In 2011, Daniel Nocera stood in front of a room of chemists at MIT and held up something the size of a playing card. Drop it in a glass of water, shine sunlight on it, and it split the water into hydrogen and oxygen — an “artificial leaf,” ten times more efficient at the trick than a real one. It was going to run the developing world’s homes off a bottle of water and the sun. National Geographic put him on a list of maverick inventors. His startup, Sun Catalytix, had Department of Energy money and a mission that sounded like science fiction.

The artificial leaf never shipped. The economics of splitting water for fuel refused to pencil out, and by 2012 Sun Catalytix had quietly retooled its chemistry into something far less romantic: a rechargeable battery you build out of two tanks of liquid. In 2014, Lockheed Martin — the company that builds the F-35, the THAAD missile interceptor, and the satellites that watch for launches — bought the whole thing.

That acquisition looks, in hindsight, like the single best-timed move in American grid storage. Because if you count US patents, the deepest flow-battery portfolio in the country doesn’t belong to an energy startup. It belongs to a weapons maker.

The number

Pull every US patent grant since 2015 whose title or abstract describes a flow battery — the class of grid storage that keeps its energy in external tanks of liquid electrolyte rather than in sealed cells — and sort by who owns them. The top of the list is a surprise on its own. Sumitomo Electric, the Japanese wire-and-cable giant, leads with 77 grants. Second, combining its two energy subsidiaries, is Lockheed Martin, with 73. Third, with 55, is ESS Tech, the Oregon company that is the closest thing the sector has to a household name.

So the two largest flow-battery patent estates in the United States belong to a Japanese cable manufacturer and an American defense prime. The actual grid-storage startups trail both. And Lockheed is not alone among the contractors: United Technologies holds 20 grants and Raytheon another 9 — the two companies merged in 2020 to form RTX, which means a second aerospace-and-missiles conglomerate is sitting on nearly 30 flow-battery patents. Two of the three deepest US corporate portfolios in a grid-storage technology are held by companies whose day job is national defense.

Lockheed has been filing steadily the whole time, not dabbling. Its flow-battery grants run from 2014 through 2025, peaking at 16 and 17 in 2018 and 2019, with fresh grants still issuing this year. This is a decade-long program, not a science-fair booth.

What’s actually inside the patents

The coherence test for any patent-count story is simple: delete the shared phrase and see if the inventions still belong together. Here they do, because a flow battery is a specific piece of engineering. You have two tanks of liquid, a pump, and a stack of cells with a membrane down the middle. Charging shoves electrons into the molecules dissolved in one tank; discharging pulls them back out. The elegant part is that power and energy are decoupled — want more hours of storage, add a bigger tank; want more kilowatts, add more stack. That is exactly the shape of the problem a solar farm or a military microgrid has: store cheap afternoon electrons for eight, ten, twelve hours until they’re needed.

The three leaders solve it with three different liquids, and reading the claims tells you who is betting on what.

ESS built its chemistry to be almost insultingly cheap. Its electrolyte is iron dissolved in saltwater — the company’s patents cover “alternative low cost electrodes for hybrid flow batteries” and methods for “iron preformation,” the housekeeping tricks that keep an all-iron system from gunking up. Iron, salt, water. Nothing flammable, nothing scarce, nothing that shows up on a conflict-minerals list. If you were designing the perfect earth-abundant grid battery from scratch, you would draw something close to what ESS drew.

Lockheed went the other direction, toward chemistry that is clever rather than cheap. Its portfolio is built on what the patents call “metal ligand coordination compounds” — transition metals wrapped in organic cages of catecholate or cyanide. One 2024 grant, titled “High solubility iron hexacyanides,” is a small masterpiece of practical chemistry: it claims mixing sodium and potassium salts of ferrocyanide together so that the resulting solution holds more dissolved iron than a saturated solution of either salt alone. More iron per liter of tank means more energy per dollar of infrastructure. That is the entire ballgame in long-duration storage, and Lockheed patented a way to cheat the solubility limit to get there.

Both approaches work. The question the patent data can’t answer, but the market can, is which company survives long enough to matter.

The twist

Here is where the story stops being a fun trivia fact about defense contractors and starts being a warning about how this technology reaches the grid.

ESS Tech — the pure-play, the one with the earth-abundant chemistry and the second-deepest patent stack in the country — is nearly out of money. In its own filings, the company ended the second quarter of 2025 with roughly $800,000 in unrestricted cash and issued a going-concern warning: substantial doubt about its ability to survive another year. The New York Stock Exchange has sent it two delisting notices, one for a market capitalization under $50 million and one for a share price stuck below a dollar. The company furloughed a large share of its staff and, as reported by Energy-Storage.News, described the stretch ahead as an “operational reset.” It is now pivoting toward sodium-ion, chasing a different market to stay alive.

This is the pattern that keeps repeating in flow batteries, and it explains the patent map. United Technologies had a serious vanadium flow program; it got spun off, became a company called Vionx, and went bankrupt. The graveyard of independent flow-battery firms is crowded, because the technology’s defining virtue — cheap, boring chemistry in big steel tanks — is also a commercial curse. There is no Moore’s law here, no exponential cost curve to ride down while investors wait. There is just a long, capital-hungry slog to build factories and win utility contracts, and startups keep running out of runway before the slog ends.

A defense prime does not run out of runway. Lockheed Martin reported over $70 billion in revenue last year. It can carry a grid-storage program through a decade of thin demand as a rounding error, file patents the whole way, and wait for the long-duration storage market that everyone from the Department of Energy to Colorado Springs Utilities keeps insisting is coming. Its GridStar Flow system already broke ground on a 1-megawatt, 10-megawatt-hour pilot at the US Army’s Fort Carson in Colorado in late 2022, running a two-year assessment under protocols written by Pacific Northwest National Laboratory. A ten-hour battery on an Army base is not a coincidence. It is the natural first customer for a company that has spent a century selling resilience to the Pentagon.

So the adjacent-possible move here isn’t chemical, it’s financial. The liquid that stores the electrons was figured out years ago, in a dozen flavors, by startups and national labs and one Harvard chemist’s failed dream of an artificial leaf. What was missing was a balance sheet patient enough to wait for the grid to want it. For an R&D director or a utility scout, the map is worth reading closely: the patents that will still be enforceable, and the company that will still be standing, when long-duration storage finally arrives, may belong to the same people who build the missiles.

Method note. Counts come from US utility patent grants (9.3M documents sourced from USPTO bulk grant data), filtered to grants published between January 2015 and July 2026 whose title or abstract describes a flow battery or redox flow battery. Assignee totals combine an organization’s variant spellings and subsidiary filings — Lockheed Martin’s figure sums its two energy-storage entities; RTX’s combines pre-merger United Technologies and Raytheon filings. Patent counts measure filing activity and portfolio depth, not deployed megawatts or commercial success; a large portfolio at a struggling company (ESS) and one at a well-capitalized incumbent (Lockheed) count the same here. Financial details on ESS Tech are drawn from the company’s 2025 SEC filings and reporting by Energy-Storage.News; the Sun Catalytix history from MIT News, MIT Technology Review, and IBTimes; the Fort Carson pilot specifications from Energy-Storage.News. Company histories and any private-strategy characterizations reflect public statements and filings only.