The Crescent Dunes Solar Energy Project in Nevada was a 110-megawatt concentrating solar plant near Tonopah that stored heat in 10,347 tons of molten salt and could run turbines for about 10 hours after sunset. It reached commercial operation in 2015, never met its promised output, lost its only customer in 2019, and entered bankruptcy in 2020. Its real legacy is technical proof that utility-scale molten-salt storage works, paired with a cautionary tale about first-of-a-kind contracts.

What the Crescent Dunes Project Was Built to Do

Crescent Dunes sits on roughly 1,670 acres of federal land in Nye County, about 14 miles northwest of Tonopah, roughly halfway between Las Vegas and Reno on US 95. The site is high desert β€” around 4,800 feet elevation β€” which matters because clear air and strong direct sun are the two things a power-tower plant cannot fake. The project was developed by SolarReserve and built under a U.S. Department of Energy loan guarantee program. It was designed to send electricity to NV Energy under a 25-year power purchase agreement.

The engineering premise was straightforward. About 10,347 heliostats β€” each roughly 1,200 square feet of mirrored glass on a steel frame β€” track the sun and bounce light onto a receiver atop a 640-foot tower. Molten salt circulates through that receiver, absorbs the concentrated heat, and then either generates steam immediately or sits in insulated tanks until the grid wants power. That dispatchability was the whole sales pitch: solar that behaves like a gas peaker.

Specification Crescent Dunes value
Gross capacity 110 MW
Net capacity 110 MW (nameplate)
Heliostat count 10,347
Heliostat mirror area each ~1,200 sq ft
Tower height ~640 ft
Thermal storage medium Molten nitrate salt
Salt inventory 10,347 tons
Storage duration at full load ~10 hours
Site acreage ~1,670 acres
Commercial operation 2015
Power purchaser NV Energy (25-year PPA)

Every one of those figures comes from project disclosures SolarReserve filed with the DOE and Nye County, and they're the numbers the plant's engineers were held to. The gap between the table and reality is the story.

How the Molten-Salt Power Tower Actually Works

Three loops do all the work, and understanding them explains both the plant's promise and its fragility. The first loop is sunlight: heliostats reflect direct normal irradiance onto the receiver. The second is salt: "cold" salt at about 550Β°F is pumped up the tower, heated to roughly 1,050Β°F, and returned to a hot tank. The third is steam: hot salt flows through a heat exchanger that boils water, and the steam spins a conventional Rankine turbine.

The advantage over photovoltaic panels is that heat is cheap to store. A PV plant either sends electrons to the grid instantly or loses them. A salt plant decouples collection from generation, so it can sell into the evening peak when Nevada's air conditioning load stays high and the sun is already down. The disadvantage is that every one of those loops is a place to fail, and the salt loop in particular is unforgiving.

Freeze protection is the constraint outsiders rarely hear about. Nitrate salt solidifies around 430Β°F, so every pipe, valve, and tank needs trace heating and a fallback plan. If a freeze event takes hold during an outage, you don't replace a fuse β€” you deal with tons of rock in your plumbing. That single operational risk shaped everything that followed at the site.

The 2015–2016 Startup: First Power and First Failures

Crescent Dunes reached commercial operation in November 2015, and the milestones came fast enough to look like success. By early 2016 the plant had synchronized to the grid and begun delivering under its PPA. NV Energy's contract priced the output well above the going wholesale rate, which was fine as long as the plant delivered. The trouble is that a first-of-a-kind plant almost never delivers on schedule, and this one didn't.

A molten-salt leak in the receiver during 2016 forced a shutdown and repairs that stretched for months. Heliostat performance and receiver efficiency both came in below design assumptions, which is the classic first-of-a-kind penalty in concentrating solar. SolarReserve later disclosed that the plant was producing well under its rated capacity during its early operating period. NV Energy, meanwhile, kept paying for a plant that wasn't generating the energy the contract assumed.

The contractual structure made this worse, not better. The 25-year PPA obligated NV Energy to buy the plant's output at a fixed price, but it also exposed the utility to the cost of the plant's shortfall through replacement power. That detail β€” a buyer forced to cover a seller's underperformance β€” is the piece most summaries of Crescent Dunes skip, and it's the piece that eventually ended the relationship.

Why NV Energy Terminated the Power Contract

By 2019, NV Energy had enough. The utility moved to terminate the PPA, citing the plant's failure to produce at contracted levels, and the dispute went to arbitration. The plant shut down in 2019 and has not generated power for the grid since. SolarReserve, the developer, had already been reworking the business and its financial position was deteriorating.

