US power grid infrastructure is straining to keep up with AI energy demands. Maximizing existing power grid capacity is critical to maintaining reliability and supporting long-term modernization.
As the United States races to expand its artificial intelligence (AI) capabilities and accelerate electrification, the nation faces a significant challenge: ensuring its aging power grid can keep pace with rapidly escalating energy demands. To respond to this challenge, incentives are needed to encourage infrastructure upgrades and support power capacity enhancement in the long run. Any viable roadmap takes time, and this means that for now, policymakers and industry leaders must focus on maximizing the use of existing power grid capacity.
Growing Strain on the Power Grid
The scale and speed of anticipated AI-driven electricity needs—alongside the impact of decarbonization and industrial growth—are putting extraordinary pressure on the power grid. Much of the US power grid infrastructure is over 25 years old, and key upgrades will require at least a decade to implement. Meanwhile, data centers and other critical sectors will drive electricity demand upward, risking shortfalls and reliability threats. The estimated electricity for AI needs alone may well exceed the net energy capacity in the US power grid by 2030. If US companies cannot access reliable and affordable power, the option of moving operations abroad becomes more appealing. With America’s AI leadership at stake, stakeholders in government and industry could expect the AI energy race to be one of the immediate battles to fight.
Strategies for Maximizing Existing Power Grid Capacity
The United States can still take actions, starting from better management and clever investments, to maximize utilization of existing power grid capacity. A RAND research team identified several actionable strategies to do so by 2030, buying valuable time for grid modernization:
- Prioritize completion of high-probability projects, especially in storage and flexible generation, to realize near-term capacity gains. This strategy focuses on advancing projects with high-capacity accreditation factors (CAF), such as battery storage and gas. We estimate the potential increase in available capacity from prioritizing the interconnection of high-CAF projects to reach 64 gigawatts (GW) by 2030. At the same time, fast-tracking projects that have already reached advanced milestones in the interconnection process—like finalized construction services agreements—can unlock significant additional capacity quickly. Policies that prioritize shovel-ready projects can deliver up to 63 GW of available capacity if completion rates are maximized.
- Delay planned retirements of existing gas and coal plants where it is safe and responsible. Where feasible, postponing the retirement of older thermal plants can immediately boost power grid reliability and resource adequacy. For example, authorizations such as California’s extension of certain natural gas and nuclear facilities granted short-term relief to grid stability. Delaying half of the scheduled retirements could result in an additional 22 GW of available capacity. Planners must also consider the environmental and health impacts associated with these older resources, but short-term actions are feasible.
- Integrate battery storage with stand-alone renewables to transform intermittent resources into a reliable supply. Co-locating battery storage with existing or new solar and wind projects enables intermittent generation to be used more reliably and dispatched when needed most. Hybridizing all solar and wind resources with storage could unlock up to 30 GW of additional capacity between 2025 and 2030.
- Make improvements to the interconnection process, both in the short term (e.g., temporary staffing, outsourcing, queue rationing) and medium term (e.g., surplus interconnection service, generation replacement, energy-only service), to reduce project backlogs and accelerate new capacity additions. Streamlining the interconnection process can significantly shorten project lead times. Measures such as prioritizing projects with higher readiness, utilizing surplus interconnection rights, and developing expedited pathways can address the current backlog and accelerate integration of capacity. Our analysis of interconnection queue data indicates that a 1-percentage-point improvement in all completion rates is equivalent to a 5-GW increase in available capacity. If completion rates increase by up to 5 percentage points, an additional 26 GW may become available, increasing net available capacity to approximately 58 GW. Longer-term reforms should focus on more flexible interconnection service models and enhanced coordination between power grid operators and developers.
Laying the Foundation for the Future of Energy and AI
Implementing these recommendations would have a measurable impact on available net capacity by 2030 and delay the necessity of sourcing power from outside the United States. These actions are just the beginning for sustainable solutions to support AI-driven energy demands, but taking the right first steps now can buy valuable time for infrastructure upgrades and build the foundation for longer-term competitive edges.
About the Authors: Ismael Arciniegas Rueda and Robin Wang
Ismael Arciniegas Rueda is a senior economist at RAND and a professor of policy analysis at the RAND School of Public Policy. Robin Wang is a student at the RAND School of Public Policy and an assistant policy researcher at RAND.
Robin Wang is a Masters of Philosophy candidate at the RAND School of Public Policy and an assistant policy researcher at RAND. His research focuses on applying data to climate resilience. He holds an MPA from the London School of Economics and Political Science and a BA in political science from the University of Rochester.
Image: metamorworks/shutterstock
