Quantile time–frequency connectedness and spillover between artificial intelligence, clean energy, and traditional asset classes: insights and portfolio implications

Abstract The unprecedented rise of artificial intelligence (AI, hereafter) equity and the increasing popularity of clean energy investments have raised concerns among numerous researchers and market participants who are seeking to assess market risks and dependencies. Thus, we explore the connectedness and spillover effects across AI stocks, clean energy, and traditional asset classes under different market conditions (bearish, normal, and bullish). For the empirical analysis, we employ quantile vector autoregression (QVAR, hereafter) and the quantile connectedness approach on daily data spanning from January 2012 to December 2025. The findings of this study demonstrate that within the network, AI stocks, particularly Alphabet Inc Class C (GOOG1), Microsoft Corporation (MSFT), and NVIDIA Corporation (NVDA1), consistently serve as dominant transmitters of return spillovers. On the other hand, findings suggest that clean energy stocks are more susceptible to downside risks and their responses to market environments are asymmetric. Furthermore, the findings demonstrate that the short-term spillovers prevail when the market is bearish and normal, and the long-term spillovers are more prominent during bullish market conditions. Findings further demonstrate that portfolio strategies provide effective hedging benefits, whereas dynamic portfolio approaches may amplify risk in volatile market environments. Additionally, the findings provide significant implications for investors, portfolio managers, and policymakers by identifying the various effects of AI-based stocks and clean energy assets and the necessity to consider market conditions and investment horizons.