AI energy costs are unsustainable: shocking European Environment Agency report

The rise of artificial intelligence in Europe is accelerating rapidly, and its environmental costs can no longer, and must no longer, be ignored: The European Environment Agency's new report, "Artificial intelligence and sustainable consumption in Europe", highlights the growing pressure that AI infrastructures are exerting on the continent's resources. The alarm bells are ringing.

At the heart of this problem are data centers: projections indicate that, largely stimulated by AI, electricity demand in this sector is set to almost double by 2030. Europe already accounts for 15% of the world's electricity consumption by data centers, and the concentration of these infrastructures around major urban centers is putting a strain on local networks.

But, as the report points out, the impact goes beyond the energy question alone:

• water: AI-dedicated data centers could boost annual water consumption for cooling and power generation to around 1,068 billion liters by 2028, an eleven-fold increase on 2024 estimates ;
• raw materials: AI infrastructures depend on a small number of critical raw materials, notably gallium, germanium and tantalum, the extraction of which entails considerable environmental and social costs;
• governance: current EU regulations focus mainly on the drive of AI systems, only partially regulating the energy consumption linked to inference (daily use phase), which now accounts for 80-90% of AI computing power.

The need to revise public policies

With the rapid transformation of the European economy under the impact of AI, our continent has a duty to analyze how targeted policies can both guarantee our economic competitiveness and manage the trade-offs associated with the deployment of these technologies, at a time when environmental pressures must imperatively diminish.

The Agency's latest report emphasizes precisely this point: rapidly expanding AI technologies, capable of reshaping systems, have the power to redefine how economies work, how consumption decisions are made and how value chains are organized.

Without clear political guidance, these changes risk increasing demand for energy and materials, reinforcing resource-intensive business models resource-intensive economic models, accentuate strategic dependencies and exacerbate social inequalities. It is unlikely, therefore, that efficiency gains alone will be enough to reduce global environmental pressure.

Data presented in the document show that the rapid growth of data centers is in turn fuelling increased demand for energy, water and critical raw materials. The report underlines that data centers, networks and terminals, taken as a whole, generate a growing environmental footprint that energy efficiency improvements alone will struggle to offset.

The analysis comes against a backdrop of heightened geopolitical competition, economic uncertainty and strategic interdependencies, says the agency. In this context, digital technologies and artificial intelligence are increasingly seen as central to Europe's competitiveness, resilience and strategic autonomy.

Meeting the dual challenge of the combined green and digital transitions is therefore not just an environmental issue, but also a strategic one, requiring carefully considered choices on how to guide and regulate innovation.

These findings are particularly relevant (or at least should be) to the implementation of key EU legislative and policy frameworks, which link digital transformation to sustainability and competitiveness.

Some of these results should clearly be incorporated into the EU Artificial Intelligence Act as soon as possible. The majority of its provisions will only be fully applicable from August 2, 2026.

This regulation defines the rules governing the development and use of AI systems throughout the EU, as well as the major strategies that place digitization at the heart of economic competitiveness, while reinforcing the objectives of the ecological transition.

Better alignment between digital policies, consumer measures and environmental objectives, the EEA stresses, will be essential to ensure that Europe's digital transformation supports climate neutrality, resource efficiency and long-term resilience.

Sources: European Environmental Agency / European Environmental Agency/Facebook