Pumped hydro grid storage could be poised for a comeback

Look out, high-concept battery tech: Water and gravity may be on the brink of a resurgence. 

The Cleuson dam above Haute-Nendaz in southern Switzerland (Fabrice Coffrini/AFP via Getty Images)
  • Link copied to clipboard

Venture capitalists are pouring buckets of cash into novel ways to store clean energy, even though we already have a technology that does this, and does it well.

That would be pumped-storage hydropower, which simply lifts water to an elevated reservoir for storage, and then releases it to spin turbines and generate electricity when needed. This mechanism has been in use for more than a century and constitutes some 95 percent of grid-scale storage in the U.S. today, according to the Department of Energy.

The problem is, nobody’s built a major new pumped-hydro project in the U.S. since the Clinton presidency (though newer projects have been built elsewhere). Modern environmental laws make it much harder to devastate rivers than it was in the dam-building frenzy of the New Deal era. And while much of the innovation in energy storage targets modular and mass-produced products, pumped hydro is big, old-school infrastructure, with high upfront costs and that pesky need to pour concrete competently.

Subscribe to receive Canary's latest news

Nonetheless, an extremely patient set of pumped-hydro entrepreneurs asserts that the roar of new wind and solar construction will resuscitate this long-dormant sector.

There’s a need for just insane amounts of storage,” said Paul Jacob, CEO of Rye Development, a company that just might break the decades-long streak of inaction on new U.S. pumped hydro. The grid will need many technologies, he noted, but this one has a key advantage: The pumped storage works — it’s proven, it’s dependable.”

Rye has developed one of the three pumped storage projects in the U.S. that could now be considered shovel-ready, in that they have completed all necessary federal permitting and environmental reviews. Another 75 or so new proposed projects are wending their way through permitting, amounting to more than 50 gigawatts of storage that could someday get built. 

That’s still a hypothetical outcome.

Today’s market for storing electricity is utterly dominated by lithium-ion batteries. But they’re better at sprints than at marathons, because their costs scale unfavorably when delivering power over many hours. Batteries today work as power plants for four hours straight, as in California, where they help meet evening demand after solar power production plummets. But those batteries aren’t competitive if the power is needed for durations of eight hours or beyond.

There’s little commercial traction for that longer-term storage today, but several states say they’ll need it within the decade. A menagerie of emerging and Bill Gates–endorsed technologies could win those contracts — if they’re ready in time. 

If you squint just right, you can watch the slow-motion race afoot between novel long-duration storage technologies running from lab to the field and new pumped-hydro projects hopping from blueprint into the 3D landscape. 

Rye, at least, has all the requisite federal permits in place for a project in Oregon that won’t disturb any rivers. And the developer already sold it, preconstruction, to Copenhagen Infrastructure Partners, a reputable infrastructure investor. All the two companies need now is a customer willing to buy the plant’s output. 

Easy on rivers, heavy on jobs

On Oregon’s Klamath River, a coalition of Indigenous groups, environmental advocates and other stakeholders will soon remove four hydropower dams erected between 1903 and 1962. The goal is to revitalize the ecosystem — and the surrounding economy.

Upriver and roughly 10 miles northeast of Klamath Falls, as the crow flies, Rye Development hopes to build a new type of hydropower plant altogether.

The Swan Lake Energy Storage Project would have no connection to surrounding rivers. Rye plans to excavate two 60-acre holes and fill them with water it has already secured. The water would cycle up and down the closed system as the plant charges and discharges electricity. The total surface area of the ponds would be about the same as the Tidal Basin near the National Mall in Washington, D.C.

(Rye Development)

Two ponds may not seem radical, but they mark a decisive break with the past.

There are no closed-loop or off-river [storage] facilities in the U.S. right now,” said Cameron Schilling, VP for market strategies and regulatory affairs at the National Hydropower Association industry group. But this is the way of the future for pumped hydro: The vast majority of the approximately 75 new projects working through federal permitting today have opted for this design, he added.

