The artificial intelligence (AI) landscape is rapidly evolving, with significant implications for the energy industry. Growth in AI capabilities across industries is fueling operational efficiencies and advanced data analytics, but these improvements come at a cost. After a decade of flat demand growth, the Department of Energy estimates 5-6% load growth, driven in large part by data center power demand, which is expected to double from 2022 levels to 35 GW by 2030. Tech companies like Google, Microsoft, and Meta are developing data centers with connection requests for 300-1,000 MW or more, challenging utility resource plans and clean energy goals.

Recent months have seen a flurry of activity from federal and state governments, regulatory bodies, and industry leaders. Key topics and themes include grid planning, siting and permitting, grid operations and reliability, cybersecurity, and grid resilience. This blog explores some of the emerging initiatives to explore and regulate AI for the energy sector.

Federal Initiatives and Reports

In October 2023, the Biden Administration kicked off new interest in AI by releasing Executive Order 14110, creating a framework for the ethical and safe deployment of AI. In April of 2024, several tasks were fulfilled by the Department of Energy (DOE), including:

1. The “AI for Energy: Opportunities for a Modern Grid and Clean Energy Economy” report, which identifies AI applications in planning, permitting, operations, and reliability.
2. The “Advanced Research Directions in AI For Energy” report, outlining long-term opportunities and challenges for national laboratories.
3. A new website showcasing DOE-developed AI tools.
4. A new cybersecurity initiative to convene stakeholders on the risks and potential of AI, in addition to a report on the use of AI for securing critical energy infrastructure.
5. Other initiatives to tackle load growth, data center infrastructure, and interconnection challenges.

The DOE’s focus on AI has only just begun, especially as funding from the CHIPS and Science Act and Bipartisan Infrastructure Law continues to be allocated. In July 2024, the DOE launched an agency-wide effort via the Frontiers in Artificial Intelligence for Science, Security and Technology (FASST) program, focusing on advancing AI tools and workforce readiness. Shortly after launching the FASST program, the DOE released a report to the Secretary of Energy on “Recommendations on Powering AI and Data Center Infrastructure.” The report, developed by a broad group of stakeholders, recommended the creation of a DOE-run AI data center to study ways to lessen energy usage and continued stakeholder discussions to address supply bottlenecks, data sharing considerations, and new frameworks for orchestrating grid services. The Biden Administration’s 2024 Report on the Cybersecurity Posture of the United States also highlights growing interest in – and concern for – AI within the security space.

Regulatory Responses and State Initiatives

While many of DOE’s initiatives are broad, lawmakers have begun contemplating more concrete ways to safeguard this new technology. Both the Senate and the House have established bipartisan AI working groups, and in May the Senate “AI Gang” released a 31-page AI roadmap, calling for $32B per year in non-defense funding for AI research and development. Sen. Joe Manchin focused more keenly on AI energy issues via the bipartisan Department of Energy AI Act (S. 4664), aiming to expedite interconnection processes and study data center growth. Other legislators have requested information from the Federal Energy Regulatory Commission on how it is using AI to address growth in electricity demand from AI data centers.

State lawmakers are also stepping up to regulate AI:

1. Colorado: Colorado was the first state to enact comprehensive AI legislation earlier this year, seeking to regulate “high-risk” AI systems, including new reports, risk management plans, and other requirements.
2. California: California, home to 35 of the world’s top 50 AI companies, may be the next frontier for AI legislation, as dozens of bills were introduced this year. At the end of August, California lawmakers passed a law to the Governor that would require the largest (>$100M) AI systems to test their models and publicly disclose Safety and Security Plans.
3. Texas: Lawmakers are preparing new AI oversight initiatives for the January 2025 legislative session, following AI-related hearings and the establishment of the Texas AI Advisory Council.

Looking Ahead: Challenges and Opportunities

The rapid expansion of AI and data centers presents both challenges and opportunities. As the energy demands from AI continue to rise, the need for innovative solutions to manage this growth becomes critical. Utilities will be at the forefront of managing the AI revolution, especially as AI technology enhances customer service, outage detection, automatic diagnostics, orchestration of distributed energy, and other use cases whilst concurrently driving dramatic demand growth. While many investigatory reports and initiatives have been launched across the U.S. government, the adoption of true regulation for AI is, so far, minimal. Nonetheless, we can expect that interest in, and concern about, AI will continue to grow in the energy space.