For two decades the case against renewable power rested on one word: intermittency. The sun sets. The wind drops. Grid operators need firm capacity, and only coal, gas, or nuclear could supply it. That argument collapsed this year, and it collapsed on cost, not ideology.
Solar-plus-storage delivers 24/7 electricity at $54–82/MWh in prime resource regions, below new coal in China and new gas almost everywhere.
Firm costs have fallen from above $100/MWh in 2020, with another 30% expected by 2030 and 40% by 2035.
The binding constraint has moved from generation economics to grid access, permitting, and project finance.
The benchmark behind that claim is new. IRENA calls it the firm levelised cost of electricity — the cost of delivering power on demand, not merely generating it. It adds storage, overbuild, and complementary generation to the familiar LCOE. The result is the number a buyer of 24/7 power actually pays.
That distinction matters for anyone underwriting an asset. A low LCOE tells you a plant is cheap to build and run. A low firm LCOE tells you the electricity is there when the contract says it will be. The second number is the one that competes with a gas turbine.
The number that ended the reliability argument
IRENA's 2026 study models co-located solar PV, onshore wind, and battery storage against a 95% reliability target. At that threshold, solar-plus-storage in high-quality regions delivered firm power at $54–82/MWh in 2025. The same report puts new coal in China at $70–85/MWh and new gas above $100/MWh almost everywhere.
The gap is structural, and it inverts the old trade-off. Renewables used to win on energy and lose on capacity. Now they win on both, provided the resource is good and the storage is sized correctly.
IRENA's Director-General, Francesco La Camera, put it plainly. "The long-standing argument that renewables lack reliability no longer holds," he said. The claim is no longer rhetorical. It rests on a project-level cost model that any developer can replicate.
The economics of the entire energy system have shifted: the battery revolution has driven down costs while accelerating advances in storage.— Francesco La Camera, Director-General, IRENA
The scale of the move is easy to miss because the headline number looks modest. A range of $54–82/MWh sits close to wholesale power prices in many markets. That is the point. Firm renewables now compete with incumbent generation inside the same price band.
What the firming premium buys
Firm LCOE is built from two parts. The first is the standalone cost of generation. The second is the firming premium — the extra capital needed to make output reliable. That premium is where the engineering happens.
Batteries handle daily variation. They are excellent at shifting midday solar into the evening peak. They are poor at covering a multi-day wind lull. Beyond roughly 85% reliability, the marginal cost of adding batteries alone rises sharply. At 99% and above, storage-only designs become economically irrational.
The fix is portfolio design. Pair wind with solar to smooth seasonal and diurnal gaps. Add modest overbuild. Then size storage for the residual. IRENA finds that overbuild cuts firming costs substantially, because it reduces the storage capacity needed to hit the same reliability target.
The market evidence is already in the price. Firm wind-plus-storage ranged from about $59/MWh in Inner Mongolia to $88–94/MWh across Brazil, Germany, and Australia in 2025. By 2030, those same markets are projected at $49–75/MWh.
| Benchmark | Firm solar + storage | Firm wind + storage | 2030 outlook |
|---|---|---|---|
| China, best sites | ✔ ~$30/MWh at 90% reliability | ◐ ~$59/MWh (Inner Mongolia) | Lower |
| UAE, Al Dhafra | ✔ ~$70/MWh for 1 GW firm | — | Lower |
| Brazil, Germany, Australia | ◐ $54–82/MWh | ◐ $88–94/MWh | $49–75/MWh |
| New coal, China | ✗ $70–85/MWh | — | Flat |
| New gas, global | ✗ Above $100/MWh | — | Fuel-linked |
Firm LCOE estimates at 95% reliability unless noted · IRENA, 2026
The table hides as much as it shows. Each row is a modelled project, not a market average. Resource quality dominates the spread. A site with strong wind and strong sun can carry a smaller battery for the same reliability, and that difference compounds across a portfolio.
Why China sets the cost floor
China defines the bottom of the global cost curve, and the reason is scale, not subsidy. IRENA analysed 252 utility-scale solar PV projects commissioned in 2024. A majority can already deliver firm electricity below $100/MWh. The best sites reach about $30/MWh at 90% reliability, and roughly $46/MWh at 99%.
That last figure is the one to watch. It says reliability is no longer the expensive part. At China's best sites, moving from 90% to 99% firm costs about $16/MWh. A gas peaker cannot compete with that for baseload replacement.
Outside China, the picture is uneven but converging. Saudi Arabia's firm solar is approaching parity with gas-fired generation even where fuel is cheap. Brazil, India, South Africa, and Australia are all moving toward fossil parity. The Gulf has the resource. Europe has the demand. The United States has both, in patches.
