Dynamic Network Connectedness and Risk Spillovers Among DeFi, AI-Based, Islamic and Commodity Assets

Against the backdrop of rapid technological innovation and the growing use of alternative investment instruments, this study examines the dynamic connectedness among decentralized finance assets, AI-based stocks, Islamic stocks and commodities. Covering the period from December 2019 to June 2022, we use the time-varying parameter vector autoregression (TVP-VAR) model to measure the magnitude, direction and evolution of return spillovers across Chainlink, Maker, Basic Attention Token, NVIDIA, Amazon, Google, Microsoft, DJIM World, DJIM EM, gold, crude oil and Global X Lithium and Battery Tech. The connectedness literature has examined spillovers across different asset classes during crisis periods. However, much of this literature focuses mainly on pairwise relationships among traditional asset classes, with limited attention to how emerging, alternative and technology-driven assets interact within a single network. We further assess the role of investor sentiment and network topology in identifying systemic transmitters and receivers. The results show strong interconnectedness, with an average total connectedness index (TCI) of 68.81%. Notably, AI-based stocks, especially Microsoft and NVIDIA, consistently emerge as net transmitters of return shocks, while commodities like gold and crude oil serve as absorbers of shocks. The portfolio results show that network centrality improves risk-adjusted performance by reducing volatility and downside risk. These insights have practical implications for policymakers and market participants, offering guidance for developing effective regulatory frameworks, investment strategies and risk management approaches in an increasingly interconnected financial landscape.