21 October 2025

The European Internet Forum and the European Energy Forum held a joint debate on the interlinkages between artificial intelligence and energy systems. The discussion explored how AI technologies can support the clean energy transition while addressing their own growing electricity needs. Bringing together policymakers, industry leaders, and experts, the debate examined the opportunities and challenges at the intersection of Europe’s digital and green transformations. Speakers reflected on the policy frameworks, investment priorities, and standards required to ensure that AI contributes to energy efficiency, competitiveness, and sustainability across the Union.

From Algorithms to Energy: How AI and the Sustainable Transition Drive Each Other Forward

Opening Remarks – Tsvetelina Penkova MEP and Brando Benifei MEP

Tsvetelina Penkova MEP opened the discussion by underlining the growing interdependence between artificial intelligence and energy, noting that AI has become an invisible yet transformative force in everyday life, shaping business models, industrial decisions, and investment priorities. She stressed that this rapid technological progress brings not only efficiency gains but also rising energy demand. A single large-scale AI model, she noted, can consume as much electricity as one hundred European households in a year. MEP Penkova called for a balanced policy approach that recognises AI both as a driver of higher energy consumption and as a key enabler of system optimisation - from forecasting renewable generation and managing smart grids to improving industrial and building efficiency. She framed the debate as an opportunity to explore how Europe can leverage AI for a more sustainable and efficient energy future.

Brando Benifei MEP emphasised that “there is no AI without energy, and no energy transition without AI.” He pointed to the forthcoming EU strategic roadmap on digitalisation and AI in energy, expected early next year, which will align data access, grid efficiency, and governance standards with the AI Act, the NIS 2 Directive, and the Cyber Resilience Act. He highlighted the need to measure and reduce AI’s environmental footprint while fostering innovation and competitiveness. MEP Benifei called for stronger public-private cooperation to improve grid services, encourage efficient data-centre siting, and promote the reuse of waste heat. He also urged the launch of an “AI for Efficiency” initiative at EU level to deploy AI across factories and public buildings, achieving measurable energy savings.

The Commission’s Perspective

Vincent Berrutto, Head of Unit for Energy Efficiency at DG ENER, noted that digital and energy transitions are increasingly inseparable. He outlined the Commission’s upcoming Strategic Roadmap on Digitalisation and AI in the Energy Sector, to be presented in early 2025, which will seek to unlock the benefits of digital technologies while addressing energy-consumption challenges. He stressed the urgency of tackling the rising electricity demand of data centres, which already represent 2–3 percent of EU consumption, with higher local shares in certain Member States. The roadmap will prioritise the creation of a European energy data space, greater investment in smart grids and digital twins, and the development of foundation models tailored to the energy sector. Mr Berrutto also announced a forthcoming Data Centre Energy Efficiency Package, including a common rating scheme and measures to improve grid access. He underlined that these initiatives complement the Affordable Energy Action Plan and other ongoing Commission strategies aimed at delivering cleaner, more resilient, and competitive energy systems across Europe.

Industry Perspectives

Mallika Ishwaran, Chief Economist at Shell, argued that AI could become as transformative for global productivity as earlier industrial or digital revolutions, provided that resource efficiency and cost-effectiveness improve. She warned that the technology’s growing electricity need, potentially a fivefold increase by 2050, must be balanced with its long-term potential to reduce emissions through better integration of renewables, electrification, and system optimisation. She emphasised that AI’s benefits will depend on regional readiness, skills, and governance capacity, with advanced economies likely to benefit first before gains spread globally.

Josh Parker of NVIDIA highlighted that AI’s current share of total electricity use remains relatively small but is expanding rapidly. He reported that advances in hardware have delivered dramatic energy-efficiency gains - up to thirtyfold improvements within two years - and stressed that performance per watt has become a key industry metric. He described examples of green data centres powered by renewables and integrated into local heating networks, illustrating how innovation can align digital growth with sustainability. Mr Parker argued that broad deployment of AI could yield a net positive impact on emissions, supporting grid modernisation, demand response, and efficiency across manufacturing, buildings, and transport.

Expert Perspectives

Thomas Spencer from the International Energy Agency presented IEA analysis showing that global data-centre electricity use could double by 2030, reaching roughly 3 percent of global consumption. He underlined the local grid challenges posed by data-centre concentration and the need for smarter, more flexible European networks. Europe, he noted, remains a key actor in the data-centre value chain, with significant industrial capabilities in cooling, transmission, and automation technologies. Mr Spencer also highlighted AI’s potential to accelerate innovation in energy technologies, from advanced materials to hydrogen production, and to enhance industrial competitiveness through digital twins and process optimisation.

Dr. Arti Garg, Chair of the IEEE Working Group on the Environmental Impacts of AI, focused on the importance of establishing consistent global measurement standards for AI’s energy and environmental footprint. She explained that the complexity of AI life cycles - from hardware manufacturing to data storage and model training - makes quantification difficult, yet essential for sound governance. Dr Garg noted that the IEEE initiative aims to develop transparent, repeatable metrics to guide policymakers and industry towards more sustainable AI deployment.

Videos

  • Online piracy of live sports and other live events
  • Fake science
  • From Algorithms to Energy: How AI and the Sustainable Transition Drive Each Other Forward
View more

Related content

European Internet Forum ©2025. All Rights Reserved.