The policy lever matters more than the technology now. Markets that prequalify hybrid plants, publish firm-capacity rules, and shorten interconnection queues are converting cost advantage into deployed capacity. Markets that do not are leaving the cheapest megawatt-hours on paper.
Al Dhafra: the proof case
The United Arab Emirates' Al Dhafra complex is the clearest working example. It pairs 5.2 GW of solar PV with 19 GWh of battery storage to deliver 1 GW of firm clean electricity at around $70/MWh. That is a utility-scale plant, not a pilot.
The configuration is instructive. The solar array is roughly five times the firm output. The battery is sized for hours, not days. The system sells reliability, not just electrons. That is the product data-centre operators and industrial buyers are now shopping for.
IRENA's report is explicit that these systems are positioned to serve the most demanding electricity users, including AI and data-centre operators that require power without interruption. The commercial logic is straightforward. A firm renewable contract hedges both fuel price and carbon price. A gas contract hedges neither.
Compare that with the alternative now being signed across the United States. Gas plants paired with batteries are winning contracts because they can be permitted and built inside the load-growth window. They are not winning on lifetime cost. They are winning on time-to-power.
The investor's reframe: technology risk becomes grid risk
For a decade, renewable investment carried technology risk. Would the panels degrade? Would the batteries last? Those questions are mostly settled. Panel degradation curves are well understood. Battery warranties now run 15 to 20 years, and installed battery costs fell about 93% between 2010 and 2024, from $2,634/kWh to $197/kWh. Industry data cited by IRENA suggests another 30% drop in 2025 alone, to roughly $140/kWh.
The risk has migrated. It now sits in interconnection queues, permitting timelines, and offtake contracts. Construction is fast — typically one to two years from securing permits and grid connection. A new gas plant takes longer in most markets. Speed is now a renewable advantage, not a liability.
That is why the grid, not the generator, is the bottleneck. As we wrote in October, grid-forming capability has moved from a vendor feature to an interconnection condition in several markets. The inverter is now a regulated component of system stability, and the cost of compliance is becoming part of the firming premium.
Firm LCOE below $50/MWh at the best sites, now projected for 2035.
Interconnection reform and prequalification rules for hybrid plants.
Battery pack prices pushing toward $100/kWh.
Corporate 24/7 clean-power contracts with hourly matching.
Hourly matching is the quiet change in procurement. A company that buys annual renewable credits can claim clean power on paper while consuming gas-fired electricity at night. A firm contract removes that gap. For buyers with Scope 2 and Scope 3 commitments, the shift from annual to hourly accounting turns reliability into a reporting requirement.
That is where the data-centre demand meets the cost curve. A hyperscaler needs power around the clock, and it needs the supply to be defensible to auditors and regulators. Firm solar-plus-storage now satisfies both conditions at a price below new gas in the best locations. The remaining obstacle is land, grid, and time.
What would break the thesis
The benchmark has limits. Firm LCOE is a project-level metric. It does not include transmission upgrades, curtailment risk, or the system-level cost of balancing a grid with a very high renewable share. Those costs are real, and they vary by market. A project can be cheap and still fail if the grid cannot absorb its output.
Three things could slow the crossover. The first is land and permitting. A hybrid plant needs more acreage than a gas turbine of equal firm capacity, and permitting is local, slow, and political.
The second is cannibalisation. As solar penetration rises, midday prices fall. A plant that earns most of its revenue from energy arbitrage loses margin as more capacity arrives. Firm contracts insulate against this, but only if the offtaker accepts a fixed price above the merchant curve.
The third is the 99.9% problem. IRENA's model targets 95% reliability by default. Some loads — semiconductor fabs, hospitals, certain AI clusters — need more. At those levels, storage alone is insufficient. Gas, nuclear, or long-duration storage still has a role, and the economics shift back toward dispatchable thermal capacity.
None of these reverses the direction of travel. They set the pace. The cost curve is now on the side of firm renewables, and the constraint is institutional. That is a very different investment problem from the one the sector faced five years ago.
The worst energy crisis in decades has exposed the true cost of fossil fuel dependence. But another path is now possible. Renewable power is increasingly the most affordable, reliable and secure option.— António Guterres, United Nations Secretary-General
What to underwrite now
The practical question for an investor is not whether firm renewables are cheap. It is where the cheap megawatt-hour can actually reach a buyer. That means grid capacity, not generation cost, should anchor diligence.
Three checks matter. First, does the project have a firm interconnection agreement, or only a queue position? Second, is the offtake contract structured around reliability, or around energy volume? Third, does the hybrid configuration have a demonstrated firming premium, or is it a paper model?
IRENA's benchmark gives a reference price. It does not give a project. The gap between $54/MWh and a bankable asset is filled by grid access, contract structure, and execution. That gap is where returns are made and lost.
The reliability argument is over. The grid argument has just begun.