Over the past decade, artificial intelligence (AI) has been increasingly used in firm innovation. While AI has contributed to innovation improvement, direct evidence of its effectiveness in radical innovation is limited. This study fills this gap by empirically investigating the impact of AI on radical innovation and how this relationship is shaped by digital foundation and government subsidy from the perspectives of technological synergy and the external institutional environment. Using panel data from Chinese A-share listed firms from 2007 to 2023, this study empirically tests hypotheses through regression analyses. The findings reveal that AI adoption significantly promotes radical innovation, and this relationship is moderated by the characteristics of a firm’s digital foundation (i.e., degree and rate) as well as government subsidy. Specifically, a high degree of digital foundation hinders AI-driven radical innovation, while a fast rate enhances it. In addition, government subsidy strengthens the positive impact of AI adoption on radical innovation. A heterogeneity analysis further shows that both the timing (early vs. late) and pace (fast vs. slow) of AI adoption exert nuanced impacts: firms that adopt AI later and at a slower pace tend to achieve greater gains in radical innovation. This study advances research on radical innovation in the era of intelligence and provides managerial implications regarding the interplay of AI with internal digital foundation and external government subsidy.

Under geopolitical conflicts, the deployment of emergency green energy in energy-import-dependent countries is constrained by limited upfront capital and uncertainty regarding the conflict duration. This study evaluates financial leasing as a policy instrument by combining a benchmark decision model under conventional procurement with a tripartite evolutionary game involving the energy demand side, financial leasing companies, and the government. The results show that conventional procurement creates an underinvestment-idle-capacity dilemma governed by the critical conflict-persistence probability. Financial leasing alleviates financial constraints and redistributes the risks associated with demand uncertainty, producing market-driven and policy-guided stable equilibria. A global grid search shows that the policy-guided basin expands as the conflict-persistence probability increases. However, government subsidies have an inverted-U-shaped effect because their effectiveness is jointly constrained by the demand-side participation condition and the government’s net-benefit condition. The numerical example shows that combining financial leasing with government subsidies reduces the critical conflict-persistence probability for adopting financial leasing from 0.286 to 0.127, a decrease of approximately 56%, and substantially relaxes the feasible upper bound of the risk-sharing coefficient. These findings provide a theoretical basis for the use of financial leasing as a policy instrument to enhance energy security in energy-import-dependent countries and for the design of complementary incentive mechanisms.