Unlocking Carbon Emission Efficiency Through Corporate AI Computing Power Accessibility: Empirical Evidence from China

Faced with dual challenges of reducing carbon emissions and improving corporate operating performance, companies increasingly need access to favorable external computing resource conditions that may support information processing for low-carbon transformation. Drawing on organizational information processing theory, this study explores the influence of corporate AI computing power accessibility (AICPAcc) on carbon emission efficiency (CEE). Using Chinese A-share listed companies from 2012 to 2023, the results reveal that AICPAcc has a significantly positive impact on corporate CEE. Mechanism analyses indicate that this effect operates via the improvement of both the quantity and quality of green innovation. Furthermore, high levels of analyst attention and environmental penalties are found to moderate this relationship. Heterogeneity analyses suggest that the positive influence of AICPAcc on CEE is more pronounced in companies with higher technological intensity, higher low-carbon energy share, and those located in regions with greater communication optical cable density. These findings extend organizational information processing theory to the decarbonization context and provide practical and policy implications for improving carbon emission efficiency through better potential accessibility to regional AI computing resource conditions.