Energy Dome and Salt River Project are bringing a 19 MW, 10-hour carbon dioxide battery to Arizona, with Google helping fund the project through its long-duration storage collaboration with the utility. The system is planned for SRP's Coronado Generating Station site in St. Johns, where it will use existing power infrastructure to test a storage design that does not depend on lithium-ion cells. The project is small compared with the giant four-hour lithium systems now entering service across the U.S. West. Its influence could be larger than its nameplate rating. A 190 MWh pilot attached to a real utility site, backed by a hyperscale buyer, and monitored with the Electric Power Research Institute gives the long-duration sector something it badly needs: operating data under a bankable commercial structure. AI-generated image The Arizona project puts non-lithium long-duration storage next to existing generation and transmission assets. 19 MW Planned discharge capacity 190 MWh Ten-hour storage duration 2029 Expected commercial operation What SRP Is Buying Under the announced structure, Energy Dome will own and operate the system while SRP dispatches its output under a 20-year tolling agreement. That matters because it puts the CO2 battery into a familiar utility procurement format. SRP is not simply hosting a science project. It is buying the right to call on stored energy when the grid needs it, while the technology provider carries operating responsibility. The project was selected through SRP's 2024 request for proposals for long-duration energy storage pilots. Google will fund a portion through a cost-sharing agreement tied to the company's work with SRP on non-lithium-ion long-duration technologies. The companies say the system can store enough electricity to power about 4,275 homes for 10 hours. That figure is useful because it frames the project in the right category. This is not a short-duration battery built to shave a single evening peak. It is meant to test whether a mechanical storage cycle can provide firm capacity across longer solar ramps, high-demand evenings, and reliability events when four-hour systems may not be enough. AI-generated image Utilities are testing storage that can carry energy beyond the four-hour window that dominates lithium-ion projects. How the CO2 Battery Works Energy Dome's system uses a closed thermomechanical cycle. When electricity is available, the system compresses carbon dioxide and stores it in liquid form. When electricity is needed, the CO2 is heated and expanded through a turbine to generate power back to the grid. The process uses familiar industrial equipment rather than electrochemical cells, which is why the company argues it can avoid bottlenecks tied to lithium, nickel, cobalt, graphite, and cell manufacturing. The technology is designed for eight to 24 hours of discharge. Energy Dome has said its round-trip efficiency is about 70 percent for long-duration applications. That is lower than many lithium-ion systems, but duration changes the economics. A storage asset that can discharge through a long evening or overnight shortfall can be valuable even if each kilowatt-hour cycles less efficiently, provided the installed cost, operating life, and reliability fit the grid need. The Arizona project is also a site selection story. Coronado is a legacy power station location with grid interconnection and transmission already in place. Reusing those assets can reduce one of the slowest parts of U.S. storage development: getting into the queue, obtaining studies, and paying for network upgrades. If long-duration projects can reuse old thermal plant sites, they may find a faster route to market than greenfield storage projects waiting years for interconnection. Why the location matters Old power plant sites already have roads, grid equipment, operating history, and community familiarity. Those advantages can matter as much as chemistry when new storage technologies try to reach commercial scale. Why Google Is Involved Google's role gives the project a second audience beyond SRP's service territory. Data centers are pushing utilities to add firm clean power, not just more annual renewable energy credits. Solar and wind can produce low-cost energy, but hyperscale computing loads need dependable power around the clock. Long-duration storage is one of the tools that can shift clean energy into the hours when demand remains high and renewable output drops. Google invested in Energy Dome in 2025 and described the CO2 battery as part of a portfolio of advanced technologies needed for 24/7 carbon-free energy. The Arizona project now gives that partnership a named U.S. utility site. It also gives SRP and Google a way to evaluate a non-lithium storage design before larger procurement decisions arrive later in the decade. The data-center angle is not only about corporate climate targets. It is about grid planning. Utilities serving fast-growing computing demand need resources that can support reliability during sustained evening peaks and seasonal stress. Gas turbines still dominate that conversation in many regions. A successful long-duration pilot would not replace gas capacity overnight, but it could widen the set of tools available to planners. AI-generated image Data-center load growth is pulling long-duration storage from research programs into utility procurement. The Long-Duration Test The hard part for long-duration storage has never been producing a promising diagram. It has been proving cost, efficiency, reliability, maintainability, and dispatch value in systems that utilities can finance. Energy Dome has operated a demonstration facility in Sardinia and has announced other U.S. projects, including a Wisconsin deployment for Alliant Energy and a Texas data-center agreement. The SRP project adds another step because it connects the technology to a major public power utility and a large corporate energy buyer. EPRI's role is worth watching. Independent monitoring can turn a one-off pilot into a reference case for other utilities. Developers, regulators, and buyers will want to see how the system performs during hot weather, how quickly it responds, how much maintenance it needs, and whether the real-world efficiency matches expectations. Those answers will shape whether CO2 batteries remain a niche option or compete for larger grid portfolios. Lithium-ion will keep dominating storage additions in the near term because the supply chain is mature and costs keep falling. The Arizona project does not challenge that dominance directly. It targets the hours where lithium's economics become harder, particularly when a grid needs eight, 10, or 12 hours of capacity instead of two to four. AI-generated image Long-duration systems need to prove not only storage duration, but maintenance, dispatch speed, and lifetime economics. What Comes Next The project is expected to enter service in 2029, which means it will arrive in a different market from the one that selected it. By then, U.S. storage deployment will be larger, interconnection backlogs may still be painful, and data-center demand will likely be an even bigger driver of utility planning. That timing makes the pilot both urgent and patient. The industry needs proof soon, but the project will not deliver operating data tomorrow. For Energy Dome, SRP, and Google, the value is in creating a credible bridge from demonstration to procurement. If the system works as planned, it could help utilities justify larger non-lithium storage solicitations. If it struggles, the result will still be useful because it will show where the technology must improve before it can compete against batteries, pumped hydro, compressed air, hydrogen, and gas-backed reliability options. The bottom line: Energy Dome's 19 MW CO2 battery with SRP and Google is not a scale record. It is a commercial proof point. The storage industry already knows how to build four-hour lithium projects. The harder