Abstract
This study conducts a quantitative international comparison of AI innovation networks using patent data and social network analysis. AI innovation differs from traditional technological development, relying on modularity, ecosystem-driven structures, and extensive inter-organisational collaboration. The study examines network structures in the United States, Europe, and China, revealing distinct patterns. The US exhibits a competitive ecosystem with central organisations acting as platforms, the EU has a loosely structured network of central firms, and China shows signs of an emerging sectoral platform-based system. Regression analyses indicate that weak ties and structural holes theories apply in the US and China, while the EU relies more on partnerships among central firms. The study acknowledges limitations, including autocorrelation in the US dataset, omission of Japanese firms, and the lack of generative AI patents. Future research should incorporate longitudinal analysis to capture evolving AI innovation trends and the impact of generative AI on inter-organisational networks.