Why Data Readiness is One of Utilities’ Most Urgent AI Challenges

Realizing AI potential starts with getting data right.

By Chris Warren

In a relatively short period of time, artificial intelligence (AI) has become the centerpiece of discussions across most industries, including the utility industry. Much of the recent focus has been on utilities’ ability to meet the growing demand for electricity coming from data centers powering AI applications. In August 2025, for example, EPRI and Epoch AI released a report forecasting that the total AI power demand in the U.S. could grow from 5 GW at the time of publication to more than 50 GW by 2030.

Meeting the electricity demand from these large new customers is understandably a top utility priority. But utilities are equally motivated to leverage AI to improve everything from their operations and maintenance (O&M) to their planning and investment decisions. In March of 2025, EPRI joined with national labs, universities, utilities such as Duke Energy and Exelon, and technology firms such as NVIDIA, AWS, and Microsoft to launch the Open Power AI Consortium.

“The mission of the Open Power AI Consortium is to transform the electric sector by leveraging AI, as appropriate, to improve how electricity is made, moved, and used,” said Jeremy Renshaw, director of the Open Power AI Consortium. “We need to utilize the members of the consortium to think big and push the boundaries of what is possible as well as help to inform regulators and policymakers by bringing them closer to the development process.”

The Foundation of AI

There is a dizzying array of already feasible and emerging applications of AI in the power sector. EPRI is pursuing an ambitious AI research agenda, exploring everything from how it can improve wildfire detection, grid management, predictive maintenance, and cybersecurity across hundreds of use cases in the utility industry.

It is already clear that AI’s ability to help prevent outages, accelerate wildfire response, and manage an increasingly complex grid comes down to data.

A survey of Open Power AI Consortium members found that data readiness, or lack of it, is a primary obstacle in scaling AI.

The report AI Readiness in Utilities: Turning Data into Strategic Advantage was released in August 2025 and validates the critical importance of data readiness in transforming AI potential into value, and it provides a five-point framework to achieve it. Developed with input from the utilities Constellation Energy and PPL and the consulting companies Accenture, PwC, and EY, the report defines data readiness as a scalable data architecture that provides clean, structured, and governed data, enabling advanced AI models to achieve specific business outcomes.

Put simply, data readiness is foundational to tap the potential of AI to improve grid resilience, customer engagement, forecasting, and to manage and orchestrate an increasingly distributed grid. “When our clients ask for help with their AI strategy, our first question is: what is your data strategy?” said Mike Juchno, a principal and consulting partner in EY’s data and AI energy sector practice.

Currently, most utilities do not yet have the level of data readiness needed to fully realize AI’s value. Common challenges include limited data accessibility across the enterprise, unresolved privacy and security concerns, and inconsistent data quality.

“It’s well-known in the industry that many utility data sources, including GIS (geographical information systems), are not 100 percent accurate,” said Matt Wakefield, EPRI’s director of Information Communication and Cyber Security (ICCS) and one of the co-authors of the report. “The real question is whether AI can still support sound business decisions when data is incomplete or inconsistent. In our experience, it can but only happen when utilities understand those limitations and design AI use cases accordingly.”

A Roadmap to Data Readiness

Collaboration is an AI Catalyst

Ultimately, the steps necessary to become AI data-ready must take place within utilities. But collaboration among utilities, technology providers, academics, national laboratories, and others can accelerate progress for everyone.

The Open Power AI Consortium provides many opportunities for collaboration, including:

• Launching pilot projects and co-innovation via AI sandboxes
• Gathering input from industry surveys on the most valuable use cases to undertake. Utilities and technology providers can share their insights in one of the two following survey links:
• Utility survey: https://www.surveymonkey.com/r/OPAI-Use-Cases-Utilities
• Technology provider survey: https://www.surveymonkey.com/r/OPAI-USe-Cases-Tech-Providers
• Establishing working groups on specific topics. To date, four working groups have been launched (see www.openpowerai.org for more information on how to engage)
• Applying shared data standards and interoperability frameworks for data exchange across systems
• Building industry-specific datasets, knowledge graphs, and foundation models
• Investing in AI literacy and workforce reskilling

EPRI has already worked with consortium members to identify open-source datasets and AI models, and has published white papers on how to prepare data to be AI-ready; these are all available on the Open Power AI Consortium website. In addition, EPRI surveyed consortium members and presented findings about over 250 AI use cases across the power sector.

Future consortium goals include expanding its global membership, developing new iterations of power industry-specific AI models, and evaluating high-priority use cases identified in the survey. “We plan to scale deployments, host global and regional AI workshops, formalize partnerships with academia and national labs, and maintain AI sandboxes with technology companies for collaborative experimentation and co-innovation,” Renshaw said. “These efforts collectively aim to establish the consortium as the global hub for AI in the power sector.”

---