Rye already has the OK it needs from the Federal Energy Regulatory Commission, which has jurisdictional authority to evaluate this type of project. Construction can commence as soon as project owner Copenhagen Infrastructure Partners gives the go-ahead; preliminary permission is expected early this year.

If that happens, the facility will be built with union labor, following a project labor agreement between Rye and the Oregon building trades. At its construction peak, Swan Lake would employ up to 2,500 or 3,000 people in the rural area, said Michael Rooney, Rye’s vice president of project management. The work would involve excavating reservoirs, preparing them to hold vast amounts of water and installing turbines in the powerhouse for generating electricity. 

It’s a massive investment in those communities,” he said. It looks and feels like a big, heavy civil project.”

The storage capacity would be similarly impressive: Swan Lake would be able to generate 400 megawatts of power, for as long as nine hours straight, if needed. That’s about as much instantaneous power as the most powerful lithium-ion battery can deliver, but the pumped plant can keep it flowing nearly five times as long. 

There’s a catch, of course. Even if the project owner greenlights construction this year, the curtain wouldn’t rise on Swan Lake until 2026. The size of the construction workforce reflects the labor-intensiveness of pumped hydro, compared to the relatively light touch and speedy timelines for laying stable foundations and dropping battery containers atop them.

If and when the plant is up and running, though, there’s little doubt it would help the Northwest absorb its rapidly growing renewable electricity supply and meet demand when wind and solar aren’t generating power.

Utilities with pumped-storage plants love them — they’re awesome,” said Paul Denholm, who models pathways to clean up the electrical grid at the National Renewable Energy Laboratory. They’re super-flexible assets that can provide tons and tons of services.”

Denholm recently oversaw a two-year effort to help the Los Angeles municipal utility plan its transition to carbon-free electricity. That involves installing lots of batteries to handle peak demand periods of around four hours. But when the system taps out on battery capacity, it has a secret weapon: the Castaic plant, a legacy pumped-storage facility north of the city with around 1,250 megawatts of capacity. 

Castaic ranks very high on the list of really nice things that L.A. is fortunate to have” to help ease the process of grid decarbonization, Denholm explained. If L.A. didn’t already have a pumped-storage plant, we definitely would have been looking at building more long-duration storage.” 

A break from past inaction

Utilities that don’t have a Castaic in their back pocket may feel compelled to build new pumped hydro for on-demand clean electricity. The tricky part is nailing the right moment to cut a deal for one of these capital-intensive, multiyear construction projects.

Most places don’t need much more than four hours [of storage duration] now, but they will at some point,” Denholm said.

Nobody’s put a bet on a large new pumped-hydro plant since Georgia’s Rocky Mountain Hydroelectric Plant came online in 1995. Many of the big ones that were built up to that point were intended to store nuclear power at night; the decline of new nuclear construction coincided with the dormancy of new pumped hydro. Rye and like-minded developers must make the case that now’s the time to break the quarter-century hiatus.

This will be a major milestone for pumped storage in this country, to get a new project done,” Jacob said. That’s a marked contrast to the battery industry, which keeps breaking records for annual installations.

The pumped-hydro plants of yore got built by centralized authorities capable of socializing costs over decades of operations — much like monopoly utilities or the federal government. A model in which an independent, non-utility developer builds a pumped-hydro plant from scratch and operates it hasn’t been done yet in the U.S.,” Jacob acknowledged. 

A successful developer needs to show investors it can take their money and actually build a plant that will operate profitably over many years. 

As with every recent game-changer in the power sector, most private investors want nothing to do with it until someone else puts their money on the line and succeeds.

Rye has some legacy backers who’ve gamely stuck with the company as it toiled for 15 years without actually building anything. Those early backers — including Crestline Investors and U.S. Renewables Group — finally got long-awaited satisfaction in 2020, when Copenhagen Infrastructure Partners bought the Swan Lake project from Rye, along with the Goldendale project in Washington state, which has yet to secure all its needed permits.

