Why Sustainability-Oriented Resources Translate Differently Into Sustainable Performance and Durable Advantage: A Socio-Technical View of Green Innovation Ambidexterity

Firms pair artificial intelligence with green innovation to cut environmental harm and hold off rivals, yet whether these resources buy lasting advantage or only passing gains remains unsettled. We answer it with a socio-technical reading. We trace how each shapes exploratory and exploitative green innovation and how those two modes feed sustainable performance, green competitive advantage and sustainable competitive advantage. Covariance-based structural equation modelling of 318 firms in the logistics and supply chain sector of the United States shows that Knowledge Integration Capability drives both modes more forcefully than AI Infrastructure does. The two resources do not directly achieve these outcomes. Each works only once green innovation converts it, and the conversion is selective rather than uniform, because two of the routes that run from AI Infrastructure do not reach significance. Exploration supports all three outcomes. Exploitation does not. It lifts current performance and presents green standing yet works against durable advantage, with the path from knowledge integration through exploitation running significantly negative. We read this as a ‘competency trap’ inside green innovation: Refinement pays now, but only sustained exploration holds an advantage that lasts.