When the Sun Sets but the Grid Doesn't: How Solar-Plus-Storage Is Quietly Eating the Peaker Market

For nearly half a century, the open-cycle gas turbine has been the unloved but indispensable workhorse of the modern power grid. It does not run often — perhaps two to six percent of the hours in a year — but when it does, it earns its keep. The gas peaker is the unit that keeps the lights on at 6:47 p.m. on a still August evening when air conditioners are at full song and the wholesale price is sprinting past four hundred dollars a megawatt-hour. For grid operators from ERCOT to the United Kingdom's National Grid ESO, peakers have been the answer to the question: what runs when nothing else can?

That answer is changing, and it is changing faster than almost anyone in the industry forecast even three years ago. According to BloombergNEF's most recent figures, global energy storage deployments are on track to reach roughly 158 gigawatts in 2026, a 41 percent increase over 2025's record 112 gigawatts. In the United States alone, federal data show that solar and battery assets together will account for nearly 80 percent of all new utility-scale capacity additions planned for the grid this year. The numbers are no longer marginal. They are the headline.

What is happening, in mechanical terms, is straightforward. A four-hour lithium-iron-phosphate battery, charged on cheap midday solar electrons, can now discharge into the evening peak at a delivered cost that — in sun-rich, gas-expensive markets like California, Texas, and increasingly Spain and Australia — undercuts the marginal cost of running an existing gas peaker, let alone the levelized cost of building a new one. The asset that wakes up at sundown is no longer a turbine. It is a battery cabinet.

Three reinforcing trends pushed the inflection point forward. The first is the collapse in lithium-iron-phosphate cell pricing. After a brief inflationary detour in 2022 and 2023, average installed costs for utility-scale four-hour storage systems have fallen below $245 per kilowatt-hour in 2026, with leading developers reporting bids closer to $210 in competitive auctions. LFP chemistry, less energy-dense than its NMC cousin but cheaper, safer, and far more cycle-tolerant, has captured roughly four-fifths of new utility orders. It is, in plain terms, the right chemistry for the job the grid actually needs done.

The second is the maturing of the co-located solar-plus-storage project model. Where early storage projects were sited and contracted in isolation — paid for ancillary services, frequency response, or merchant arbitrage — the dominant project archetype today couples a solar farm and a battery behind a single point of interconnection. The solar arm captures the renewable PPA or the production tax credit; the battery arm captures the evening price spread; and the combined asset is dispatched as a single optimized portfolio. Operators in Texas's ERCOT market are reporting internal rates of return in the high teens on co-located projects where the solar arm alone would have struggled to clear cost-of-capital hurdles.

The third trend is the most consequential and the least-discussed: gas peakers are getting more expensive to keep around. New-build combustion turbines now face turbine-supply backlogs of three to five years (a function of the sudden global rush to gas-fired capacity for AI data-center load), insurance premiums that have hardened materially since 2023, and an increasingly hostile permitting environment in coastal and urban load pockets. The implicit option value of a battery — built in eighteen months, sited on a few acres, modular, and politically uncontroversial — has risen even faster than the explicit value of its arbitrage spread.

The result is a quiet but unmistakable rewriting of the dispatch stack. ERCOT's own operational data show that during the evening ramp on most spring and summer days in 2026, batteries — not gas — are now the price-setting unit. California's CAISO reports the same pattern. The United Kingdom's Balancing Mechanism is fast converging. The peaker is not dead. But it has, for the first time in its long career, a credible challenger that is younger, cheaper, and growing exponentially.

The strategic implications are sharpest for three constituencies. For independent power producers with peaker fleets, the question is no longer whether to retrofit, sell, or run the units to depletion — it is which combination of the three, and on what timeline. A peaker with a remaining technical life of fifteen years is, increasingly, a depreciating financial asset rather than a depreciating physical one. The forward capacity revenues that historically backstopped these assets are being competed away in auction after auction by bidders quoting four-hour and now eight-hour battery durations.

For power traders, the change is even more fundamental. The marginal flexibility resource on most modern grids is no longer thermal; it is a state-of-charge decision made by an optimization algorithm. The skill that pays in 2026 is forecasting nodal price spreads in five-minute blocks and dispatching storage against them — closer, in temperament and toolset, to a fixed-income desk than to a traditional commodities pit. Several of the largest European independent power traders have, in the last twelve months, reorganized their desks around exactly this principle.

For the upstream gas complex, the news is more nuanced than the headlines suggest. Peaker displacement reduces the number of running hours gas turbines accumulate; it does not, in the near term, materially reduce the installed capacity of gas the grid still requires for multi-day weather events, winter reliability, and the residual seasonal mismatch that no four-hour battery can plausibly cover. The longer-duration storage technologies — iron-air, vanadium flow, sodium-ion, thermal-rock — are advancing rapidly but are not yet at the cost or scale to displace gas's seasonal role. The gas peaker's twilight is real. The gas system's twilight is not, at least not on this decade's clock.

What is unmistakable is that the boundary between the renewable sector and the conventional generation sector is dissolving in front of us. The asset that competes with a gas turbine in 2026 is, increasingly, a solar farm with a battery bolted on. The asset that competes with a battery is, increasingly, a different battery with a different chemistry. The grid is becoming a single, integrated optimization problem — and the operators, regulators, and investors who internalize that earliest will have a meaningful head start over those who continue to treat solar, storage, and gas as separate industries.

The sun still sets every evening. The grid, for the first time in its history, no longer has to notice.