Shiyu Li, Rongkai Chen, Wenji Zhou
Reducing energy intensity is crucial for China’s dual-carbon goals and sustainable urban development. Although numerous studies examine factors affecting energy intensity, the role of innovation networks remains underexplored. This study investigates how urban energy innovation networks influence energy intensity in Chinese cities by establishing a two-layer framework that integrates collaborative innovation structures with energy technology knowledge flows. By mapping collaborative energy technology patents to cities and their technological fields, this study employs social network analysis to reveal the evolution and dynamic interactions of collaboration and knowledge networks. Two-way fixed effects models are used to examine both the direct effects of network position on energy intensity and the underlying mechanisms through knowledge network characteristics. The findings show that China’s urban energy innovation network has evolved from a highly centralized pattern into a more distributed system. Cities with advantageous network positions exhibit significantly lower energy intensity. This result remains robust after a series of robustness tests and endogeneity tests. Mechanism analysis reveals that network position influences energy efficiency through three pathways: enhancing knowledge network openness, increasing technological complexity, and strengthening knowledge connectivity. Furthermore, substantial heterogeneity exists across regions and collaboration types, with the Eastern and Northeastern regions showing the strongest effects and industry-university-research networks demonstrating superior performance.