Does AI Really Drive the Grid? A Four-Decade Test of the U.S. Energy Footprint

Nademi, Younes; Ebtia, Majid; Khochiany, Ramin; Hoseini, Sayyed Mohammad

doi:10.22050/pbr.2025.525046.1394

Does AI Really Drive the Grid? A Four-Decade Test of the U.S. Energy Footprint

Document Type : Original Article

Authors

Younes Nademi ¹

Majid Ebtia ²

Ramin Khochiany ¹

Sayyed Mohammad Hoseini ²

¹ Department of Economics, Faculty of Humanities, Ayatollah Boroujerdi University, Boroujerd, Iran.

² Gahar Artificial Intelligence Research Group, Ayatollah Boroujerdi University, Boroujerd, Iran

https://doi.org/10.22050/pbr.2025.525046.1394

Abstract

The recent surge of artificial-intelligence (AI) activity has sparked concern that large-scale model training, cloud inference, and data-centre expansion could accelerate national energy demand. We marshal a 21-year annual panel for the United States (2004–2024) that couples multiple AI proxies—technology-stock valuations and a ChatGPT-era dummy—with four aggregate energy series (fossil fuels, nuclear, renewables, total primary energy). A five-stage empirical protocol implemented in Python combines Engle–Granger cointegration testing, higher-order ADF stationarity checks, linear and nonlinear dependence diagnostics (Pearson, Dynamic Time Warping, mutual information), multicollinearity screening (variance-inflation factors), and out-of-sample forecasting with linear regression, decision trees, random forests, and support-vector machines augmented by SHAP explainability. Across all tests we find no evidence that AI developments imprint on national energy use: AI variables cointegrate only with one another, their short-run correlations with energy vanish once trends are removed, their mutual-information scores remain near zero, and their inclusion never improves predictive accuracy beyond a parsimonious macro model driven by GDP, inflation, and population. SHAP rankings confirm that AI features carry negligible explanatory weight relative to conventional fundamentals. We conclude that, to date, AI’s macro-level energy footprint is statistically invisible—any electricity it consumes is either too small or offset by efficiency gains within the wider economy. Policymakers should therefore continue to anchor long-range energy scenarios to established economic drivers while monitoring localised data-centre hotspots that national aggregates obscure.

Highlights

Four-decade U.S. data show AI leaves national energy demand unchanged.

No AI–energy cointegration detected via Engle–Granger or Johansen tests.

Classic factors explain 90 % of demand; AI adds zero forecast skill.

SHAP ranks GDP, population high while AI features show near-zero impact.

Keywords

Artificial Intelligence

Energy Consumption

Cointegration

Machine-Learning Forecasting

SHAP Explainability

Subjects

Energy Management and Engineering

Ahmad, T., Chen, H., & Shah, W. A. (2019). Effective bulk energy consumption control and management for power utilities using artificial intelligence techniques under conventional and renewable energy resources. International Journal of Electrical Power & Energy Systems, 109, 242–258.

Bogmans, C., Ganpurev, G., Gomez-Gonzalez, P., Melina, G., Pescatori, A., & Thube, S. (2025). Power hungry: How AI will drive energy demand (IMF Working Paper No. WP/25/81). International Monetary Fund.

Burian, V., & Stalla-Bourdillon, A. (2025). The increasing energy demand of artificial intelligence and its impact on commodity prices. ECB Economic Bulletin, 2(2025).

Chen, C., Hu, Y., Karuppiah, M., & Kumar, P. M. (2021). Artificial intelligence on economic evaluation of energy efficiency and renewable energy technologies. Sustainable Energy Technologies and Assessments, 47, 101358.

Dauvergne, P. (2022). Is artificial intelligence greening global supply chains? Exposing the political economy of environmental costs. Review of International Political Economy, 29(3), 696–718.

Engle, R. F., & Granger, C. W. (1987). Co-integration and error correction: Representation, estimation, and testing. Econometrica, 55(2), 251–276.

Fang, Y., Lee, C.-C., & Li, X. (2025). Assessing the impact of artificial intelligence on the transition to renewable energy: Analysis of U.S. states under policy uncertainty. Renewable Energy, 246, 122969.

Himeur, Y., Ghanem, K., Alsalemi, A., Bensaali, F., & Amira, A. (2021). Artificial intelligence-based anomaly detection of energy consumption in buildings: A review, current trends and new perspectives. Applied Energy, 287, 116601.

Lee, C.-C., Zou, J., & Chen, P.-F. (2025). The impact of artificial intelligence on the energy consumption of corporations: The role of human capital. Energy Economics, 143, 108222.

Liu, X., Cifuentes-Faura, J., Zhao, S., Wang, L., & Yao, J. (2025). Impact of artificial intelligence technology applications on corporate energy consumption intensity. Gondwana Research, 138, 89–103.

Müller, M. (2007). Dynamic time warping. In Information retrieval for music and motion (pp. 69–84). Springer.

Tao, S., Wang, Y., & Zhai, Y. (2023). Can the application of artificial intelligence in industry cut China’s industrial carbon intensity? Environmental Science and Pollution Research, 30(33), 79571–79586.

Thompson, C. G., Kim, R. S., Aloe, A. M., & Becker, B. J. (2017). Extracting the variance inflation factor and other multicollinearity diagnostics from typical regression results. Basic and Applied Social Psychology, 39(2), 81–90.

Veyrat-Charvillon, N., & Standaert, F.-X. (2009, September). Mutual information analysis: How, when and why? In International Workshop on Cryptographic Hardware and Embedded Systems (pp. 429–443). Springer.

Wang, Q., Li, Y., & Li, R. (2024). Ecological footprints, carbon emissions, and energy transitions: The impact of artificial intelligence (AI). Humanities and Social Sciences Communications, 11, Article 1043.

Yunyun, F., Yongchang, S., Malin, S., & Weiyu, W. (2024). Does artificial intelligence reduce corporate energy consumption? New evidence from China. Economic Analysis and Policy, 83, 548–561.

Zhao, P., Gao, Y., & Sun, X. (2022). How does artificial intelligence affect green economic growth? Evidence from China. Science of the Total Environment, 834, 155306.