Digital transformation is associated with lower ecological footprint across developed and developing economies

The rapid expansion of the digital economy is transforming global production and consumption systems, yet its environmental implications remain insufficiently understood. This study examines the relationship between the Digital Economy Index (DEI) and ecological footprint (EF) across 77 developed and developing countries from 2001 to 2020. An integrated empirical framework combining fixed-effects and System GMM estimation with machine learning techniques, Random Forest (RF), and Support Vector Machine (SVM), is employed to provide complementary econometric and predictive evidence. Econometric results show that digitalization significantly reduces EF on average, with stronger environmental benefits observed in developed economies. Moderation analysis reveals that industrialization and trade openness weaken these environmental benefits, whereas artificial intelligence, ICT development, and renewable energy transitions strengthen them. The transmission-channel analysis further suggests that environmental technological innovation and changes in the energy structure may represent potential pathways linking digitalization to environmental outcomes. The machine learning analysis provides complementary predictive evidence and identifies a nonlinear pattern in the DEI–EF relationship, potentially exhibiting an inverted U-shape in which ecological pressure increases at earlier stages of digitalization but declines at more advanced stages. RF and SVM achieve predictive accuracies of 94% and 80%, respectively. Overall, the findings indicate that digitalization is not inherently green and that its environmental benefits depend on the stage of digital development and broader structural and technological conditions, highlighting the importance of clean energy integration and sustainable digital transformation.