Wise Europa
Europe wants to catch up in the global race for artificial intelligence. The European Commission’s AI Continent Action Plan includes up to five AI gigafactories, backed by multi-billion in investment, while the EU aims to at least triple its data centre capacity within the next five to seven years.
There are good reasons for this ambition. Computing infrastructure is becoming as important for European competitiveness as energy infrastructure, semiconductors or access to critical raw materials. Yet there is a risk that AI continues to be treated mainly as a digital policy issue, while its very physical dimension is overlooked. AI needs servers, grid connections, transformers, cooling systems, electricity and, depending on the technology used, significant quantities of water. The development of AI infrastructure is therefore relevant also for its climate and energy policies.
In a communication published earlier this year, the Commission pointed to the possible synergies between the energy and AI sectors. AI-driven solutions may help reduce some of the investment needs associated with the development of the energy system, for example by improving grid management, forecasting or the use of existing infrastructure. The difficulty is one of timing. Bringing them into widespread use across the energy sector will take time, and it is far from certain that energy companies and system operators will be among the first beneficiaries of Europe’s AI expansion. In the meantime, the growth of AI infrastructure will place additional pressure on the energy system which are already facing increasing constraints. The costs may therefore appear much earlier than the benefits.
The International Energy Agency expects global electricity consumption by data centres to roughly double by 2030, reaching around 950 TWh, with AI being the most important driver of this growth. At the same time, energy infrastructure develops much more slowly than the digital infrastructure it is expected to supply. A data centre can be built within a few years, while new grid infrastructure or generation capacity often requires considerably longer development cycles.
This expected increase in electricity demand also creates an interesting parallel with the EU’s hydrogen strategy. When the Union developed the regulatory framework for renewable hydrogen, policymakers were concerned that electrolysers could consume existing renewable electricity and indirectly increase fossil-fuel generation elsewhere in the system. The response was the additionality principle, together with requirements linking hydrogen production to new renewable generation. However, applying the same approach to data centres is not an option. The case of the hydrogen market demonstrates how easily detailed sustainability requirements can also increase costs and become an obstacle to investment.
At the same time, data centres should not be viewed only as an additional burden on the system. As the Commission has highlighted, AI can improve renewable generation forecasting, support grid management and demand response, optimise storage and help predict infrastructure failures. Data centres can also become a source of flexibility, while the heat they generate can potentially be recovered and used in district heating systems. The relationship between AI and energy is therefore genuinely two-sided.
Realising the benefits of AI while limiting these costs will therefore require more than separate digital, energy and climate strategies. It calls for a more horizontal approach, linking infrastructure planning, energy policy, climate objectives and the development of the AI sector. It also requires a more transparent, evidence-based discussion about who bears the costs of this transition, who captures its benefits and how both are distributed between companies, consumers and local communities.
Ultimately, Europe’s AI ambitions and its energy transition are no longer separable policy tracks. The PRISMA project’s modelling makes clear that AI-driven data centre growth, while significant, is manageable within the EU’s broader electrification pathway — but only if grid planning, climate policy and digital strategy are designed together rather than in isolation. Getting this right will determine whether AI becomes a genuine accelerator of the energy transition, or simply another source of strain on a system already under pressure.
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No. 101081604 – PRISMA. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Climate, Infrastructure and Environment Executive Agency (CINEA). Neither the European Union nor the granting authority can be held responsible for them.