The pervasive integration of Artificial Intelligence (AI) is creating unprecedented demands on global energy infrastructure, driving a growing imbalance between burgeoning needs and existing, often aging, supply networks. As energy demand from data centers is projected to triple by 2028 — potentially capturing up to 12 percent of total U.S. national electricity consumption—this paper defines the magnitude of the AI-Energy Nexus. We propose a physically grounded policy framework that establishes the necessity of the Atomic Energy Transition. Our analysis highlights the critical role of new nuclear generation as the only viable pathway, leveraging its vast energy density and innovative fuel breeding capacity. The paper concludes by exploring how AI itself can serve as an essential catalyst for this transition, utilizing its power to enable resource allocation and streamline reactor deployment, ensuring the long-term viability of technological growth.
Pekny et al. (Mon,) studied this question.