The rapid development of Artificial Intelligence (AI)[1] is transforming our daily lives and impacting our economies, including the energy sector. On the one hand, data centres hosting the computing power and databases, as well as the development, training and running of AI systems, have become a major source of additional electricity demand and congestion at the local level. On the other hand, AI might support new tools to improve the operation and planning of the electricity system, increasing its efficiency and reliability.

Several questions arise from the emerging context of the AI-Energy Nexus. How do AI systems work, and what are their current and potential applications in the electricity system? Which type of energy data is required to train and develop such AI systems? Can energy data be reused beyond its original operational or transactional purpose to train AI systems? What is the applicable legal framework for a trusted, transparent and secure energy data exchange?

In this Topic of the Month, we dive deep into these questions to unpack the relationship between Artificial Intelligence and the EU Electricity System.

In the first instalment, we situate the discussion within the broader EU policy agenda, focusing on the Strategic Roadmap for Digitalisation and AI in the Energy Sector, recently adopted as part of the European Tech Sovereignty Package. We proceed in three steps: first, we introduce the AI-Energy Nexus; second, we explore the concepts of strategic autonomy and technological sovereignty; third, we map the Roadmap’s Pillars and Flagship Actions.

The AI-Energy Nexus

AI is increasingly considered as a potential enabler of a more efficient, flexible and decarbonised electricity system.[2] Its potential applications span multiple layers of the system. In grid management, AI can support real-time monitoring, predictive maintenance, and grid balancing. On the demand side, it enables smart energy services, dynamic pricing, and demand-side flexibility. It also plays a growing role in improving forecasting and facilitating the integration of variable renewable energy sources.[3] According to SmartEN and DNV’s 2022 prediction, digitalisation of the EU’s energy systems could deliver €71 billion per year in direct consumer savings and more than €300 billion in wider system benefits.[4]

These benefits, however, depend on the availability, accessibility and quality of data. Energy-specific AI systems require large volumes of heterogeneous data, which must typically be accessible, standardised and interoperable. This data includes electricity grid operational data, wholesale market data, smart metering and consumption data, but also non-energy data such as weather and climate information. Training AI systems for energy applications requires large-scale, high-quality data that are pooled, representative, and statistically robust – alongside a trusted and secure environment for data sharing.

At the same time, AI itself is energy-intensive, as the computational intensity of AI systems’ training and deployment, particularly in large-scale data centres, translates into significant electricity demand. This creates a fundamental dilemma: AI is used to improve the functioning of the energy system, while simultaneously placing additional strain on it. This dual relationship is commonly described as the ‘AI–energy nexus’.[5]

Strategic Autonomy and Technological Sovereignty

The development of AI is embedded in highly globalised supply chains. AI key components include the software and hardware level, encompassing the data infrastructures to store and process data and the compute infrastructure that powers AI systems training. These components are creating new forms of structural dependency, particularly around the control of hardware production, software design, standards setting, and data-flow infrastructures management. What began as a private-sector innovation has evolved into a key factor shaping global economic competition and geopolitical power dynamics, where the US and China are by far the dominant players.[6] Currently, non-EU companies control most of the critical layers of the European digital stack, holding the intellectual property of ‘choke points’ for operating systems, cloud platforms, chip architectures and machine learning frameworks.[7]

The picture becomes even more complex if we consider the critical nature of energy infrastructure. While the integration of digital technologies such as AI can enhance system performance and efficiency, it also introduces new risks, including safety concerns, hybrid threats, and cybersecurity vulnerabilities.[8]

It is in this context that the concepts of strategic autonomy and technological (or digital) sovereignty have gained prominence in the EU policy discourse. The concept of EU strategic autonomy originated in the field of security and defence[9] but has since expanded to encompass broader domains, including energy and digital policies.[10] It generally refers to the EU’s capacity to act independently in strategically important areas, reducing reliance on third countries for critical goods and services, but it does not imply autarky. Closely related (and slightly overlapping) is the concept of technological sovereignty, which emphasises the ability of a State or group of States to develop or access key technologies necessary for its welfare, competitiveness, and political agency, without becoming subject to one-sided structural dependencies. [11]

The Strategic Roadmap for Digitalisation and AI in the Energy Sector

Strategic autonomy and technological sovereignty are framed as “mutually reinforcing goals” within the European Tech Sovereignty Package, published by the European Commission on 3 June 2026. The package consists of the Strategic Roadmap for Digitalisation and AI in the Energy Sector, the EU Open Source Strategy, and two legislative proposals, the Chips Act 2.0 and the Cloud and AI Development Act.

