Leveraging AI for Energy Security: A Transatlantic Perspective
Larissa Iannella Oliveira
Introduction
Artificial Intelligence (AI) is transforming the global energy sector on both sides of the Atlantic. In the wake of Russia’s 2022 invasion of Ukraine, the strategic risks of overreliance on a single energy supplier became clear, with Europe's dependence on Russian gas leading to a severe crisis that NATO classified as a "threat to allied resilience.” At the same time, the urgency of decarbonization has never been greater, as the United States and European Union (EU) strive to meet ambitious climate targets while maintaining energy security. These dual challenges shows the need for innovative solutions that optimize both fossil fuel-based and renewable energy systems.
AI-driven technologies are emerging as critical tools in this transition. By leveraging machine learning, predictive analytics, and automation, AI optimizes energy consumption, strengthens grid resilience, and enhances supply chain security. Recognizing these opportunities, transatlantic leaders are calling for stronger U.S.-EU cooperation to adopt AI for energy security.
This paper examines AI’s role in enhancing energy security, from traditional legacy energy to renewable and fusion technologies. It explores how AI is being deployed to optimize resource management, enhance grid stability, and improve predictive maintenance, by using key case studies from both sides of the Atlantic. Finally, the paper argues that deepened U.S.-EU cooperation in AI governance and energy innovation is essential to ensuring a secure, sustainable, and technologically competitive energy future.
I. AI’s Role in Energy Security
Optimizing Legacy Energy Systems (Coal, Gas, Nuclear)
When discussing energy security, legacy energy sources (including coal, natural gas, oil, and nuclear power) have long played a crucial role and are expected to remain significant in the near future. AI-driven technologies are increasingly being used to improve the efficiency, reliability, and sustainability of these systems.
First, AI optimizes oil and gas exploration by enhancing reservoir identification and management. For example, the Italian energy company ENI S.p.A. invested in the HPC5 supercomputer in 2020—at the time, the most powerful industrial supercomputer in the world—to improve the efficiency and accuracy of resource discovery. HPC5 played a crucial role in Egypt’s energy security by enabling the discovery of the Zohr gas field, the largest in the Mediterranean, in an area that had been previously explored three times unsuccessfully. So by leveraging AI-powered digital twins of geological strata, ENI was able to analyze subsurface patterns with unprecedented precision, leading to a breakthrough in resource identification. Additionally, this added precision, which reduces leaks and pressure spikes, also aligns with broader sustainability goals by making oil and gas exploration safer, more cost-effective, and less environmentally disruptive.
Second, in terms of coal-fired power, AI has also been instrumental to optimize efficiency and reducing emissions. A notable example is Vistra Corp.'s Martin Lake Power Plant in Texas, where the company partnered with McKinsey & Company to implement a machine-learning model that analyzed two years' worth of plant data. The AI system (developed with QuantumBlack AI) identified optimal “heat-rate conditions” by assessing both external factors (such as temperature and humidity) and internal plant operations. Operators then received real-time efficiency recommendations every 30 minutes, leading to a 2% improvement in fuel efficiency, saving $4.5 million annually, and cutting 340,000 tons of CO₂ emissions. This shows how much AI can extend the viability of coal power by making it cleaner and more cost-effective.
Third, AI holds significant potential in enhancing nuclear energy production. Similar to other legacy energies, machine learning softwares are increasingly being used for predictive maintenance of nuclear power plants (NPPs). Dr Jeremy Renshaw, senior technical executive at the Electronic Power Research Institute (EPRI), illustrates this by saying that incorporating advanced machine learning systems allows for inspectors to only look at the relevant data: “Instead of searching for the ‘needle in the haystack’, we remove the haystack.” This leads to an increased precision and a reduction in time and costs and an overall increase in efficiency for nuclear power plants operations. However, in terms of decision-making, AI (specifically generative AI), is still far from operating NPPs. This is due mainly to the “black box issue” in neural networks, where the methods by which artificial networks reach specific conclusions are not fully understood. Given the lack of transparency in how these models arrive at their conclusions, it is reasonable that they are not yet responsible for managing NPPs. However, as our understanding of AI systems grows and transparency improves, it is safe to anticipate that generative AI will have broader applications in decision-making for NPPs in the future.
