Companies that strategically invest in AI development and integration will likely see long-term benefits
January 29, 2025
Artificial intelligence (AI) is everywhere in the news, and companies worldwide are rushing to capitalize on its opportunities. At the other end of the spectrum, utility companies are capital-intensive organizations with low-profitability, long-term returns and risk profiles. While constantly innovating, they provide a critical service to the economy and favor stability over disruption for obvious reasons. The introduction of a paradigm-changing technology like AI is thus orthogonal to their traditional approach to doing business.
The core question for these companies is how AI will affect them and whether there are use cases that can be profitable while preserving operational safety and continuity. A thorough analysis of the annual reports of the ten largest listed utility companies worldwide and the ten largest U.S. utilities reveals some interesting insights.
A first-level approach, based on the number of references to artificial intelligence in the latest annual reports available (2023), highlights a geographic discrepancy between European utilities and the rest of the world. EDF (France) and Iberdrola (Spain), notably, mention the technology more frequently than their peers. Asian and U.S. counterparts barely mention it. This is highly noteworthy as the reporting dates of these annual reports are December 31, 2023, which was quite early after the introduction of groundbreaking innovations such as ChatGPT and quite late given the rapid innovation cycle surrounding these technologies.
Focusing on the U.S., where the companies are geographically and culturally closest to the AI innovation hub, one can only be surprised by the lack of reference to AI in 2023 annual reports. Only the Tennessee Valley Authority provides substantial information on the topic. This does not imply that U.S. companies are not deploying AI solutions; it only suggests that their AI-related efforts have not reached a maturity level significant enough to be disclosed to investors. At this stage of the technology’s deployment, it remains too early to determine whether these communication differences reflect actual disparities in adoption or are simply indicative of differing communication strategies. However, the discrepancy clearly signals varying levels of bullishness on the topic to investors and competitors alike.
For companies in the sample that do discuss AI, many references focus on potential risks rather than opportunities. Seven out of fourteen companies that mentioned AI identified it as a cybersecurity threat. Utilities are particularly sensitive to these issues as they operate critical infrastructures and manage significant volumes of personal data through their distribution and retailing operations. The reported risks are often non-specific but include using AI to assist in hacking or in social engineering. These concerns underscore the need for vigilance as AI technology evolves.
On the positive side, AI use cases are rarely discussed in detail. A small number of companies—three in total—highlight the potential for AI to become a competitive advantage and its possible influence on market conditions through cost reductions. Four companies report that AI solutions are being deployed or are under consideration to improve operational and maintenance (O&M) performance. These statements raise the question of what is meant by “AI,” as the term encompasses a broad array of technologies. In the O&M context, well-established use cases include using machine learning to predict faults and implement preventive maintenance measures, as well as leveraging large datasets of weather, operational, and market data to optimize the production of renewable energy sources. While these technologies share similarities with the Large Language Models currently in the spotlight, their capabilities and purposes differ significantly.
A notable trend is that AI initiatives are often conducted at the subsidiary level rather than at the parent company. This is particularly evident in specialized subsidiaries that operate in smart building or smart city markets. For example, some companies are testing AI-driven solutions in subsidiaries established through acquisitions of startups or investments in incubation programs such as EDF Pulse. Intrapreneurial programs can also be used to foster innovation (for example with Yxir, an internal start-up of EDF, focused on industrial quality management through the use of AI). This approach allows utilities to experiment with innovative solutions on a smaller scale without exposing their core business to undue risks. Moreover, it creates an option value: successful solutions can be scaled and replicated across the organization.
In their current form, annual reports only disclose materially significant aspects of operations. AI-based solutions, many of which are still in the early stages, do not yet meet this threshold for disclosure. This does not imply a lack of activity; on the contrary, utilities are likely experimenting with AI across various segments of the value chain. The true extent of these efforts remains hidden but promises to become more apparent as the technology matures.
AI is poised to play a transformative role in the utility sector. Companies that strategically invest in its development and integration will likely see long-term benefits, ranging from enhanced efficiency to improved sustainability. The sector’s inherent conservatism, while a constraint in the short term, may enable a more measured and ultimately successful adoption of AI. As the industry evolves, the 2024 annual reports, expected in early 2025, will provide valuable insights into how utilities worldwide are positioning themselves within this technological paradigm shift. The next few years will reveal whether AI will become a cornerstone of utility operations or remain a niche technology.
You May Also Like
