Industrial Chain Intellectual Property Empowerment and Ecological Development of the Intelligent Economy and Carbon–Energy Metabolic Control Capacity: Causal Inference Based on Spatial Difference in Differences and Double Machine Learning Using Chine

The central challenge of the energy transition lies in whether an economy possesses the institutional capacity to systematically regulate its own energy inputs and carbon emissions. Drawing upon social metabolism theory, this study constructs an indicator of carbon–energy metabolic control capacity (CMCC). Building on business ecosystem theory, it conceptualizes the intelligent economic ecosystem (IEE) and incorporates industrial chain intellectual property empowerment (IP) into a causal framework of institutional provision → ecosystem development → enhancement of metabolic control capacity. Using panel data from 30 provincial-level administrative regions in China covering the period 2010–2022, this study employs a spatial Durbin difference-in-differences (SDID) model and a double machine learning (DML) framework for empirical analysis. The results indicate that industrial chain intellectual property empowerment significantly enhances carbon–energy metabolic control capacity and generates positive spatial spillover effects on neighboring regions through the public diffusion of patent information. Furthermore, intelligent economic ecological development serves as a significant partial mediator between intellectual property empowerment and carbon–energy metabolic control capacity, with the indirect effect accounting for more than one-third of the total effect. This mediating mechanism remains robust after replacing machine learning algorithms, altering sample-splitting ratios, controlling for concurrent innovation policies, and excluding the impact of the COVID-19 pandemic. Path-specific mediation analysis further reveals that computing power acquisition and value transformation together with digital substrate robustness constitute the dominant transmission channels, whereas innovation metabolic flux contributes a relatively smaller mediating effect due to the long gestation period required for translating fundamental research into practical applications. Heterogeneity analysis further demonstrates that the transmission mechanism exhibits full mediation in the dimension of metabolic structure, indicating that the contribution of industrial chain intellectual property empowerment to the clean substitution of energy structures depends almost entirely on the mediating role of the intelligent economic ecosystem. These findings provide clear actionable guidelines for three specific policy-making domains to advance low-carbon transitions. First, intellectual property authorities should transition from quantity-driven patent creation to establishing cross-regional patent navigation and industrial chain IP pooling. Second, digital economy and industry regulators need to prioritize computing power value conversion (CCV) over raw infrastructure expansion to mitigate energy rebound effects. Third, energy and environmental agencies ought to integrate real-time algorithmic dispatching with green finance incentives. Ultimately, this study demonstrates that achieving deep low-carbon transformation requires leveraging institutional public goods to catalyze digital ecosystems, which in turn enable precise, dynamic carbon–energy metabolic control.