Why Dispatchable Solar Is Reshaping How Grids Buy Clean Power

Most investors and grid planners treat solar as interchangeable. But there is a version of solar power that can be switched on at 10pm on a Tuesday to meet evening peak demand, without a single battery cell involved.

That distinction is quietly reshaping how governments procure clean capacity across three continents. As grids absorb more wind and photovoltaic generation, the missing piece is no longer more energy. It is firm, schedulable capacity: electricity that arrives when the system needs it, not when the sun happens to shine.

Capacity constraints in modern grids are increasingly the binding problem rather than fuel costs or generation volume, which is why procurement frameworks in high-growth markets are being redesigned to reward resources that deliver power precisely when demand peaks rather than resources that simply contribute energy across the day.

Concentrated solar power towers with molten salt storage solve exactly that problem. And the markets where sunlight is most intense are beginning to design procurement frameworks built to reward the capability.

Here is what you will take from this: how the technology physically works, why its economics look different from standard solar comparisons, where it is being built and the grid logic driving each market, and what the current pipeline signals for anyone tracking clean infrastructure over the next decade.

How a solar tower actually works after dark

Picture a field of thousands of mirrors arranged in concentric rings around a single tower. At the top of that tower sits a receiver glowing orange, because every mirror in the field is redirecting sunlight onto that one point. The temperature there climbs high enough to melt salt. That is the entire trick, and it is why this version of solar behaves differently from everything else sold under the same name.

The point worth holding onto is that this plant stores heat, not electricity. Everything that makes its economics unusual flows from that single fact.

Concentrating the resource: heliostats and the central receiver

Each mirror is a heliostat, a reflector that tracks the sun individually and aims its beam at the receiver. Thousands of them working in concert concentrate enough solar energy to reach the temperatures molten salt requires.

Precision is what makes this viable rather than a physical curiosity. Good pointing precision distributes heat evenly across the receiver surface, which limits the risk of unplanned outages and equipment damage while also reducing the total number of mirrors a plant needs to hit its output target, lowering both capital cost and land use.

That precision was enabled by compact, high-torque permanent magnet motors. In desert deployments, high-temperature magnet grades enhanced with dysprosium and terbium keep the motors reliable in extreme heat, a supply chain detail that becomes a risk factor later in this piece.

Storing heat, not electrons: the two-tank molten salt cycle

The storage system uses two tanks of nitrate salt. Cool salt is pumped up to the receiver, heated by the concentrated sunlight, then held in a hot tank. When the grid needs power, that hot salt is drawn down through a heat exchanger to raise steam, the steam drives a turbine, and the now-cooled salt returns to the cold tank to be reheated.

This is thermal energy storage, and it is the operative concept for understanding the economics. Because the hot salt can sit in its tank for hours, the generator can run whether or not the sun is currently shining.

The operational sequence runs in four steps:

• Solar collection: heliostats concentrate sunlight onto the receiver, melting the salt.
• Heat transfer: hot salt moves to the hot storage tank.
• Dispatch: hot salt passes through a heat exchanger, raising steam to drive the turbine.
• Return: cooled salt flows back to the cold tank to be reheated.

Modern commercial plants store 8 to 17.5 hours of energy. Noor Energy 1 in Dubai pairs a 100 MW tower with up to 15 hours of molten salt storage, and China’s Shouhang Dunhuang Phase II runs a 100 MW state-of-the-art molten-salt tower.

Capacity factor: 40% to 70% In high-DNI locations, tower CSP plants achieve capacity factors between 40% and 70%, against the typical of utility-scale PV without storage. That gap means far more hours of annual generation per unit of installed capacity, and it is why grid planners treat these plants as firm capacity rather than an intermittent resource.

The 15-hour storage figure is not an engineering footnote. It means a plant can run on stored solar heat from noon until 3am, dispatching into the evening peak long after sunset.

Why dispatchability changes the economics entirely

Here is a comparison that trips up most people reading solar cost data. Put a tower CSP plant next to a PV-plus-battery system and compare their cost per megawatt-hour, and the answer you get depends entirely on how many hours of storage you are asking for. At short durations, the comparison favours batteries clearly. At long durations, it flips. Both can be true at once, and understanding why is the point.

For short storage durations, PV combined with lithium-ion batteries is unambiguously more cost-competitive. That is not a close call, and positioning CSP as a universal winner would be wrong.

