Thi Loan Le
Purpose This study examines how manufacturing SMEs convert sustainability-oriented digital capability into sustainable performance. Drawing on the Dynamic Capabilities View, it develops and tests a model in which green artificial intelligence (AI) capability, resource and coordination flexibility, and inbound and outbound open innovation jointly shape circular economy practices and, ultimately, sustainable performance. Design/methodology/approach Guided by a positivist philosophy and a quantitative survey design, data from 211 Vietnamese manufacturing SMEs were analyzed using PLS-SEM and polynomial regression with response surface methodology. Findings The results indicate that green AI capability enhances both resource and coordination flexibility. Resource flexibility supports inbound open innovation, whereas coordination flexibility drives outbound open innovation. Both forms of open innovation enhance circular economy practices, with balanced configurations yielding superior results. In turn, circular economy practices significantly contribute to sustainable performance. Practical implications The study provides actionable insights for managers and policymakers on leveraging green AI and innovation ecosystems to enhance resource efficiency, reduce waste, and meet sustainability regulations, particularly in emerging economies. Originality/value This study conceptualizes green AI as a sustainability-oriented dynamic capability and highlights the interplay between flexibility and open innovation in advancing circular and sustainable performance among manufacturing SMEs.