A Hillside Battery That Does Not Need a Mountain

Frontier Signal · Deep Dive · Energy & Investing · 7 min read · Friday, 18 September 2026

At a kaolin mine outside Plymouth, three fibreglass pipes run up a hillside. A pump pushes liquid to a reservoir at the top. When power is wanted, a valve opens and gravity sends it back down through a turbine. No lithium, no hydrogen, no chemical reaction, nothing that degrades with cycling.

This is pumped hydro, the oldest and largest form of energy storage on Earth, and the fussiest about geography. Conventional schemes want hundreds of metres of drop, which means mountains.

RheEnergise did not change the machine, it changed what is in the pipe. Its working fluid is a mineral suspension about 2.5 times as dense as water, so the same volume falling the same distance carries about 2.5 times the energy. Make the liquid heavier and the hill can be smaller.

Neat physics. The harder question is whether a heavier liquid produces an asset a lender will finance.

500 kW REACHED · 59% ROUND-TRIP EFFICIENCY · 15 MINUTES OF DURATION
Measured at the Devon demonstrator. Efficiency excludes parasitic loads. The design target was four hours.

The first two numbers are the ones the company leads with. The third decides whether this is a long-duration storage technology yet, and it comes from the same government report.

On 27 January 2026 RheEnergise said its first High-Density Hydro project had reached full power. The plant sits at Sibelco's Cornwood kaolin mine in Devon, built under the UK government's Longer Duration Energy Storage demonstration programme. Grant support rose from £8.24 million to about £9.8 million, alongside a €2.5 million European Innovation Council award.

It was designed for 500 kW over four hours and it repeatedly reached 500 kW. A purpose-built turbine, controls and hundreds of cubic metres of suspension worked together outside a laboratory, which had never been shown. It was a Technology Readiness Level 7 demonstrator, it was not grid connected, and it established no lifetime, financing cost or levelised cost of storage.

The old way: pumped hydro stores energy by lifting mass, and what you get back rises with the mass moved and the height it falls. Water is not heavy for its volume, so conventional schemes chase height, which means mountains and decade-long planning.

The new way: RheEnergise attacks the other term. Its closed-loop fluid is a suspension of abundant minerals at about 2.5 times the density of water, so a project delivers the same energy from less liquid or a smaller drop. The company targets 75 to 300 metres against about 500. Those are design claims, not rules, and a water turbine cannot handle a slurry, which is why this one was built from scratch.

In plain English, the company did not invent a better battery. It changed the weight of the water.

Batteries are good at minutes and hours. Grids running on wind and solar also need assets that cover an evening peak and a still, grey week. That is the long-duration storage problem, and a crowded field is competing for it, from a battery that rusts on purpose to hold a grid for days to one that stores power as liquid carbon dioxide.

High-Density Hydro targets 4 to 8 hour modules of 10 to 20 MW and an asset life beyond 60 years. The likely first market is industrial: mines and quarries already have disturbed land, roads, grid connections and elevation. Consent, geology and grid queues do not disappear because the hill got shorter.

The strongest result is integration. The turbine matched RheEnergise's models, the system reached rated output, and the team made 465 cubic metres of fluid in five weeks.

The report also records the friction. Pumping efficiency was 73% and turbine efficiency 80%, for a combined round-trip of 59% before parasitic loads against an 80% target. Fluid production is where it hurt: output averaged 1.09 cubic metres an hour, about 32% of design rate, after damp feedstock slowed the powder feed and one off-specification batch took three months to diagnose.

That shortfall has a consequence the headline number hides. With too little fluid in the loop, a plant designed to hold four hours held roughly fifteen minutes at full power. For a technology whose whole pitch is duration, that is the figure to sit with. It is a manufacturing failure rather than a physics failure, which is the better problem to have, but it is not a small one.

RheEnergise owns the intellectual property and is the clearest beneficiary if this becomes a repeatable category. It is private, so there is no way to buy it. Public support has been large relative to its size: about £9.8 million plus the €2.5 million EIC award.

The model is asset-light: RheEnergise supplies technology and engineering while developers finance the reservoirs and equipment, so adoption depends on warranties, guarantees and contracts a lender will accept.

Anglesey Mining (AIM: AYM) is the only disclosed listed link, and it is thinner than it looks. In July 2025 it signed a letter of intent covering a jointly owned vehicle to explore a scheme at Parys Mountain in Wales. The agreement is exploratory, no final investment decision exists, and Anglesey remains exposed primarily to Parys Mountain itself. Treat it as an option, not a valuation anchor.

Sibelco, which hosted the site, and 42 Technology, which engineered the fluid plant, are private. No dependable listed revenue line exists today.

The competitive set is not other pumped hydro. It is everyone chasing multi-hour storage without lithium, and this is an industry read-through rather than evidence that any of them is affected by the Devon result. Energy Dome liquefies carbon dioxide in a closed loop, the closest cousin: mechanical rather than electrochemical. Form Energy takes the opposite route, iron-air chemistry. Quidnet Energy pumps water into rock and lets the geology be the upper reservoir. All face the test RheEnergise has just partly failed: not whether the physics works, but whether it can be made at rate and financed normally. The winner will be whichever gets boring first.

The demonstrator proved power. It did not prove duration, and it did not prove bankability.

At 59% round-trip efficiency the system returns materially less electricity than lithium-ion does. RheEnergise argues bigger equipment and better pumps lift it towards 80%. That is engineering logic, not field evidence, and the gap is where the economic case lives.

The fluid is a second concentration of risk. The company calls it benign and the loop is closed, but a buyer will want decades of data on stability, seal compatibility, corrosion, abrasion, settling and disposal. The demonstrator could not make it reliably for five weeks.

Smaller hills expand the map, they do not erase project development: land, two reservoirs, consent, a grid connection and civils. The estimate of 6,500 UK sites is a screening result, not a pipeline. And the regulator has not bought it yet. Ofgem's first cap-and-floor window shortlisted 16 projects across pumped storage hydro, compressed air, lithium-ion and vanadium flow. Dense-fluid hydro is not among them. That is a statement about what is financeable today, not a verdict on the science.

What is the independently verified cost per delivered megawatt-hour, after fluid manufacture, parasitic loads, maintenance and financing, at the first commercial site?

Site counts are a screening exercise and every storage company has an impressive one. The number that decides this is what a megawatt-hour costs from a plant somebody paid for, and nobody has it.

Stored energy depends on how much mass you lift and how far it falls. A liquid 2.5 times denser carries more energy per cubic metre at the same height, so a project works across a smaller drop.

The plant was designed for four hours at 500 kW and could not be filled. Fluid manufacture ran at about a third of design rate, so the inventory supported only about fifteen minutes at full power. The power rating was met, the duration rating was not.

No, it is private. Anglesey Mining is listed and has an exploratory agreement, but that is an option on a possible project, not ownership of the technology.

Operating data and measured efficiencies come from the UK government's final report and independent analysis of it. Forecasts and targets are labelled as such.

Frontier Signal explains frontier technology in plain English. This is general information, not investment advice.