Quiet Breakout: The Climate Battery Quietly Pumping More Oil

Drive an hour south of Bakersfield, past the pump jacks slow-nodding along Highway 33, and you arrive at the strangest power plant in California. A 20-megawatt solar array sits next to a windowless industrial shed. Inside the shed is a stack of bricks the size of a small house. Not exotic ceramics. Just refractory bricks, the same material a kiln-builder would recognize. The sun charges the bricks to over a thousand degrees Celsius during the day. The bricks then give that heat back, all night, as high-pressure steam.

The steam is injected into oil wells. It heats the crude underground, thins it out, and pushes more of it to the surface.

The world’s largest “heat battery” is a Bronze Age idea wired to a solar farm, and it is making fossil fuels.

That is the awkward, important opening scene of the most underreported energy story of the past three years. While the trade press has spent the EV-battery decade arguing about lithium chemistries and Megapack fires, a tiny cluster of startups has quietly built a different kind of battery. It stores electricity as heat in cheap rocks or carbon, then releases it as the kind of brutal high-temperature process heat that runs the world’s chemical plants, cement kilns, steel mills, and ethanol stills. The Holmes Western Oil Corporation installation near Taft, California, switched on in October 2025 as a 100-megawatt-hour Rondo Heat Battery, powered by an adjacent 20 MW solar array, charging to over 1,000°C with a round-trip efficiency above 97%. Seven months later, on May 19, 2026, Antora Energy and POET commissioned a 5-gigawatt-hour solid-carbon thermal battery at a bioethanol plant in Big Stone City, South Dakota. Fifty times larger, built in under twelve months. In January, MIT spinout Electrified Thermal Solutions turned on its first commercial-scale Joule Hive at Southwest Research Institute in San Antonio, charging electrically conductive firebricks to 1,800°C.

These three companies are doing the same thing with three different materials. They are also building a market that the lithium-ion industry, as far as the patent record shows, has decided to skip entirely.

The patents say the battery world isn’t watching

A search of US grants from 2020 onward for “thermal energy storage” coupled with “refractory” or “brick” returns 40 patents. Rondo Energy alone holds 29 of them. The remaining 11 are split across Antora, EnergyNest, Polar Night Energy, 247Solar, MIT itself, Electrified Thermal, and a handful of academic outliers. The same query run against the major lithium-ion grid-storage players (Tesla, Form Energy, CATL, LG Energy Solution, BYD, Fluence, Powin) returns zero. The companies that make every Megapack and BESS cabinet shipping into US data centers have not filed a single US grant on industrial heat storage in the last six years.

The story gets sharper when you read what Rondo is actually claiming. Of its 44 issued US grants, thirteen specifically target petrochemical or mineral processing applications: four patents on coupling the heat battery to steam cracking furnaces (the reaction that turns naphtha into ethylene, the molecule at the start of almost every plastic), three on alumina calcination (the front end of aluminum smelting), and several on what one Rondo abstract clinically calls “material processing.” US12234751, granted in February 2025, describes “an energy storage system [that] provides higher-temperature heat to a steam cracking furnace system for converting a hydrocarbon feedstock into cracked gas.” US12146425, granted three months earlier, describes the same brick-and-thermocline architecture applied to “an alumina calcination system used to remove impurities or volatile substances and/or to incur thermal decomposition.” This is not a battery company designing around the grid. It is a company designing around the inside of an ExxonMobil plant.

The reason is the size of the prize. Industrial heat, the energy needed to push a chemical reaction over its activation barrier or to bake a kiln load of clinker, is more than 20% of global energy consumption, of which about 80% comes from burning fossil fuels, according to the IEA’s 2025 Renewables report. The agency projects cumulative heat-related CO2 emissions of 100 gigatonnes between 2025 and 2030, “more than one-fifth of the carbon budget remaining for limiting global warming to 1.5°C.” Electrifying the car fleet, the project that consumed venture capital from 2015 to 2024, addresses a smaller share of emissions than this single overlooked sector.

