China has emerged as a major AI power, but its ambitions face a critical bottleneck: dependence on imports of advanced semiconductors. Since 2022, U.S. export controls have dramatically tightened this constraint, evolving from targeted restrictions on individual companies into increasingly comprehensive limits on China’s access to advanced AI chips and related technologies. These controls have accelerated a broad transformation in China’s AI strategy.

This talk examines how that strategy is changing. Rather than focusing solely on whether China can replace the most advanced U.S. chips, it shows how the Chinese government and firms are adapting across the whole AI stack – from chips and computing infrastructure to AI models, software, and applications. These strategies seek to work around, compensate for, and create counter-leverage against China’s chip bottlenecks by making better use of available technologies, developing capabilities across other parts of the AI ecosystem, and reducing dependence on critical inputs. The talk also considers where these changes may lead. As China adapts to technology controls and builds greater technological self-reliance, the global AI landscape may become increasingly bifurcated, with two increasingly distinct technological ecosystems centred around the United States and China.

Speaker:

Guangyu Qiao-Franco is an Assistant Professor of International Relations at Radboud University in the Netherlands, an Advisory Board Member of the Leiden Asia Centre, a Senior Researcher on the ERC-funded AutoNorms Project at the University of Southern Denmark, and an Editor of Global Governance. She is also the Principal Investigator of a Dutch government-funded project on China’s export control strategy for dual-use technologies under the China Knowledge Network. Her research focuses on China’s policies, diplomacy, and ethics surrounding artificial intelligence and other emerging technologies.