ARTIFICIAL INTELLIGENCE-DRIVEN ELECTRICITY DEMAND AND PRICE VOLATILITY IN THE UNITED STATES: THE MODERATING ROLE OF RENEWABLE ENERGY, BATTERY STORAGE AND GRID INVESTMENT

Abstract

The rapid expansion of artificial intelligence (AI) and data-centre infrastructure is creating a structural increase in electricity demand in the United States, raising concerns about electricitymarket stability, infrastructure adequacy, energy security and consumer costs. This study examined the asymmetric effects of AI-driven electricity demand on U.S. electricity-price volatility and assessed whether renewable energy, battery storage and grid investment moderated these effects. Monthly secondary data were obtained from the U.S. Energy Information Administration, Lawrence Berkeley National Laboratory, Federal Energy Regulatory Commission, Federal Reserve Economic Data and the National Renewable Energy Laboratory. A nonlinear autoregressive distributed lag (NARDL) model was employed to distinguish between positive and negative AI-related electricity-demand shocks and estimate their short- and long-run effects. The results indicated that positive AI-demand shocks exerted a significant positive longrun effect on electricity-price volatility (β = 0.284, p < .001), while negative shocks produced a smaller but significant effect (β= 0.117, p = .026). The Wald test rejected long-run symmetry (W = 9.72, p = .002), confirming that AI-related demand expansion and contraction generate unequal electricity-market responses. In the short run, positive AI-demand shocks remained significant (β= 0.176, p = .006)), while the error-correction coefficient (β = -0.417, p < .001)) indicated substantial adjustment towards long-run equilibrium. Renewable energy (beta = -0.083, p = .008)), battery storage (β= -0.116, p = .003) and grid investment (β= -0.094, p= .010) significantly moderated AIdemand effects. The study contributes to energy economics and AI economics by establishing AI infrastructure as an asymmetric electricity-market demand shock and identifying infrastructure conditions that reduce its volatility effects. The findings provide evidence relevant to U.S. federal and state energy planning, grid-investment decisions, large-load regulation and AI infrastructure policy, supporting coordinated investment in renewable generation, storage and transmission capacity to sustain AI growth without imposing disproportionate costs on other electricity consumers.