A new report by Schneider Electric has forecast that the electricity demand for AI in Europe could reach as high as 145 terawatt-hours by 2030, putting significant pressure on the grid, if the sector’s expansion remained unchecked.
The report, AI & Energy in Europe, presents four possible scenarios for AI growth across the continent, with electricity consumption potentially as low as 45 TWh under constrained development, or 90 TWh through coordinated sustainable development.
A fourth hypothetical scenario proposes as ‘volatile trajectory oscillating between crisis and recovery’, according to Schneider Electric.
‘A unique opportunity’
“Europe has a unique opportunity to lead in sustainable AI development,” commented Laurent Bataille, executive vice president, European operations, at Schneider Electric. “Currently, it has less than 5% of the world’s computing infrastructure, well below its share of global GDP..
“But to unlock the full potential of AI while meeting our climate goals, it is clear from this research that we must work together to accelerate permitting processes, facilitate a faster and simpler connection to the grid, and continue to invest in decarbonised electricity capacity. Electricity is the backbone of Europe’s digital future, so if managed in the right way, we have the chance to succeed in the digital and energy transitions together.”
As the report noted, each country in Europe is managing AI energy demand from a different starting point, with countries boasting low-carbon electricity and flexible resources capable of absorbing AI growth with minimal emissions increases, and fossil fuel-dependent nations likely to see higher emissions even under efficiency measures.
The study identifies the need for coordinated action across three key areas – developing infrastructure ahead of demand through the deployment of firm and flexible electricity capacity and adaptive systems; the implementation of ‘adaptive’ regulation; and accelerating grid decarbonisation by ensuring AI loads can operate within a low-carbon context, in line with EU-wide performance standards.
Energy trajectory
“AI’s energy trajectory is not inevitable – it depends on the choices we make today on three pillars: technology, regulation, and infrastructure,” added Rémi Paccou, director of sustainability research at Schneider Electric and lead author.
“This research shows the importance of coupling AI technology development with electricity infrastructure expansion and adaptive regulation that responds to real-world conditions. “Sustainable AI in Europe is achievable, but only through deliberate design – and the window to reach it is narrowing without conscious action across the continent.” Read more here.
