Arizona and Texas Set a U.S. Quarterly Battery Record: 20.2 GWh

U.S. battery storage set a quarterly record of 20.2 GWh in Q2 2026, according to the U.S. Energy Storage Market Outlook (ESMO) published on September 1 by the Solar Energy Industries Association (SEIA) and Benchmark Mineral Intelligence. Arizona led with 6.2 GWh and Texas added 3.8 GWh.

The report puts U.S. utility-scale capacity at 165 GWh, having grown from 88 GWh in 18 months, with 30.8 GWh added in the first half of 2026. The grid-scale segment contributed 17.9 GWh of the quarterly total; more than 10% of all storage ever installed in the country came online between April and June alone. The ESMO records that batteries delivered more electricity to the grid in the first eight months of 2026 than in all of 2025. The expansion is driven by rising electricity demand and the pursuit of reliability during peak hours, and is no longer concentrated in California and Texas: California added 3.6 GWh in the quarter, and Arizona, the country's fourth-largest solar market, is pairing batteries with its solar generation.

Canary Media's coverage calculates that the 20.2 GWh equates to 6.7 GW of instantaneous power, and notes that seven projects of at least 1 GWh came online during the quarter. Of grid-scale storage, 44% was built co-located with solar plants, with the remainder deployed as standalone systems. The report also counts 1.8 GWh from the commercial and industrial segment and 657 MWh from residential. Domestic manufacturing set records as well: cell plants opened in Ohio and Tennessee, and a 50 GWh module factory came online in Texas. Benchmark Mineral Intelligence raised its cumulative projection to 683 GWh by 2030, 11.5% above its previous estimate.

For Mexico and Latin America, which are beginning to tender utility-scale storage, this cost and supply-chain trajectory is the regional benchmark to watch. The next quarter's ESMO will reveal whether the new manufacturing capacity in Ohio, Tennessee, and Texas sustains the record pace.

This article was written with artificial intelligence assistance based on verified sources and reviewed by a human editor before publication.