The split emerges at longer durations. Battery cost scales with electrochemical capacity: more storage means more cells, and each cell carries a manufacturing cost. Molten salt storage scales differently. Adding more hours means adding more salt and building larger tanks, which is a materials and civil engineering cost curve, not an electrochemical one. Beyond longer durations, that difference compounds in CSP’s favour.

Long-duration storage alternatives to molten salt, including compressed air and hydrogen stored in underground caverns, follow a similar cost logic: the storage medium itself is cheap and abundant, so capacity scales at marginal material cost rather than at the electrochemical cell cost that makes battery systems expensive at longer durations.

Chinese cost trajectory According to Frost & Sullivan industry data, the levelised cost of CSP in China fell from RMB 1.15/kWh in 2016 to RMB 0.55/kWh in 2024, with projections toward RMB 0.38/kWh by 2035.

State-of-the-art molten salt towers in high-DNI regions now achieve a levelised cost of energy below 0.10 USD/kWh. Shouhang Dunhuang Phase II holds the lowest recent reported value at 0.08 USD/kWh.

That 0.08 USD/kWh figure is not a benchmark you can lay against a PV project. It is the cost of delivering controllable, schedulable electricity across a 15-hour window, which is a fundamentally different product from what a solar panel sells to the grid.

The right frame for a grid planner is cost per unit of firm, dispatchable capacity, not cost per megawatt-hour of energy. Compare CSP’s LCOE directly against PV’s without accounting for dispatchability, and you are comparing prices in two different markets. That is the error this section exists to prevent.

Where high-DNI grids are procuring tower CSP and why

The economics only matter where a grid has a problem that dispatchable solar actually solves. Three very different markets, China, the MENA region, and Chile, are building tower CSP for reasons that look unrelated on the surface but share one underlying logic: each has a grid need no alternative currently meets at scale.

China is the dominant scale story. In December 2025, policy directives established a target of 15 GW by 2030, alongside an ambition to reach LCOE equivalence with coal-fired power. This is the single largest national deployment commitment in the technology’s history, driven by heavy midday PV curtailment and evening peak demand pressure across northwestern grid regions.

The MENA region offers a structural fit rather than a scale story. High direct normal irradiance meets rapidly growing electricity demand, much of it from air-conditioning loads that concentrate in the evening, exactly when PV output fades. Noor Energy 1 in Dubai, with its 100 MW tower and 15 hours of storage, is the operational anchor for the region.

Chile mirrors the MENA logic at smaller scale. The Atacama Desert delivers some of the world’s best DNI, and the country’s copper mining industry needs stable overnight power. A dispatchable clean resource that runs through the night fits that demand profile directly.

For context, global installed CSP capacity reached approximately 8.4 GW by the end of 2025, according to the Energy Institute Statistical Review of World Energy 2026, up from 6.7 GW at the end of 2023. Parabolic trough systems, an older design, still account for roughly 75% of the existing fleet, while tower technology represents the growth share of the new pipeline.

The IEA and NREL identify two grid conditions that most strongly favour tower deployment:

1. High curtailment of existing PV, where midday solar is routinely wasted.
2. Constrained transmission that prevents cheap PV from being exported elsewhere.

China’s 15 GW target, if achieved, would nearly double global installed CSP capacity from its current base. For the next five years, the Chinese domestic market is effectively the demand story for this technology, which is both its strength and its concentration risk.

The REN21 Global Status Report on CSP tracks China’s pipeline at 8.1 GW across various development stages, underscoring why the Chinese domestic market functions as the primary demand driver for tower technology through the end of the decade.

What tower CSP technology is built on, and where it can break

The mechanism works and the markets are real, but a reader who only knows the bull case is poorly positioned. The risks here are specific, quantifiable, and worth understanding before anyone builds an investment thesis on dispatchable solar.

The commercial track record carries real lessons. The US Department of Energy’s Solar One programme in California’s Mojave Desert served as the experimental foundation for tower technology, while its successor Solar Two confirmed that molten nitrate salt could function simultaneously as heat-transfer fluid and storage medium. At commercial scale, Crescent Dunes and Ivanpah produced the most significant operational data on receiver reliability and performance shortfalls, lessons that shaped the engineering approaches used today.

The three primary risk categories are worth holding as a checklist:

• Construction and commissioning: Tower projects are complex thermal infrastructure with long build cycles and notable commissioning risk.
• Water consumption and cooling: Wet-cooled plants demand substantial water in the arid regions where DNI is highest, and dry cooling avoids the water use at the cost of efficiency.
• Rare earth supply chain: High-temperature heliostat motors depend on dysprosium- and terbium-enhanced magnets, tying CSP to the critical minerals conversation.

