Events

Professor Jiasun Li
Robert D. Johnston Associate Professor of Finance
Costello College of Business
George Mason University

Cryptocurrency mining and AI data centers have opened live policy discussions over their grid and carbon impacts. We build a model of an integrated grid with endogenous renewable and fossil capacity serving large flexible loads (LFLs)—intensive electricity demands that can scale elastically in real time without disrupting subsequent operations. We show that when the amortized capital costs of LFL hardware and renewable capacity fall below characterized thresholds, introducing LFLs with demand response yields a triple win: higher utility profit, lower blackout probability, and lower carbon emissions. This occurs as LFLs monetize otherwise uncompensated renewable surplus and substitute for fossil backup by absorbing renewable intermittency. Legacy capacity may however reverse the carbon leg. The result is robust across market structures, e.g., deregulated competition or rate-of-return regulation at cost recovery (an explicit condition signs the direction of Averch–Johnson over-investment). Calibrating every input to measured values at disclosed U.S. mining sites, we find the market alone essentially
never lands in the carbon window (the rare sites inside have commission-set tariffs); Averch–Johnson padding points toward fossil; the curtailment rule and the host grid’s joint distribution of demand and renewable output move the boundary several-fold; and reward dilution closes the window everywhere so the carbon channel works at the margin of adoption and cannot scale. The triple win is thus a designed outcome, and the model supplies the levers that deliver it.