Concentrated siting of AI data centers drives regional power-system stress under rising global compute demand
Authors: Danbo Chen, Zijun Zhou, Yongyang Cai, Jiahong Qin, Ani Katchova, Lei Chen
Organizations: The Ohio State University, Columbus, OH 43210, USA. · Meta Platforms Inc., Menlo Park, CA 94025, USA. · Zhejiang A&F University, Hangzhou, Zhejiang 311300, China.
The rapid rise of generative artificial intelligence (AI) is driving unprecedented growth in global computational demand, placing increasing pressure on electricity systems. This study introduces an AI-energy coupling framework that combines large language models (LLMs)-based analysis of corporate, policy, and media data with quantitative energy-system modeling to forecast the electricity footprint of AI-driven data centers from 2025 to 2030. Results show that the new AI infrastructure is highly concentrated in North America, Western Europe, and the Asia-Pacific, which together account for more than 90% of projected compute capacity. Aggregate electricity consumption by the six leading firms is projected to increase from roughly 118 TWh in 2024 to between 239 TWh and 295 TWh by 2030, equivalent to about 1% of global power demand. Regions such as Oregon, Virginia, and Ireland may experience high Power Stress Index (PSI) values exceeding 0.25, indicating local grid vulnerability, whereas diversified systems such as those in Texas and Japan can absorb new loads more effectively. These findings demonstrate that AI infrastructure is evolving from a marginal digital service into a structural component of power-system dynamics, underscoring the need for anticipatory planning that aligns computational growth with renewable expansion and grid resilience.
Figures & tables
Figure 1: (a) Global distribution of data-center sites showing the transition from legacy to AI-specific infrastructure. Small blue markers represent pre-existing enterprise and cloud data centers, while larger color-coded circles denote newly identified AI data centers with varying estimated probabilities of siting. The ringed overlays summarize dominant regional determinants.(b) Firm–determinant intensity heatmap of AI data-center expansion. Color gradients represent the relative importance of the estimated key siting factors.
Figure 2: (a) Firm-level electricity consumption of data centers from 2025 to 2030 under three site-expansion scenarios. Each trajectory is derived from firm-specific expansion rates (15%, 25%, and 35% for conservative, neutral, and optimistic scenario, respectively) combined with the projected share of AI-intensive workloads by firm. (b) Aggregated global electricity consumption of data centers from 2018 to 2030, showing historical trends and scenario-based forecasts. The solid grey line represents historical values, while the colored dashed lines denote the mean and range across the three scenarios.
Figure 3: (a) Regional distribution of AI data-center electricity demand by firm in 2030, combining U.S. states and overseas host countries. Error bars denote scenario uncertainty across conservative, neutral, and optimistic scenarios. (b) Regional Power Stress Index (PSI) in 2030, defined as the ratio of data center electricity demand to total regional generation capacity. Dashed horizontal lines indicate stress thresholds of 0.10, 0.25, and 0.40 corresponding to industry reliability benchmarks reported by the International Energy Agency, EPRI, and IEEE. Values above 0.25 represent regions where industrial digital loads may require dedicated transmission expansion or peak-support resources, while values above 0.40 indicate extreme concentration of demand relative to local generation.
Figure 4: The Workflow Diagram
Figure 5: Data includes selected AI models and is shown in log scale. Estimates are based on publicly reported numbers and calculations by Epoch.ai. Source: Epoch.ai, adapted from Financial Times ( https://ig.ft.com/ai-data-centers/ ).
Figure 6: Computation power of leading AI systems, measured in petaflops (10 15 floating-point operations per second). Source: Epoch.ai, adapted from Financial Times ( https://ig.ft.com/ai-data-centers/ ).
Jun 8, 2026·Mohammad Hemmati, Gbemi Oluleye, Vassilis M. CharitopoulosPower Systems
Department of Chemical Engineering, Sargent Centre for Process Systems Engineering, University College London (UCL), Torrington Place, WC1E 7JE, London, United Kingdom · Centre for Environmental Policy, Imperial College London, London SW7 1NE, UK