporting, and there is no enforcement mechanism attached to the proposal; it remains a diplomatic ask rather than a framework with any binding force.
The Channel That Might Actually Work
The realistic path for this kind of agreement, if it happens at all, likely runs through formal U.S.-China diplomatic channels rather than direct lab-to-lawmaker outreach -- Treasury Secretary Scott Bessent has separately signaled U.S. openness to discussing AI "shared risks" with China, a parallel track that carries more institutional weight than one representative's direct ask to individual companies. Until a formal bilateral or multilateral mechanism exists, this stays a notable signal of policymaker intent rather than a change in how fast any Chinese lab is actually building.