3 insights on the role of AI in grid transformation
Understanding and addressing rapid power demand growth forecasts within the context of the clean energy transition will require collaboration from federal, state and local, private sector, and utility leaders. Looking ahead to 2025 and beyond, leaders are facing challenges of planning for and managing load growth from data centers, artificial intelligence, and electrification, while also meeting established decarbonization goals and managing energy costs (particularly for low-income customers).
While AI is contributing to these challenges, what’s the potential role of it in the solutions?
ICF convened a roundtable discussion on the role of AI in promoting a clean, reliable, and resilient electric grid, featuring DOE’s Chief AI Officer Helena Fu and Senior Advisor Keith Benes. The roundtable also featured leaders in the D.C. region from the federal government, state and local governments, utilities, and the private sector. Here are three key takeaways from this important discussion.
1. Rising power demand is creating significant challenges for utilities and meeting decarbonization goals.
ICF’s latest report projects a sudden surge of U.S. electricity demand after two decades of relatively flat growth. The report leverages EnergyInsite, ICF’s cloud-based renewable energy analytics platform.
U.S.-wide average annual electricity demand is projected to rise 9% by 2028 and 18% by 2033 compared to 2024 levels. By 2050, demand is projected to increase by 57% compared to 2024.
Rising power demand is also occurring at the same time as states and utilities are working to decarbonize and make energy more affordable. This creates a challenge for how to meet the need for increased power generation to satisfy AI and data centers while transitioning away from fossil fuel generation.
In addition, this is occurring at a time when utilities are challenged by the time it takes to build and connect new clean energy sources—particularly with infrastructure siting and permitting. It can be difficult for planners to identify the ideal locations for clean energy projects, and while queues for interconnections are starting to advance, delays in permitting can increase the time it takes to bring clean energy projects online. There’s also the issue of shrinking electricity reserves, which impacts reliability and is contributing to rising electricity prices.
These factors may mean that legacy fossil assets remain in use for a longer duration than previously expected, which could delay progress towards decarbonization and clean energy goals.
2. Increasing power demand also presents unique opportunities for the clean energy transition.
Despite the concerns, data centers and AI also present a key opportunity for state and local governments, utilities, and the private sector. Hosting data centers and AI training within the U.S. is seen as a competitive advantage, and data centers can create unique opportunities for the clean energy transition.
Rising power demand and clean energy goals also leave room for and will accelerate innovation, such as Dominion Energy’s and Amazon’s partnership to explore small modular nuclear reactors in Virginia. Microsoft, Google, and others are also making significant investments in nuclear given that it’s a zero emissions, reliable energy source.
Another potential opportunity exists in reusing heat generated by data center operations. Leaders are discussing how to take excess heat generated by data centers and use it to benefit a surrounding community, such as for heating a school or community center.
3. AI has the potential to help address challenges from rising power demand.
AI adoption can be a key tool to help leaders navigate these challenges. AI is already being used across the energy industry in the development of new technologies, to advance research, and improve existing processes. Examples include using AI to identify new materials for long duration energy storage and using AI to speed up the NEPA environmental permitting process. Leaders are also starting to consider how to use AI to improve grants management processes.
In addition, utilities are using AI for some preventative work, such as helping prevent wildfires and survey transmission lines to prevent outages. Utilities are also using AI to aid with dynamic line rating, improve customer outreach and program engagement, and meet building decarbonization goals. Some utilities are implementing digital twins to help with planning around Distributed Energy Resources (DERs), like solar panels and EV charging stations within the grid, and to reduce energy costs for low-income communities.
DOE’s proposed FASST (Frontiers in Artificial Intelligence for Science, Security and Technology) initiative is also focused in part on solutions to address energy demand challenges. FAAST’s goal is to speed up the U.S. government’s AI capabilities and create an integrated, scientific AI system. The proposed initiative calls for leveraging DOE’s existing infrastructure to address energy challenges such as unlocking new clean energy sources, optimizing energy production, and improving resilience of the electric grid.
These actions and initiatives are accelerating at a critical time to address rising power demand, while attempting to meet essential clean energy goals. While data center demand and the need for decarbonization pose hurdles, AI's potential to optimize energy production, improve grid resilience, and support clean energy initiatives offers a promising path forward.