A rendering of the Goldendale pumped hydro storage project in Washington state (Rye Development)

The Danish infrastructure firm knows how to invest in megaprojects that generate predictable returns over decades. It has developed a taste for new pumped hydro: In 2019, CIP backed the Gordon Butte pumped-hydro venture being developed by Absaroka Energy in Montana.

CIP is a very sophisticated investor, and they’re taking a bigger-picture view of the needs of the marketplace,” Jacob said. They’re not doing this lightly; it’s not on a whim.” 

Private investors have bought existing pumped-hydro plants to run for profit. LS Power owns 1,620 megawatts of pumped storage in the mid-Atlantic. FirstLight Power is an independent merchant plant operator in New England that controls Northfield Mountain, the largest pumped storage in the region.

But there’s far less risk in buying a plant someone else has already built. CIP, which did not respond to questions for this article, is taking on more of a bet by buying projects that don’t yet exist.

Still missing a customer

Rye just needs to fill in the blank space in the business plan between build pumped hydro” and profit.”

The biggest obstacle [is] not really the permitting,” Jacob said. The biggest obstacle is market acceptance, making sure you have the offtakers to buy that power.” 

The early major lithium-ion batteries in California got built once developers secured anchor contracts to help a utility meet its capacity needs. Rye is looking to sign similar deals for Swan Lake. One target market is West Coast utilities that need power at times when there isn’t solar generation but are barred by climate legislation from building new natural-gas plants, the previous industry default for that role.

Since the Swan Lake project ties into a major Western transmission grid hub, it could sell to customers beyond Oregon. California’s regulators explicitly decreed that the state will need to procure 1 gigawatt of long-duration storage by 2026. Utility PacifiCorp, which operates in California, Oregon and Washington, said it’s examining new pumped hydro as a means of balancing a low-carbon grid. 

The world’s biggest tech companies say they want round-the-clock clean energy to power operations. It’s not entirely clear what those contracts would look like in practice, but some combination of renewables production paired with pumped-hydro storage for the remaining hours could do the trick.

Meanwhile, the newfangled inventions for long-duration storage are already competing for utility deals and corporate contracts. Hydrostor just raised $250 million from Goldman Sachs’ private-equity arm to excavate a novel form of underground compressed air storage. The Canadian startup is developing two projects in California with similar capacity to Swan Lake.

If you’re a utility and you need some long-duration storage, you’re in a tough spot,” said Schilling, from the hydropower industry group. You’ve got the promise of these new technologies. […] But [pumped hydro] is the only proven technology that we have.” 

That proven technology has just three fully permitted test cases right now: Swan Lake, Gordon Butte and NextEra Energy’s Eagle Mountain near Joshua Tree National Park in California. If any of them actually get built, they will fuel the ambitions of those other developers still kicking around in federal-permitting purgatory.

The National Hydropower Association is lobbying to reform federal licensing for off-river storage, given its drastically lower environmental impact compared to on-river dams. Off-river licenses currently take around five years, Schilling said, but could be sped up to as little as two years. Long-term utility planning that considers new pumped hydro would also be a boon.

Otherwise, the hydropower association’s policy wish list sounds remarkably similar to that of the battery storage industry: a general storage tax credit is the top ask. Current policy lets storage claim the 30 percent Investment Tax Credit if it gloms onto a solar project, privileging such hybrid plants over storage that stands on its own. A universal credit for any type of grid storage would level the playing field, for batteries or pumped systems. That policy is currently included in the Build Back Better legislation stalled in Congress.

Meanwhile, lithium-ion batteries continue their march of record deployments. If the current supply and logistics crunch subsides and their costs continue to fall, batteries will compete for longer and longer durations. They could eventually get cheaper than pumped storage for eight-hour applications, NREL’s Denholm said.

In other words, if the century-old storage technology can’t get its act together and build a project soon, it’ll find itself relegated back to the dustbin of history.

Julian Spector is senior reporter at Canary Media.