Research on the influence mechanism of cooperative models in China’s autonomous vehicle industry

Abstract
Autonomous vehicles (AVs) represent a disruptive technological trajectory that fundamentally reshapes the global automotive industry. Their complex, interdisciplinary, and cross-sector nature makes collaborative innovation both essential and inevitable for industrial advancement. Within collaborative innovation, selecting appropriate cooperation models is critical for establishing stable partnerships and fostering sustainable industrial ecosystems. This study categorizes AV cooperation models into five types—kinship-related, geo-related, industry-related, academic-related, and hybrid—based on relational ties. Using multidimensional proximity as antecedents and interorganizational trust as a mediator, we employ PLS-SEM to examine how innovation actors select cooperation models in China’s AV industry. Results show that institutional, geographic, and organizational proximity directly influence both current and future model selection. Social proximity affects only future selection—a null effect attributed to four countervailing mechanisms: active avoidance of over-embeddedness, substitution by other proximity dimensions, selection-stage filtering, and strategic openness to novel partners. Cognitive proximity has no direct effect but operates fully through trust. Mediation analysis further reveals distinct pathways: Cognitive proximity predominantly operates through trust, whereas institutional proximity exerts primarily direct effects. Grouped analyses show that different cooperation models are driven by distinct proximity configurations: kinship models by organizational and geographic proximity, geo-related models by institutional and organizational proximity, industry-related models by institutional and cognitive proximity, academic models by geographic and institutional proximity, and hybrid models by institutional and organizational proximity. These findings extend proximity theory to cooperation model selection in the AV context, distinguish current versus future temporal dynamics, and offer actionable guidance for practitioners navigating China’s rapidly evolving AV ecosystem.