A 1920s technology that no one needed until now

The engineering does not look new because it isn’t. Firebricks are roughly 3,500 years old, first fired in the iron-smelting kilns of the Hittites. Resistance heaters that turn electricity into heat are early-twentieth-century technology. The novel part is the configuration: stack the bricks into a thermally insulated box, wire a high-voltage DC heater through it, and shape the internal cavities so that radiation from the heated zones rebounds uniformly through the mass. Then push room-temperature air or gas through structured channels in the brick array. A thermocline forms; the gas comes out the other end at the same temperature it would have come out of a gas-fired boiler, with no flame and no carbon.

The conceptual leap was made not at a startup but at MIT in 2017, by Charles Forsberg, a senior nuclear engineer, who pitched a system he called FIRES (Firebrick Resistance-heated Energy Storage). Forsberg told MIT News at the time that the technology “could have been developed in the 1920s, but there was no market for it then.” His student Daniel Stack completed a PhD thesis in 2021 on tuning firebrick chemistry so the bricks themselves became electrically conductive, letting them serve as their own heating element and eliminating the resistance wire altogether. Stack spun the work out as Electrified Thermal Solutions and has since lined up backing from three of the most carbon-intensive companies on earth: Holcim (cement), Vale (iron ore), and ArcelorMittal (steel). Rondo, meanwhile, was founded by John O’Donnell, a former Princeton Plasma Physics Laboratory engineer who had previously co-founded the solar-thermal companies Ausra and GlassPoint and watched both get squeezed by tumbling photovoltaic prices. The thesis at Rondo is that PV finally got cheap enough to make Forsberg’s twenty-year-old observation profitable.

The customer list is the disorienting part

The Holmes Western installation generates emissions-free steam to extract more oil. Big Stone City uses the same physics to evaporate water out of corn mash for ethanol. The Joule Hive in San Antonio is sized to displace gas burners in glass kilns, cement plants, and steel reheat furnaces. None of these is a feel-good headline. All of them are real demand. A heat battery shipped to an ethanol plant in May 2026 cuts that plant’s gas bill on day one and removes 13,000 tonnes of CO2 a year on the Holmes site alone, according to project documents. A project that closes on those economics ships, while a project that requires a customer to wait for “the hydrogen economy” does not. The Rondo CEO Eric Trusiewicz told reporters in October that the company is “operating heat batteries across four continents and five industries”; Andy Lubershane of Energy Impact Partners predicted in the same release that heat batteries will open “an even larger new market” than the EV-driven grid storage boom did.

For an R&D director at a chemical major, a corporate-development scout at a steel company, or a VC who watched the lithium-ion supply chain consolidate around three Chinese firms, the read is straightforward. A field that the major battery players have not entered is now putting hardware in the ground at gigawatt-hour scale, with industrial customers signing long-term offtake. The patent landscape is empty enough that a serious entrant filing today could still be a top-three holder in three years. The materials science is from the Hittites. The physics is from the Edison era. The arbitrage is between a $20-per-megawatt-hour solar PPA at noon and a $40-per-megawatt-hour natural gas boiler at midnight.

The strangest power plant in California is the leading indicator. Watch the ones in Big Stone City and San Antonio next.

Method. Patent counts in this brief come from a full-text search of US utility grants (~9.3 million documents sourced from USPTO bulk grant XML) issued on or after January 1, 2020, matching “thermal energy storage” together with “refractory” or “brick” in the title, abstract, or claims. Each company’s tally combines variant spellings and known subsidiaries. Industrial-emissions figures are from the IEA’s 2025 Renewables report. Company deployment details and quotes come from primary press materials linked inline and from coverage in Energy-Storage.News, Electrek, MIT News, and Latitude Media. Caveat: a patent grant is a claim on paper, not a deployed system; the Holmes Western, Big Stone City, and San Antonio installations are the three commercial-scale units publicly confirmed as of mid-May 2026.