Energy and Artificial Intelligence: A Customer That Is Also a Tool, and the Race Between Demand and Grid Capacity

Purpose. Energy has a double relationship with artificial intelligence. AI is a growing consumer of electricity, through data centres, and a set of tools that can make energy systems cheaper, cleaner and more reliable. This article asks where AI stands in the energy sector in 2026 and where the sector may be by about 2035. Design and methods. A review combining the share of the energy research literature that involves AI (OpenAlex, 2010–2025), Eurostat's survey of AI use in EU electricity, gas and steam supply enterprises (2021–2025), the International Energy Agency's analysis of energy and AI, and peer-reviewed reviews of machine learning in energy systems. Claims taken from the literature were checked against the sources cited. The scenarios are judgements built on this evidence. Findings. The share of energy works that mention AI rose from 2.8% in 2018 to 11.5% in 2025 (4.1 times as much), and language-model works from 3 to 534, but language models are a small part of the field (3.7% of AI-related works). In the EU, 33.6% of electricity, gas and steam supply enterprises used AI in 2025, up from 17.7% in 2021 and well above the 19.9% for all enterprises, with text mining (20.9%), language generation (17.2%) and workflow automation (14.9%) leading. The International Energy Agency estimates that data centres used 415 TWh in 2024, about 1.5% of world electricity, and will use about 945 TWh in 2030 and about 1,200 TWh in 2035 in its base case, with the United States accounting for nearly half of the growth to 2030. On the supply side, it estimates that AI-based fault detection could cut outage durations by 30–50% and that remote sensing and AI management could unlock up to 175 GW of transmission capacity without new lines, but that broad AI adoption would cut energy-related emissions by only about 5% in 2035. We expect AI tools to become routine in forecasting and grid operations by 2028, the race between demand and grid capacity to decide the pace of data-centre growth, and the net effect on emissions to remain disputed. Originality. The article treats AI in energy as a demand story and a tool story together, sets measured adoption in utilities beside the agency's modelled potential, and marks which numbers are measurements and which are estimates.