ChatGPT & DALL-E panoramic image of a cavern underground full of water with an engineer in a hard hat looking at it and scratching her head.

Compressed Air Storage Redux — LightSail & Hydrostor

Sign up for daily news updates from CleanTechnica on email. Or follow us on Google News!

The other day, I published my doorstop assessment of the thermodynamics and operational challenges that limit compressed gas storage solutions, pushing them into the 100 GW of also-ran capacities. In comments in various places, people reminded me of the long-defunct LightSail and the now-getting-press Hydrostor. A bit of followup with some comments and assessments on them seemed worthwhile.

As a reminder, there are only a couple of reasonable-sized compressed air storage solutions actually on grids, and they have been operating for decades without any more being built. The fundamentals of physics, thermodynamics, electrical generation, and compression technologies related to compressed gas solutions haven’t changed because they are absurdly mature. We’ve been compressing gases for energy storage, energy transfer, and multiple other purposes for hundreds of years. Asserting disruptions that change the game in this space is an extraordinary claim and hence requires extraordinary evidence. It’s not forthcoming.

LightSail was one of the Silicon Valley, Peter Thiel acolytes, ‘energy disruptors’ that came to nothing because the grid fundamentally isn’t like the internet, electricity can’t be packetized like data, and there are no equivalents to the absurd data transmission multiplications that fiber multi-plexing brought to the internet. Thiel, and in this case Danielle Fong, thought that the grid and utilities were ripe for the same disruption that they wrong on legacy media and communication companies, but they didn’t start from a realistic understanding of the physics and economics, and so have sunk into the depths of internet time.

I became aware of LightSail in 2011, per my Quora history, when I answered a couple of questions about it and had a very brief interaction with Fong. The premise of their ‘innovation’ was carbon fiber tanks capable of holding 200 atmospheres of pressure, along with some promised ancillary techniques and approaches that would enable relatively efficient compression to those levels. For context, that’s about what air in scuba tanks is compressed to in order for divers to spend more than a few minutes underwater.

A metric ton of air occupies about 840 cubic meters at sea level at 20° Celsius, the usual starting point for my quick workups. At 200 atmospheres, the ton would occupy about 4.3 cubic meters. Compression would take about 2.35 MWh. One of the big problems of compressed air storage is that it requires big caverns which have to be airtight. That’s possible with salt and hard rock cavern construction (more on the latter when we get to Hydrostor), but maximum operational pressures are frequently in the 40 to 80 atmosphere ranges, with one Chinese salt cavern having potentially 180 atmospheres, because the higher the pressure, the more sophisticated the compression requirements are and the more difficult it is to keep seals intact throughout the system.

As a reminder, that’s one of the many big problems with hydrogen for transportation, and in fact any energy solution, in that hydrogen has to be compressed to 350 to 700 atmospheres, well above these pressures, and it’s a much tinier molecule. A nitrogen molecule is about 2.6 times larger across than a hydrogen molecule, with an oxygen molecule being a bit smaller than nitrogen, but close.

We’ve been compressing air to 200 atmospheres for an awfully long time, so Fong and team’s claims that they were doing something groundbreaking there was unlikely. It’s not Theranos-scale overstatement, but really, these pressures and the thermal management of them have been industrialized for yonks. The big thing was avoiding the requirement for caverns and the risks and limitations that they bring by bringing pressures up higher than traditional compressed air solutions. The bad thing was very expensive, bespoke, spun-fiber tanks. That was in an era when carbon fiber was high on the hype scale, so it undoubtedly didn’t hurt, just as having energy-illiterate Thiel’s connections to venture capitalists and the Valley in general didn’t hurt, and just as it didn’t hurt that Fong was photogenic and well-spoken.

Of course, LightSail ran into exactly the same end-to-end efficiency problems that I outlined in the big assessment piece — the cost of the tanks and related compression equipment exceeded the value of the energy, Fong and team were only aware of a slice of the electricity market in one part of the USA, and so they couldn’t find anyone willing to buy it.

