From algorithms to electrons: return transmission dynamics between AI and energy tokens

The convergence of AI and the energy sector represent a paradigm shift in the financial landscape and has given rise to a new class of digital assets, including AI and energy tokens. This study examines the return connectedness between AI and energy tokens in the time–frequency domain using the TVP-VAR frequency connectedness framework. Our results indicate a moderate level of return transmission between AI and energy tokens. The frequency analysis shows that approximately 83% of this transmission occurs in the short term (1–5 days), with the remainder at medium- to long-term horizons, suggesting that the token market absorbs information rapidly. In net terms, AI tokens generally emerge as net transmitters of return shocks to the system, whereas energy tokens are predominantly net receivers. Our dynamic analysis shows that connectedness between AI and energy tokens strengthens during periods of market stress. Our portfolio analysis reveals that the multivariate minimum-connectedness portfolio reduces the variance of single-token positions in WPR and GRID by approximately 80% at the weekly horizon, whereas hedging effectiveness deteriorates markedly for several tokens at longer horizons, underscoring the value of dynamic rebalancing. These findings offer practical guidance for investors, portfolio managers, and regulators navigating digital asset markets.