Battery storage firms eye AI demand

Battery-storage developers in the United States are getting more inquiries from AI data-centre operators that need flexible power and backup options, according to a Reuters report published on May 18. Developers told Reuters that demand is rising, but the same projects still face long waits for grid interconnection and a supply chain that remains heavily tied to Chinese components. Battery energy storage systems are already widely used in renewable-heavy markets such as California, where they absorb electricity when supply is abundant and discharge it later when demand rises. For data centres, the appeal is different: batteries can help smooth spikes in power demand, support reliability planning and make better use of limited transmission capacity. ### Why are AI data centres suddenly part of the storage sales pitch? (money.usnews.com) EPRI said in its 2026 “Powering Intelligence” work that data centres now account for about 4% to 5% of U.S. electricity demand and could reach 9% to 17% by 2030. The U.S. Department of Energy’s 2024 advisory group also said requests for hyperscale facilities of 300 MW to 1,000 MW with one- to three-year lead times were already stretching local grids. Those numbers help explain why storage companies are hearing from data-centre developers earlier in the planning cycle. (finance.yahoo.com) Reuters reported that batteries are emerging as a potential tool for front-of-the-meter load management as operators look for ways to secure power faster than the grid can expand. ### If demand is strong, what is holding projects back? Lawrence Berkeley National Laboratory said about 10,300 projects were seeking U.S. grid interconnection at the end of 2024, including roughly 890 GW of storage. (powering-intelligence.epri.com) The lab said the median time from interconnection request to commercial operation has risen to more than four years for projects built in 2018-2024, up from less than two years in the early 2000s. That queue matters because a battery project aimed at serving a data centre still has to secure transmission studies, upgrade assignments and an interconnection agreement. (money.usnews.com) Reuters said developers see those delays, alongside dependence on China-linked battery inputs and equipment, as two of the main constraints on how fast the sector can scale. ### How big is the storage market already? (emp.lbl.gov) SEIA and Benchmark Mineral Intelligence said the U.S. installed a record 57.6 GWh of new energy storage capacity in 2025. Their February 2026 outlook said 137 GWh of utility-scale storage had been installed in the country as of 2025, with more than 600 GWh expected by 2030. Those figures show the market is expanding even before AI-linked demand becomes a major standalone segment. Reuters described the current moment as one in which commercial interest is increasing faster than the underlying grid and supply-chain bottlenecks are easing. (money.usnews.com) ### What does this look like on the ground? Yellowstone County officials in Montana heard a proposal on May 22 for the McFarland Energy Storage Project, a 10-acre battery facility introduced by representatives from NextEra Energy, according to KTVQ. (seia.org) The local report linked the storage proposal to broader reliability needs and to planning around a separate proposed AI data-centre development nearby. (money.usnews.com) KTVQ reported in January that Quantica has proposed a 5,000-acre AI data-centre development near Broadview in Yellowstone County. Quantica Chief Executive John Chesser told the station the project was designed to address concerns about power use by generating its own electricity, while residents raised questions about water, taxes and transparency. ### Why are developers locking down project assumptions earlier? The combination of heavy AI-related load growth, long interconnection timelines and imported-component exposure is pushing battery teams to define key project assumptions sooner. (ktvq.com) In practice, that means owners and contractors have less room to leave interconnection arrangements, fire-safety scope and operating modes unresolved late in development, because delays in any of those areas can move the schedule. This is an inference based on the reported constraints and queue data. (ktvq.com) The next signals will come from local permitting and grid-processing milestones. Yellowstone County’s review of the McFarland proposal and additional interconnection data from Lawrence Berkeley National Laboratory will show whether storage projects tied to large new loads can move faster than the queue around them. (ktvq.com) (money.usnews.com)