Like a lot of startups, they pivoted to their natural market, which was to say compressed fossil gases. After all, virtually every cavern we have manufactured globally, just over 2,000 salt caverns and only about 200 hard rock caverns — hmmm, foreshadowing — is being used to store natural gas, crude oil, or liquified petroleum gas. Bit of a turnaround from disrupting electricity to marketing to the gas industry, but it’s not like it worked in the end. It was asserted to be a pivot, but it was more of a fire sale of their existing tanks as part of shutting the business down.

As an amusing side note, LightSail’s original business model was to power an urban scooter, indicative that they weren’t starting from a need and finding a solution, they were starting with a solution — compressed air and carbon fiber — and trying desperately to find a place to use it. Many of the usual suspects for bad climate solutions funding, including Khosla and Gates, joined Thiel and others in wasting about $70 million on something that just wasn’t all that.

But mediocre ideas that keep failing to be competitive in the real world and with no fundamental changes in technology that would enable them to work this time keep finding new investors.

One of the hotter compressed air firms right now is Hydrostor. It is a Toronto-based startup that’s managed to get a bunch of investment from various people and infrastructure investment tied to achieving specific milestones of $250 million from Goldman Sachs for one of its projects.

Like every other compressed air solution organization, the company claims to have a secret sauce, but to be clear, they aren’t making too extreme claims about round trip efficiency and are explicitly leveraging all the bog standard technologies and processes that already exist instead of pretending that they are finessing the laws of physics with new technology.

Also amusingly, Hydrostor is actually a pumped hydro solution with some additional energy-sucking steps which it claims as advantages.

Schematic of Hydrostor's advanced compressed air energy storage by author
Schematic of Hydrostor’s advanced compressed air energy storage by author

When Hydrostor’s energy storage solution has no energy, all of the water is in the underground cavern. When they pressurize air and send it down, they push the water out and up to the on-ground reservoir. Yes, this is more pumped hydro, but pumping air and generating electricity through less efficient gas compressors and turbines instead of more efficient reversible water turbines.

They are limited to the amount of energy provided by the water in the reservoir, which isn’t magically different than pumped hydro. It’s still mass times acceleration due to gravity times height — grade 7 science. They claim that they get more energy out for the same amount of water, but that doesn’t stand up to the slightest scrutiny.

Let’s make it simple. The fundamental thing that they are doing is moving mass uphill to create potential energy. That potential energy is held back by the air in the cavern. Reversing that to allow the water to flow back into the cavern can’t return more energy than the total required to move it upward.

Despite this, in one of the presentation decks maintained on the DOE’s website, Hydrostor makes an extraordinary claim:

With the same head as the A-CAES system (600m), pumped hydro requires ~5X more water than A-CAES (150m3 /MWh vs. 770m3 /MWh). At a more conservative head of 150m, Pumped Hydro requires ~20X more water than A-CAES (150m3).

As a reminder, this is mass times acceleration due to gravity times height. Five times more water is five times the mass at the same head height. Five times the mass of water equals five times the potential energy at the same height. That’s the way physics works. It doesn’t matter how they pump the water uphill, Hydrostor is still pumping water uphill as the basis of energy storage.

Given that the executives are business types with no STEM backgrounds, even though this is grade 7 science using grade 5 math, perhaps they just don’t realize how extraordinary a claim this is. It’s possible that this is simply an innocent mistake in the middle of the deck that no one caught. But it does make me suspicious of the rest of their math claims.

And underground caverns are expensive to create. A big underground storage cavern contains a million barrels of oil, which sounds like a lot, but it’s only about 160,000 cubic meters. That’s about 160,000 tons. At a head height of 400 meters, that’s only about 174 MWh, which once again sounds like lot, but pumped hydro generally starts at multiples of that. Making caverns bigger and deeper multiplies their costs in two different ways. The first is just that a lot more rock a lot further down has to be blasted out and removed, and the second is that it gives two physical directions to find out something unexpected like an igneous intrusion and time for the project to go off the rails.

The only secret sauce in this compressed air storage is that the use of water maintains the pressure of the air being released so the turbines that capture that mechanical energy operate a bit more efficiently, and the size of the cavern can be smaller because they don’t run out of useful pressure as quickly. That’s an accurate statement on their part.