The termination exposed a structural problem with single-buyer projects. When a power plant has exactly one customer and that customer walks, the asset doesn't have a fallback market at the same price. Crescent Dunes was built to feed NV Energy's grid specifically, under a contract that justified its construction cost. Once that contract was in dispute, the plant's revenue model collapsed with it.

The timing compounded the pain. Crescent Dunes was conceived when Nevada's renewable portfolio standard made long-term solar contracts valuable and when solar thermal looked like the storage answer. By 2019, utility-scale lithium-ion batteries were being paired with cheap PV, and the economics of a 110-megawatt salt plant looked increasingly like a bet on the wrong storage technology. That shift didn't kill Crescent Dunes, but it removed any argument for a rescue.

The 2020 Bankruptcy and What Happened to the Asset

In 2020, the project's owners filed for Chapter 11 bankruptcy protection, and the plant's fate moved from operating dispute to asset sale. The receiver, the salt, the tower, and the heliostats were all still physically there β€” the plant was never dismantled after shutdown, which is unusual for a failed generation asset. What changed was ownership and purpose.

The facility's post-bankruptcy life has been as a grid-connected asset looking for a role, including interest from parties studying long-duration storage and grid services. That's a real but narrow second act: the hardware can still store thermal energy, but it was designed around a contract that no longer exists. Reuse proposals have to answer a hard question β€” what do you sell, to whom, and at what price β€” that the original developers never had to face.

The bankruptcy mattered beyond one plant. It became the reference case for critics of DOE loan guarantees and for defenders of first-of-a-kind public investment, and both sides cite it. The honest read is that the technology worked at a basic level while the commercial structure failed, which is a different conclusion than either camp prefers.

Crescent Dunes Versus Nevada Solar Today

Nevada's solar landscape in 2026 looks nothing like the one Crescent Dunes was built for. Utility-scale PV plus lithium-ion storage now dominates new construction because the capital cost per megawatt-hour delivered is far lower and the construction timeline is measured in months, not years. Crescent Dunes cost roughly $1 billion for 110 megawatts, which works out to about $9,000 per kilowatt of capacity β€” a figure no PV project approaches.

Dimension Crescent Dunes (salt tower) Modern PV + lithium-ion storage
Nameplate capacity 110 MW Configurable, commonly 100–300 MW
Storage duration ~10 hours thermal 2–4 hours typical, longer possible
Approx. capital cost ~$1 billion Far lower per MW; site-specific
Construction timeline Years, first-of-a-kind Months to a couple of years
Dispatch behavior Firm evening block Firm block limited by battery hours
Primary failure mode Salt freeze, receiver leak, contract Cell degradation, supply chain

The comparison matters because it explains why the same state that hosted Crescent Dunes now builds almost exclusively PV-and-battery plants. Thermal storage still wins on duration β€” a 10-hour block is longer than most batteries economically provide β€” but duration only pays if the price spread between afternoon and evening power is wide enough to cover the enormous upfront cost. In Nevada, it wasn't.

Cost per megawatt-hour tells the sharper story. Take Crescent Dunes' roughly $1 billion cost and spread it over the plant's design output; even at full promised production, the capital charge alone lands well above what a PV-plus-battery plant needs to recover. That arithmetic is why the model didn't get copied.

What the Project Proved Technically

Strip away the contract fight and Crescent Dunes did something no U.S. plant had done at that scale: it ran a molten-salt power tower connected to a real grid and stored hours of thermal energy for evening dispatch. Engineers learned how the receiver behaves under Nevada's irradiance, how heliostat field performance drifts from modeling, and where freeze protection fails in practice. Those lessons fed into later plants in other countries.

The storage numbers are the part worth keeping. Ten thousand tons of salt holding roughly 10 hours of full-load thermal energy is a genuine long-duration asset, and the plant demonstrated the physical principle that heat can be banked and withdrawn on demand. If long-duration storage becomes economically necessary β€” say, for grids with very high renewable penetration β€” the Crescent Dunes data set is one of the few full-scale references available.

The failures were instructive too. The receiver leak showed the maintenance burden of high-temperature hardware in a remote desert. The underperformance showed the risk of trusting first-of-a-kind efficiency assumptions in a financing model. Neither lesson is glamorous, and neither appears in promotional material about the project.

Costs, Subsidies, and Who Paid for the Experiment

The project's financing leaned on a federal loan guarantee, which is a promise to repay a lender if the borrower defaults, not a grant. The distinction matters because Crescent Dunes is often described as a taxpayer-funded failure. The loan guarantee exposed the government to loss when the project failed, and the bankruptcy put that exposure in play. The DOE's loan programs office has recovered substantial value from several troubled projects; Crescent Dunes is a case where the recovery depended on what the asset could be sold for.

