One of the most pressing challenges in the U.S. energy market today is interconnection delays. According to Lawrence Berkeley National Laboratory’s Queued Up report, at the end of 2024, nearly 2,300 gigawatts of generation and storage capacity were seeking access to the grid, underscoring the depth of the crisis facing the energy market. Many of these projects, particularly solar and wind, are seeking to connect to the grid specifically to help meet growing electricity demand. The backlog represents not just a logistical bottleneck, but a fundamental tension between how fast energy demand is rising and how slowly new supply can come online.
A recent Utility Dive article explores how energy technology company GridCARE is working to ease that tension. In partnership with National Grid, one of the world’s largest investor-owned energy companies, GridCARE is using artificial intelligence (AI) to help “identify and unlock” interconnection capacity for large-load customers in New York. AI’s surging energy appetite and Big Tech’s infrastructure expansion are among the key drivers of increased electricity demand and, more consequentially, the interconnection delays now affecting several states, not just New York.
To unlock this hidden potential, GridCARE’s AI-powered platform, Energize, uses generative AI to sift through quadrillions of operating scenarios, identify conditions under which the grid could become constrained, and surface potential capacity-boosting solutions including flexibility assets like batteries and other distributed energy resources. The idea of finding spare grid capacity may sound surprising given how difficult it has become to secure interconnection for new large-scale projects, but that is precisely the point. According to GridCARE, spare capacity could reduce interconnection timelines from several years down to as little as six to twelve months.
If successful, Energize has the potential to benefit more than just grid operators and developers. For large commercial and industrial energy buyers, persistent interconnection delays contribute to tighter supply conditions and upward pressure on electricity prices, making it worth asking whether long-term procurement strategies, including power purchase agreements, should account for where and how quickly new generation can reach the grid. GridCARE’s own modeling suggests that when large energy users like data centers incorporate even a modest amount of flexibility into how they consume power, the benefits extend beyond their own operations. In one hypothetical scenario, that flexibility could reduce electricity costs by 5% across all customer classes or free up more than $1.35 billion in utility capital.
GridCARE is not alone in bringing AI tools to bear on this challenge. Constellation Energy and GridBeyond are tackling the problem from the demand side, using an AI-driven demand response platform to reduce peak load and ease grid strain. As Constellation CEO Joe Dominguez put it when announcing that partnership, the platform is “a better, cheaper, and faster way to meet America’s growing energy needs, help win the AI race, and lower costs for consumers.” Their work is a reminder that smarter grid management does not always mean building more infrastructure; sometimes it means using what already exists more intelligently.
Calpine Energy Solutions’ CORE (Capacity Obligation Reduction Effort) program is a good example of how demand flexibility is already embedded in programs available today. The program notifies business customers when system peak load days are likely to occur, giving them the opportunity to voluntarily curtail usage during the top hours of highest ISO system demand each summer. By reducing their Peak Load Contribution during those critical hours, customers can lower their capacity charges on a dollar-for-dollar basis in the following planning year. Programs like CORE illustrate that flexibility does not always require cutting-edge AI or major infrastructure investment. Sometimes it starts with a notification and a willingness to adjust, and the more intelligently we manage when and how power is consumed, the more capacity the grid can unlock for everyone.