But the use of water means that they can’t use cheap and low-risk salt caverns for compressed air storage either. The water would end up eroding away the salt and cause the system to fail.

Flyvbjerg cost overruns table
Flyvbjerg cost overruns table

There is a reason why there are ten times as many manufactured salt caverns in the world than hard rock caverns for fossil fuel storage, and that’s cost and risk. As Professor Bent Flyvbjerg’s 16,000+ data set on megaprojects shows, every time we build infrastructure underground, risks shoot up. Tunnels are halfway down the 25-category list of projects by schedule and budget because a lot of things get discovered only when we start digging. Mining isn’t as far down the list, but it has a lot more risks than aboveground assets.

Hydrostor is correct that this solution can be built in places where pumped hydro can’t, as if this is actually something interesting. The siting conditions for its solution require good rocks, good groundwater to keep the cavern from leaking too much by keeping water in all of the cracks in the fractured rocks, flat land, and no transmission lines.

The company has two working facilities right now, one a basic lab version and a tiny operational one in Goderich, Ontario. It’s 10 MWh of storage, which is to say a bit more than a couple of Tesla Megapacks.

They have two in development, one Broken Hill, NSW in Australia, and one in California. The one in Australia is at least in a flat part of Australia with close to no transmission of any significance, and they call out the lack of transmission as the cost justification for building energy storage there. That’s fine, and storage is often built to defer or avoid entirely transmission and distribution upgrades. It’s legacy Australian mining country, so there is a good deal of understanding of what’s under the ground there, so that risk is somewhat mitigated. But Broken Hill has a semi-arid climate, just a bit better than an all-out desert, so I suspect if the solution actually gets built they’ll find air leakage rates eating into the efficiency.

The California one is in central California near Bakersfield. California isn’t remotely challenged as far as hills go, and its grid is a lot more robust than Australia’s Outback. Two strikes against this location, but California has more money to waste on bad projects than most. Any other strikes?

Yes, Hydrostor keeps finding things underground that are making cavern construction difficult as they drill boreholes.

Hydrostor does claim to be more efficient roundtrip than other compressed air storage solutions. Existing ones are running at 42% and 54% efficient, so better than that is required. But it is only claiming 60% efficiency, with a hopeful plus sign following that point.

ANU global pumped hydro site map courtesy ANU
ANU global pumped hydro site map courtesy ANU

What are they really competing with? Pumped hydro. This is the Australian National University greenfield atlas of closed loop, off-river pumped hydro sites with greater than 400 meters of head height, reservoir siting options within three kilometers to limit tunnel length, close to transmission, not on any existing waterways, and off of protected land. The big blank spots in Alaska, Canada, and Russia are because virtually no one lives there and there’s no transmission, similarly the empty middle of Australia.

Note all of the pumped hydro resource locations available on the west coast of the USA, within easy electron-spitting distance of California.

As another reminder, pumping water and generating electricity from it with spinning turbines is more efficient than pumping air and spinning turbines with air. Gases vs liquids have different properties that make it work better. 80%+ roundtrip efficiencies are easily achievable, and water just sits in the reservoirs.

That’s why the biggest store of energy on the grid is pumped hydro today, we’ve been building pumped hydro since 1907 — China has 19 GW in operation and 365 GW in construction and planning — yet compressed air is limited to a couple of decades-old sites in Germany and the USA.

Hydrostor is just an inefficient pumped hydro solution with different siting limitations. Maybe it will get a few built, but having reviewed the company, it remains firmly in the also-ran category of grid storage, just like every other compressed gas solution I’ve looked at over the past 13 years.

Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.

Latest CleanTechnica.TV Videos

CleanTechnica uses affiliate links. See our policy here.

Michael Barnard

is a climate futurist, strategist and author. He spends his time projecting scenarios for decarbonization 40-80 years into the future. He assists multi-billion dollar investment funds and firms, executives, Boards and startups to pick wisely today. He is founder and Chief Strategist of TFIE Strategy Inc and a member of the Advisory Board of electric aviation startup FLIMAX. He hosts the Redefining Energy - Tech podcast ( , a part of the award-winning Redefining Energy team.

Michael Barnard has 723 posts and counting. See all posts by Michael Barnard