Dynamic interdependence between green resources, AI and semiconductor market

Abstract This study investigates return spillovers among green resource, artificial intelligence (AI), and semiconductor markets from 1st June 2018 to 30th March 2026 using the R 2 -decomposed connectedness framework and connectedness-based portfolio strategies. We decompose overall connectedness into contemporaneous and lagged components to identify the size, direction, and timing of spillovers within this rapidly evolving innovation ecosystem. Results reveal an extremely High Total Connectedness Index of 83.46%, indicating these markets function as an integrated financial network rather than independent segments. Contemporaneous spillovers dominate (81.04%) relative to lagged spillovers (2.42%), implying that information is rapidly incorporated into prices with limited scope for delayed adjustments, especially under uncertainty. Directional connectedness highlights distinct roles: hydrogen and water are net receivers of shocks, suggesting sensitivity to technology-driven developments, while AI and robotics are net transmitters, reflecting their expanding systemic influence. Semiconductor markets operate as key interfaces, both transmitting and receiving substantial shocks, underscoring their foundational role across AI, automation, and clean energy. Dynamic analysis shows connectedness intensifies during global stress events COVID-19, geopolitical conflicts, trade tensions, and financial disruptions amplifying cross-market transmission. Portfolio evidence demonstrates that a Minimum Connectedness Portfolio outperforms standard approaches, supporting the use of connectedness measures for diversification and resilience. Policy implications include bolstering semiconductor supply chain robustness and enhancing coordination across industrial, digital, and climate policies. Regulators should monitor interconnected markets in real time and act promptly with macroprudential tools given the predominance of contemporaneous spillovers. Limitations include reliance on selected market proxies; future research should extend to additional innovation sectors and apply nonlinear, regime-switching, frequency-domain, or quantile connectedness, as well as firm-level and cross-country analyses to refine understanding of sustainable-digital financial interdependence.