In sum, AI’s role is promising as it optimizes performance and reliability of traditional legacy energy facilities. Its predictive capabilities help ensure that sources stay available when needed and with fewer disruptions. These optimizations also reduce the carbon footprint of legacy energy by allowing less of it to go to waste, which aligns with long-term decarbonization goals.
Optimizing Renewable Energy
Over the past decade, renewable energy investments have grown exponentially, with a global spending of $500 billion in 2023, largely driven by solar and wind. In the U.S., the Inflation Reduction Act (2022) allocated $369 billion for clean energy investments, and in 2022 the EU launched the REPowerEU initiative to accelerate renewable energy deployment and reduce reliance on Russian gas (a dependence that had contributed to the European gas crisis following Russia’s invasion of Ukraine). Thanks to this initiative solar capacity more than doubled from 2019, and for the first time, in 2022 the EU was able to produce more electricity from wind and solar than from gas. As energy demands rise, driven in part by AI advancements and the continuous power requirements of data centers, legacy energy sources alone will not be enough to meet those demands, and renewable energies have to be part of the energy security conversation.
Unlike traditional legacy energy, solar and wind outputs fluctuate with weather and time of day, making stable energy generation a challenge. One of the first applications of machine learning in renewables is predictive analytics, which improves forecasting and resource allocation to optimize supply and demand. In 2019, Google’s DeepMind developed a deep neural network trained on weather forecasts and historical turbine performance data to predict wind farm output 36 hours in advance. By using reinforcement learning and probabilistic modeling, the system was able to optimize power dispatch and delivery scheduling. Tests on Google’s 700 MW wind farms showed that this approach increased the overall economic value of wind energy by about 20%. Advanced techniques such as recurrent neural networks (RNNs) and ensemble learning are now widely used in energy forecasting, addressing what is the biggest challenge of renewables: their variability and unpredictability.
Beyond forecasting, AI plays a critical role in integrating renewable energy into the grid by optimizing storage and balancing demand. In high-renewable energy systems, excess generation—such as surplus solar power at midday—needs to be stored and strategically released during periods of lower production, such as nighttime or wind lulls. Advanced AI enables real-time grid management, adjusting wind turbine blade angles and solar panel tilt based on weather conditions to maximize the energy output. The European Commission’s DG CONNECT and DG ENER workshop (September 2024) highlighted AI’s transformative role in optimizing renewable integration and strengthening grid resilience. Experts emphasized that AI-powered solutions can shorten grid connection times, attract investment, and lower energy costs, all while supporting the EU’s climate neutrality goals.
AI’s Role in Fusion Energy
Nuclear fusion has recently become a focal point in energy discussions due to major scientific breakthroughs and its potential as a limitless, clean power source. The link between fusion and national security is also critical, as fusion is a strategic domain where the US, the EU, and China are competing to establish technological leadership in this field. Unlike solar and wind, which depend on land and weather, fusion offers a continuous, self-sustaining energy source that could reshape the global energy landscape.
AI is fundamentally reshaping fusion energy research, accelerating breakthroughs that were once projected for 2060 to potential feasibility in the early 2030s.[1] One of the biggest barriers to achieving commercial fusion has been the challenge of managing plasma at 100 million degrees Celsius, where materials must withstand extreme conditions without degrading.[1] AI is now addressing this through advanced modeling, computational simulations, and materials discovery, significantly shortening the path to viable fusion reactors. Machine learning models are revolutionizing plasma heating simulations by predicting and optimizing plasma. A breakthrough at Princeton Plasma Physics Laboratory (PPPL) demonstrated AI’s ability to identify and correct errors in numerical codes and predicting plasma heating behavior where previous models had failed. This advancement enhances reactor stability and accelerates the development of next-generation fusion reactors.