The Strategic Roadmap builds on a set of initiatives and measures adopted in the digital and energy areas and is structured around three core Pillars: Pillar I focuses on the sustainable integration of data centres into the energy system; Pillar II outlines measures to deploy digital and AI solutions across the energy system; Pillar III addresses the governance framework for energy data, enabling smart energy services and the scaling of AI applications. These pillars are complemented by a cross-cutting section on securing the AI–energy nexus and a concluding section on implementation.

Overall, the Roadmap sets out a vision for a digitalised energy system in which AI plays a central role in delivering secure, clean, and competitive energy for all consumers. It also identifies seven flagship actions aligned with its pillars and strategic objectives. When the European Commission launched the Strategic Roadmap on 3 June 2026, it simultaneously introduced two initiatives linked to Flagship Action 1 and Action 4. These included the signature of a declaration of intent by industry associations to collaborate, under the Commission’s guidance, on the sustainable integration of data centres into the energy system, and the establishment of a Community of Practice for the development of AI models supporting grid management and planning.

Acknowledgements

The Florence School of Regulation gratefully acknowledges the financial support of the European Commission (DG ENER) for conducting the research that led to this blogpost. Views expressed in this blogpost reflect the opinion of individual author(s) and do not necessarily reflect the views of the European Commission.

[1] According to  Article 3(1) of the AI Act, ‘AI system’ means a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.

[2] For an analysis of the benefits and costs of digitalisation and use of AI in the energy sector, see Directorate-General for Energy (European Commission) and others, Support for the Preparations of a Strategic Roadmap for Digitalisation and AI in the Energy Sector (Publications Office of the European Union 2026) 19.

[3] These examples have been indicated as the three main areas which could benefit the most from digital and AI technologies by the respondents to the Open Public Consultation launched by the European Commission to prepare the Strategic Roadmap for Digitalisation and Ai in the energy sector: Directorate-General for Energy (European Commission), Spoden, Anna and Arrowsmith, Greg, ENTEC2 OPC Analysis (Publications Office of the European Union 2026).

[4] SmartEN and DNV, ‘Demand-Side Flexibility – Quantification in the EU’ (2022).

[5] IEA, ‘Energy and AI. World Energy Outlook Special Report’ (2025) .

[6] See Schettini Claudia, ‘IA di Stato: dall’algoritmo alla sovranità del calcolo’ (Istituto per gli Studi di Politica Internazionale (ISPI), 2026); Federica Marconi, ‘Reframing Open Strategic Autonomy in the EU Digital Ecosystem’ (Istituto Affari Internazionali (IAI), 8 June 2026).

[7] Vaida Gineikyte-Kanclere, Militsa Eggert and Goda Skiotyte, ‘European Software and Cyber Dependencies. Study Requested by the ITRE Committee’ (2025) 42.

[8] ibid 110.

[9] For a critical assessment on the rhetoric around these concepts in the political discourse and on the way they migrated from the defence sector to other domains, see Raluca Csernatoni, ‘The EU’s Hegemonic Imaginaries: From European Strategic Autonomy in Defence to Technological Sovereignty’ (2022) 31 European Security 395.

[10] For a discussion of the concept of strategic energy autonomy in the EU, see Leigh Hancher and Adrien de Hauteclocque, ‘Strategic Autonomy, REPowerEU And The Internal Energy Market: Untying The Gordian Knot’ (2024) 61 Common Market Law Review.

[11] Directorate-General for Research and Innovation (European Commission) and Kroll, Henning, New Challenges of Technological Sovereignty (Publications Office of the European Union 2026).