The ratepayer side of the ledger is separate and often conflated. NV Energy's customers paid above-market rates under the PPA while the plant operated, and the contract dispute raised questions about who bears replacement-power costs. Those costs don't show up on a single invoice. Name NV Energy's public filings with the Public Utilities Commission of Nevada if you want the audited trail.

For anyone budgeting a similar project today, the lesson is about contingency. First-of-a-kind plants should carry reserves for receiver repairs, salt handling, and schedule slip, and those reserves need to be priced into the PPA rather than discovered after COD. Crescent Dunes' numbers suggest the reserve was too thin.

Why No Second Crescent Dunes Got Built in the U.S.

The clearest evidence that Crescent Dunes didn't transfer is that no comparable U.S. plant followed. The technology exists, the engineering is documented, and the storage case is real, yet the pipeline for new salt-tower plants in the United States is essentially empty. Cost and risk explain it. Developers can build PV-plus-battery for a fraction of the capital and finance it with conventional debt because the technology is modular and the performance risk is well understood.

There's also a buyer problem. Utilities now procure storage through competitive solicitations that favor lowest-cost, fastest-to-build options, and a 10-hour thermal plant can't win on price per megawatt. The only path is a buyer who specifically values long duration, and few utilities have modeled a need for it yet.

Policy could change that. If grid operators begin paying separately for firm, long-duration capacity, the math shifts. Until then, Crescent Dunes functions as a museum piece with live wiring β€” proof of concept without a market.

FAQ

Is the Crescent Dunes Solar Energy Project still operating?

No. The plant shut down in 2019 and has not delivered power to the grid since the PPA termination dispute with NV Energy. The physical facility and its molten-salt system were never dismantled, which is why reuse discussions kept surfacing after bankruptcy. As of 2026, the site remains a non-generating asset. Confirm current ownership status with Nye County records if you need the latest filing.

How much did Crescent Dunes cost to build?

Approximately $1 billion, according to project financing disclosures tied to its DOE loan guarantee. That figure covers construction of the tower, the 10,347-heliostat field, the salt systems, and interconnection. Spread across 110 megawatts, it's about $9,000 per kilowatt, which is why the design wasn't replicated in the U.S. The cost included substantial first-of-a-kind engineering premiums.

Why did NV Energy terminate its power purchase agreement?

NV Energy moved to terminate the 25-year PPA in 2019, citing the plant's failure to deliver contracted energy levels. The dispute went to arbitration and the plant stopped producing for the grid. The core issue was persistent underperformance against the contract's assumptions. The termination is the single decision that turned an operating problem into a stranded asset.

Does a molten-salt plant really store 10 hours of energy?

Crescent Dunes was designed for roughly 10 hours of full-load thermal storage using 10,347 tons of salt, and it demonstrated the principle at scale. Actual delivered duration depends on tank temperatures, turbine demand, and how much heat the receiver collected that day. The design figure is a ceiling, not a guarantee. Operators can withdraw less or more depending on conditions.

What's the difference between Crescent Dunes and a solar PV farm?

A PV farm converts sunlight to electricity directly and needs batteries to shift output to evening. Crescent Dunes converted sunlight to heat, stored the heat in salt, and made electricity on demand. The storage medium is cheaper by volume; the plant is far more expensive to build. That trade-off is why PV-plus-battery won commercially.

Did the DOE loan guarantee mean taxpayers lost the full billion?

No β€” a loan guarantee is not a grant, and recovery depends on what the asset sold for in bankruptcy. The government's exposure was the guaranteed debt, and the ultimate loss is a matter of public record through DOE loan program reporting. Don't assume the headline cost equals the taxpayer loss. Check the DOE Loan Programs Office portfolio reports for the audited figure.

Could Crescent Dunes be restarted as a storage plant?

Technically the salt loop and turbine still exist, but restarting means finding a buyer for its output and funding repairs and recommissioning after years idle. Those costs are substantial and site-specific. Any restart proposal would need a new offtake contract first. Without one, the economics don't close.

What's the main lesson for someone evaluating a concentrated solar project?

Model first-of-a-kind performance conservatively, price in receiver and salt-system contingencies, and never assume a single-buyer contract will survive prolonged underperformance. Crescent Dunes failed on all three. The technology worked; the risk allocation didn't. Anyone underwriting a similar plant should start from the bankruptcy record, not the press release.