As AI accelerates fusion research and reshapes the broader energy landscape, its role extends beyond technological advancements, it is a strategic asset for U.S.-EU cooperation. Given the geopolitical stakes of energy security, AI-driven energy innovation presents a key opportunity for enhanced transatlantic cooperation in securing and optimizing energy systems.
II. Transatlantic Cooperation on AI and Energy Security
NATO’s Role in Energy Security
NATO’s Energy Security agenda, first outlined at the 2008 Bucharest Summit, has been strengthened through the NATO Energy Security Centre of Excellence (established 2012 in Lithuania). The latest Strategic Foresight Analysis (SFA23) by NATO’s Allied Command Transformation (ACT) identifies energy security as a critical challenge shaped by resource scarcity, climate change, and growing competition over energy access. The report also emphasized the need for NATO to enhance resilience by reducing dependence on traditional energy sources and investing in technological innovation.
While AI is not NATO’s primary focus, its broader mandate on cyber defense and emerging technologies enables AI-driven solutions for energy security. NATO’s Defense Innovation Accelerator for the North Atlantic (DIANA) is a key initiative in this effort, supporting transatlantic AI cooperation in critical infrastructure protection. DIANA’s mission is to develop dual-use AI technologies that enhance grid resilience, threat detection, and energy system security. So by integrating AI-powered cybersecurity tools, predictive analytics, and energy infrastructure monitoring, NATO has been playing a critical role in securing AI-driven energy networks across the Alliance.
U.S.-EU Collaborative Frameworks and Initiatives
The U.S.-EU Trade and Technology Council (TTC) (2021–2024) was a platform created to aligning US and EU policies on AI and trade issues. One of the major successes of the TTC was the Joint Roadmap for Trustworthy AI and Risk Management (December 2022), which set shared guidelines for safe AI deployment in critical sectors, including power grids and cybersecurity. As Frances Burwell notes in a 2024 Atlantic Council report, the TTC’s success reflects a broader shift in transatlantic digital policy, where the EU has led with regulatory frameworks, while the U.S. has focused on voluntary standards and innovation incentives. She highlights that while regulatory divergence remains, both sides recognize that AI-driven energy security requires coordinated governance to prevent market fragmentation and ensure secure AI deployment.
In February 2025, the AI Action Summit in Paris gathered several global leaders to discuss AI governance, ethics, and innovation. A key outcome was the "Inclusive and Sustainable AI for People and the Planet" declaration, signed by 58 countries, emphasizing transparency, accessibility, and ethical AI development. However, the U.S. and U.K. declined to sign, citing “concerns over global AI governance and security.”
However, the summit showed different approaches to AI governance between the US and the EU, especially in terms of regulation. European Commission President Ursula von der Leyen announced the EU's commitment to mobilize €200 billion for AI investments in Europe, but also stated that, "AI needs competition, but AI also needs collaboration, and AI needs the confidence of the people, and has to be safe." U.S. Vice-President JD Vance took an opposing approach, and voiced that overly strict regulations would hinder AI advancements which would unjustly limit American technology firms, and then stated that, "The US cannot and will not accept that.”[1] While differences in regulatory approaches persist, AI remains a crucial tool for transatlantic cooperation on energy security, requiring ongoing collaboration to align policies, enhance resilience, and ensure that technological advancements serve both economic and strategic interests on both sides of the Atlantic.
Conclusion
As AI continues to reshape the energy landscape, its role in enhancing energy security has become a key focus for transatlantic cooperation. AI-driven technologies are already optimizing fossil fuel exploration, improving nuclear plant efficiency, and making renewable energy integration more reliable. Advances in machine learning and predictive analytics are not only increasing energy system efficiency but also reducing greenhouse gas emissions, supporting both security and sustainability goals. The strategic importance of AI in energy security extends beyond technological optimization; it represents a vital geopolitical tool that strengthens resilience in an era of growing energy competition and climate-driven disruptions
References
[1] Marco Margheri, ENI SpA Adoption of AI, interview by Larissa Oliveira, February 7, 2025.
