On a hot evening in the summer of 2024, grid operators in Texas watched something that would have been unthinkable five years earlier: thousands of megawatts of stored electricity poured onto the grid from battery installations scattered across the state, filling the gap left by fading solar panels just as air conditioners hit their hardest stride. Across the country, similar scenes played out in California, Arizona, and Nevada. By the end of the year, the United States had crossed a threshold. Cumulative utility-scale battery storage capacity exceeded 26 gigawatts after growing 66 percent in a single year, according to the U.S. Energy Information Administration’s generator inventory. That is roughly enough instantaneous discharge capacity to match the output of more than two dozen large natural gas plants, delivered not over minutes of turbine spin-up but in seconds.
How big 26 gigawatts really is
A single gigawatt is approximately the full output of one large natural gas combined-cycle plant. At 26 gigawatts of installed capacity, the national battery fleet could, if every unit discharged simultaneously, theoretically displace the equivalent of 26 or more such facilities for short bursts. In practice, not every battery fires at once. Units serve different roles: some smooth out moment-to-moment frequency wobbles, others arbitrage price differences between cheap midday solar hours and expensive evening peaks, and still others sit in reserve for emergencies. But even partial fleet activation during a demand spike now moves the needle on wholesale electricity prices in ways that were not possible when the national total sat below 10 gigawatts as recently as 2022.
Most of today’s utility-scale batteries are lithium-ion systems housed in rows of refrigerator-sized cabinets at solar farms or standalone sites. A typical installation can discharge at full power for two to four hours before it needs to recharge. That duration is short compared to a gas plant that can run all night, but it is precisely matched to the daily “solar cliff,” the two-to-four-hour window each evening when solar generation drops while demand stays high. Filling that window is the job batteries were built for, and the 2024 buildout dramatically expanded the workforce.
Where the growth is concentrated
Texas and California dominate the national battery map, and both states now publish data that lets analysts track battery performance independently. In Texas, the grid operator ERCOT began breaking out battery output as its own category in hourly generation data starting in October 2024, a change documented in EIA’s analysis of ERCOT’s battery records. For the first time, researchers can see exactly how many megawatts of stored electricity flow onto the Texas grid during each hour, rather than having batteries lumped in with other sources.
California tracks its own capacity through the California Energy Commission’s CEC-1304 QFER database, which logs nameplate capacity for generators of one megawatt and above. The alignment between California’s state-level tables and federal EIA totals reinforces the overall growth trend, providing an independent cross-check rather than relying on a single dataset. Beyond those two leaders, Arizona, Nevada, and Florida have all seen significant battery additions tied to large solar-plus-storage projects that came online in 2024.
What is driving the surge
Two forces converged to produce the 66 percent jump. The first is economics. Lithium-ion cell prices fell roughly 20 percent in 2023 and continued declining into 2024 as global manufacturing capacity, led by Chinese suppliers, outpaced demand. Cheaper cells made four-hour battery projects financially viable in markets where they previously could not compete with gas peakers on cost alone.
The second force is federal policy. The Inflation Reduction Act, signed in August 2022, extended and expanded the Investment Tax Credit for standalone energy storage for the first time, offering a base credit of 30 percent of project costs. That incentive removed a longstanding barrier: before the IRA, batteries qualified for the ITC only if they were charged primarily from an on-site solar array. Standalone storage projects, the kind most useful for grid reliability, were left out. The law’s passage triggered a wave of project announcements that began reaching commercial operation in late 2023 and accelerated through 2024.
The EIA’s monthly generator inventory, compiled through Form EIA-860M, tracks both operating projects and a pipeline of planned additions. That pipeline remained crowded through early 2025, suggesting the current growth wave is not a one-off spike but part of a longer build cycle. Whether every proposed project clears interconnection queues, financing hurdles, and permitting timelines is another question, but the backlog signals sustained developer interest.
What the data can and cannot prove
The installed capacity figures are on solid ground. The EIA’s inventories and California’s QFER tables are official government records that count megawatts as they come online. The 26-gigawatt total and the 66 percent growth rate can be treated as reliable.
Operational evidence is thinner. No federal dataset currently publishes second-by-second national discharge records. The EIA’s electricity data portal tracks net generation at monthly and sometimes hourly resolution, but it does not capture the sub-minute response times that define battery performance during sudden demand spikes. ERCOT and the California Independent System Operator publish real-time dashboards showing batteries responding to frequency deviations and price signals, yet those snapshots cover only their own territories. No single public source aggregates battery discharge across all U.S. grid regions at the resolution needed to confirm a single national “record” discharge event.
That gap matters for precision. Saying the fleet “can” replace dozens of power plants in seconds is an engineering capability statement supported by the physics of lithium-ion systems and the verified capacity numbers. Saying it “did” on a specific date at a specific megawatt figure requires operational data that is not yet compiled nationally. The distinction is important for policymakers debating whether to retire gas peakers: the capability is proven, but the comprehensive performance record is still being built.
The question no one has fully answered yet
If batteries are growing this fast, are gas peaker plants actually shutting down? The hypothesis is intuitive: states with the fastest battery additions should see the steepest drops in peaker runtime hours and, eventually, retirements. But confirming it nationally requires a full year of post-2024 operational data that has not yet been published as of mid-2025. Local case studies in California have shown batteries shaving peaks on specific high-demand days, pushing gas peakers offline for hours at a stretch. Extending that pattern to a documented nationwide trend in retirements or reduced capacity payments is a step the data does not yet support.
There are also headwinds. Interconnection queues in several regions are backlogged, sometimes by years. Developers face rising financing costs as interest rates remain elevated. And market rules in some regions have not caught up with battery capabilities: compensation structures for fast frequency response, one of batteries’ strongest advantages over turbines, vary widely and are still being redesigned by regional grid operators.
What this means for electricity bills and grid reliability
For consumers in deregulated markets like Texas, the expansion of battery storage can translate into lower wholesale electricity prices during evening peaks, the hours when bills tend to spike the most. Batteries that charged on cheap midday solar power and discharge during the 5 p.m. to 9 p.m. window push down the clearing price that sets what generators get paid, and in competitive retail markets, those savings can flow through to customers.
For grid operators, the speed advantage is the headline feature. A gas turbine needs minutes to ramp up. A battery goes from idle to full output in under a second. When a large generator trips offline unexpectedly or a cloud bank cuts solar output across a wide area, that response time is the difference between a managed event and a cascading frequency drop. At 26 gigawatts of installed capacity, the national fleet now offers a substantial buffer that did not exist at the start of the decade.
For policymakers and utility planners, the verified capacity figures, combined with emerging regional performance data, provide a concrete foundation for planning future resource mixes. The remaining uncertainties about duration limits, fleet-wide coordination, and long-term compensation structures are real, but they are engineering and regulatory problems, not questions about whether the technology works. The 2024 buildout settled that debate. The question now is how fast the rest of the grid adapts to a resource that can deliver power at a speed no combustion turbine can match.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