Technology and operational risks

Build complexity is the first hurdle. These are sophisticated thermal systems, and the commissioning phase is where performance shortfalls tend to surface, as Crescent Dunes and Ivanpah both demonstrated.

Water is the second, and it is structural rather than incidental. The best DNI sites are deserts, where water is scarce and expensive. A wet-cooled plant consumes a lot of it; a dry-cooled plant avoids that demand but gives up efficiency. This is a site-level trade-off, not a solved binary.

Supply chain and policy risks

The rare earth exposure is not hypothetical. Dysprosium and terbium prices have historically been volatile, and a plant financed on today’s magnet costs carries real input price exposure across its asset life. Qualified heliostat component supply adds a further dependency.

Rare earth supply chain vulnerabilities that affect heliostat motor production are part of a broader pattern: dysprosium and terbium sourcing concentrations create input cost and availability risks across multiple clean energy technologies, not just CSP, and those risks have historically materialised as sharp price spikes rather than gradual trends.

Policy dependence is the systemic condition Project viability hinges on competitive auctions with firm capacity criteria, long-duration storage mandates, and fixed-price EPC contracts that enable project finance. Without procurement frameworks designed to reward firm capacity, the LCOE comparison against PV-plus-battery turns unfavourable at most durations.

A project in a market without a firm capacity procurement framework, secure water access, or a locked supply chain is a materially different risk than one that ticks all three.

What the pipeline signals for the next decade of clean infrastructure investment

The investment case for tower CSP is not a simple yes or no. It is conditional, and the conditions are specific enough that you can watch for them directly.

Three forces are converging on the demand side. Resource adequacy frameworks in an increasing number of markets are beginning to draw a formal distinction between firm capacity and energy-only resources. Chinese industrial policy is supplying the scale needed to push costs further down the trajectory already demonstrated. And auction design in MENA and Latin America is evolving toward explicit dispatchability requirements.

Capacity market design is the regulatory lever that determines whether dispatchable resources earn a price premium over energy-only generation; Spain’s 2026 ministerial order, which placed batteries and dispatchable generators inside a single remuneration framework for the first time, illustrates exactly the procurement evolution the CSP pipeline depends on.

The key is to treat this as a distinct infrastructure theme, not a slice of mainstream solar. Tower CSP is not competing with utility-scale PV for the same procurement dollars. It is addressing a different grid need, which makes it a separate thematic category for infrastructure investors.

The sceptical case deserves equal weight. At roughly 8.4 GW globally, CSP is dwarfed by PV at hundreds of gigawatts. Construction risk is real. And any scenario where battery costs fall faster than expected, or where grid operators decline to price firm capacity explicitly, is a direct headwind. The IEA frames CSP as a strategic but geographically constrained technology, a high-value niche rather than a mass-market theme.

NREL and industry analysts target the 0.05 to 0.08 USD/kWh range as the threshold for broad competitiveness in high-DNI markets.

The competitiveness threshold If Chinese projects drive LCOE toward the 0.05 to 0.08 USD/kWh band, CSP could compete even without bespoke firm capacity pricing. That is the number that would change the thesis from conditional to structural.

Three signals are worth monitoring as leading indicators:

1. Auction frameworks that explicitly price firm, dispatchable capacity.
2. China’s annual deployment progress against the 15 GW by 2030 target.
3. LCOE data from Chinese projects approaching the coal cost equivalence benchmark.

The five-year signal to watch is not CSP’s LCOE in isolation. It is whether new capacity auction frameworks in target markets explicitly price dispatchability, because that procurement design determines whether the economic advantage of molten salt storage translates into contracted revenue.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results, and financial projections are subject to market conditions and various risk factors. Forward-looking statements are speculative and subject to change based on market and policy developments.

A niche with structural teeth, in the right markets

Tower CSP with molten salt storage is not a general competitor to mainstream solar. It is a firm capacity solution for specific grid environments, and that specificity is a strength rather than a limitation for anyone who understands what they are procuring.

The advantage is durable in a defined set of conditions: markets where grid operators explicitly value dispatchable clean power, where direct normal irradiance is exceptional, and where policy frameworks reward evening peak delivery over energy-only production. Outside those conditions, PV-plus-battery usually wins, and that is the honest read.

The decade ahead will be shaped less by the technology than by how grid procurement rules evolve. Regulatory design is the primary variable to track. Watch the auction frameworks, watch China’s progress against its deployment target, and watch the cost curve.

This is a theme worth following in specific markets, under specific conditions. You now know which ones.