Using collaborative governance theory and stochastic actor‐oriented models (SAOMs), this study examines the co‐evolution of networks and participant characteristics in 1505 infrastructure PPP projects involving 3225 participants across 35 large and medium‐sized Chinese cities. The results show that collaborative networks are highly structured and exhibit small‐world characteristics. Furthermore, in the evolving landscape of China's PPPs, the collaboration network has shifted into three governance patterns: government‐SOEs (State‐owned enterprises) driven, SOEs‐consulting driven, and SOEs dominant. In addition, results reveal that network evolution is promoted by transitive triads, the similarity of organizational size, and geographic proximity, while it is hindered by outdegree, similarities of organizational age and type. The findings provide insights for partner selection and the involvement of governments in public service delivery and the development of governance strategies.