President Xi Jinping used the opening of the World Artificial Intelligence Conference in Shanghai to cast China as the standard-bearer for an open, collaborative global AI order, even as Washington tightens limits on access to its own frontier models.

Xi delivered remarks at the Shanghai summit on July 17, framing China as the champion of a new global AI architecture built on open-source technology and international cooperation. The speech arrives as Beijing moves to shape the rules of a field it considers central to its economic and strategic future.

The conference doubled as a showcase for Chinese model builders. Beijing-based startup Moonshot AI unveiled Kimi K3, described by the company as a frontier-level open model that competes with the most capable U.S. systems. Moonshot said it would release the model's open weights on July 27, allowing outside researchers to inspect, modify and self-host the technology.

The open-weights push lands against a backdrop of U.S. export controls that have restricted access to advanced models from American developers, prompting Group of Seven discussions of "trusted-partner" frameworks for approved international access. The friction has deepened a split between regulators seeking control and labs eager to distribute their work.

China's embrace of open models is also a commercial strategy. Several Chinese systems have been deployed inside U.S. firms and adopted by researchers worldwide, extending Beijing's influence through code rather than edicts. The approach contrasts with Washington's tighter stance on who may use the most advanced American systems.

For the delegations in Shanghai, the event was a stage for robotics, autonomous systems and enterprise tools, and a chance to court partners across the Global South. The coming days will test whether the lofty language of cooperation yields durable mechanism, and how far China is willing to bind itself to the international norms it now helps write.

Image source: i.ibb.co