The global race for artificial intelligence is usually described as a competition over technical matters such as algorithms, computing power and data, or in policy matters such as data protection. Yet this new chapter in human ingenuity is more complicated and connected with various other systems such as those of energy, economy and security. As governments seek to organize a strict policy framework for artificial intelligence, economists and social scientists highlight the connection between AI capacity and energy security. For the EU in particular, the next phase of the AI race may be catalytic for the adjustments of the EU to the latest challenges that humanity faces.
The connection of AI to energy becomes clearer when taking into consideration the latest data and projections. In particular, the International Energy Agency projects that the consumption of energy by data centers could double to around 945 terawatt-hours (TWh) by 2030. In this context, AI is expected to be the most significant driver of this growth. In the specific context of the current analysis, in Europe, data-centers’ electricity consumption is expected to rise by more than 70% over the same period (see IEA, 2025).
However, these insights should not be interpreted as evidence of an imminent global energy shortage but as a sign to scale up the shift towards a more extensive and efficient policy network. Specifically, the immediate problem is geographical dispersion and energy supply. Data centers disperse across the globe around locations which offer reliable sources of energy, infrastructures and connectivity. Under these circumstances, demand for energy on a global scale could generate severe pressure on electricity networks and energy markets. This would directly influence the EU’s ambitions for energy security and autonomy. On that note, there have been efforts from the European Commission to further securitize the matter through further initiatives towards a clearer and more extensive management plan. Indicatively, the Commission launched a call to establish technological sovereignty by establishing up to seven AI centers (see European Commission, 2026a). The cost of such an initiative is great, with a total of 30 billion euros being directed to materialize this plan, with the aim being to give European companies and institutions the infrastructures needed to train and operate advanced AI models. However, this still requires energy capacity, since a continent cannot be technologically autonomous, in matters of AI, if infrastructures cannot have secure, constant and affordable access to energy.
These circumstances create a difficult policy paradox. The EU wants to expand AI capacity, accelerate electrification and decarbonization, and keep energy secure, green and affordable for industries and households (see European Commission, 2026b). Even though these objectives may be compatible in the long term, they are rather complex to be balanced and organized in short term, since they compete. As a result, Europe’s AI strategy must be organized based on the reconciliation of the aforementioned parameters and how macro and microeconomic parameters are regulated and organized. This won’t be the result of policy orthodoxy, considering AI simply as an additional burden on energy systems, but through pluralization, by considering it as part of the solution. The Commission has to systematically address both sides of this focus, by acknowledging both the sustainable integration of data centers into the energy systems and the possibility that the deployment of AI could improve forecasting of the energy-market trends and the optimization of the energy flow.
In conclusion, the strategic point therefore is not whether Europe should choose between AI development and energy security. It is whether it can design a framework which would consider multiple parameters. Artificial Intelligence creates new dependencies and influences various systems such energy, by re-adjusting the consumption and cost of energy, the economy, by creating a need for a more efficient and up-to-date taxation and social welfare policy framework, and security, in its purest form. Europe and the world have spent the past several years learning that economic dependence can quickly become a security vulnerability, thus proving the possibility of subsequent domino effects. This should apply to artificial intelligence. After all, there can be no credible AI sovereignty and autonomy without an institutional system capable of organizing it and an energy system capable of powering it.
Written by Georgios Tsialas
Georgios Tsialas is a PhD Candidate in Politics (Research) at King’s College London. His research focuses on crises, polycrises, securitization, comparative politics and computational social sciences. He holds an MSc in Comparative Politics (LSE) and a BA in Political Science and History (Panteion University). He has served as Senior Political Analyst and Editor at the Association of International and European Affairs (ΟΔΕΘ) and as a Research Assistant and Research Administrator at the Hellenic Observatory Centre (LSE).