The market will resolve to YES if, at any point on or before December 31, 2032 (11:59 PM UTC), the United States and the People's Republic of China formally sign a bilateral or multilateral treaty, executive agreement, or binding international accord that includes provisions regulating the development or deployment of frontier AI.

"Frontier AI" means the most capable general-purpose AI models at the time of signing, and the systems used to develop them. Qualifying provisions include, for example: limits or thresholds on model training or capabilities; safety testing, evaluation, or incident-reporting requirements; restrictions on dangerous capabilities such as biological, chemical, or cyber weapons uplift; and inspection or verification regimes for frontier model development (including verification that relies on tracking computing hardware).

An agreement counts only if it contains at least one provision that directly regulates the development or deployment of frontier AI. Agreements covering only related matters do not count, including: export controls, tariffs, or trade terms on semiconductors, chipmaking equipment, or other AI hardware; investment restrictions; data privacy or cross-border data rules; intellectual property; and rules for narrow or application-specific AI (e.g. facial recognition, autonomous vehicles). A broader agreement that includes such matters alongside at least one qualifying frontier AI provision does count.

Joint statements, MOUs, declarations, and other non-binding commitments do not count.