Four-Hour Batteries Now Beat Gas Peakers on Cost in 43 Markets. Here's What That Comparison Leaves Out

Wood Mackenzie says four-hour battery storage now has a lower levelized cost of electricity than new open-cycle gas turbines in all 43 markets where both were modelled. Ahmed Jameel Abdullah, a principal analyst at the firm, called the shift “decisive and widening.” The finding is real as a statement about the firm's model, and it will be quoted as if batteries have beaten gas. They have not, and the firm does not quite say so. The comparison is between a battery and a specific kind of gas plant, the peaker, using a specific cost metric. Understanding what that metric includes and omits is the difference between a useful result and a misleading headline.

What was actually compared

Levelized cost of electricity, or LCOE, expresses the lifetime cost of building and running a plant per unit of electricity it delivers, here in US dollars per megawatt-hour. An open-cycle gas turbine peaker is a plant designed to run for a small number of hours a year when demand is highest. That matters for the arithmetic: a plant that runs rarely spreads its capital cost over few megawatt-hours, so its LCOE is dominated by what it cost to build and is very sensitive to utilisation assumptions. The release does not publish the capital costs, hours of operation, charging costs or financing assumptions behind either side. The comparison is also between storage and open-cycle peakers, not combined-cycle plants or baseload generation, and only for four-hour batteries.

LCOE does not capture everything a grid needs. A four-hour battery can deliver its rated output for four hours and then must recharge, so it cannot cover a multi-day shortage of wind and sun the way a gas plant with fuel on hand can. A gas turbine provides capacity for as long as fuel is available, while a battery's contribution to reliability tends to shrink as more batteries are added to the same grid, since they are all emptied by the same evening peak. Fuel security, grid connection, siting and the cost of replacing degraded cells over time are other things the headline cannot summarise. None of this means the result is wrong. It means a cost-per-megawatt-hour comparison answers the question of which source is cheaper per unit under stated assumptions, not whether one can replace the other.

The numbers, and where each comes from

Two different counts of 43 appear in the release and are easy to confuse. The storage-versus-peaker finding covers the 43 markets in which both technologies were modelled. A separate finding, that single-axis tracker solar is the lowest-cost new-build technology, covers 43 of 48 modelled markets, with onshore wind leading in the other five. They are different claims about different sets.

• Middle East and Africa: four-hour storage at $120 per MWh in 2026, forecast to fall 33% to $80 by 2035; single-axis solar at $37, falling to $24 (Wood Mackenzie release).
• Asia Pacific: the rest-of-region storage average is $134 per MWh, forecast to reach $92 by 2036; China's benchmark is more than 55% below that average (release).
• United States: four-hour storage is 65% to 75% cheaper than new open-cycle peakers for a 2026 commercial operation date, depending on state carbon pricing (spokesperson email to Utility Dive, not the report).
• Europe: battery turnkey capex rose about 2% in 2026, the first increase in three years, and is forecast to fall 12% by 2031 (release).

Some comparisons follow from these figures, and they are our arithmetic. China's benchmark being more than 55% below $134 puts it below roughly $60 per MWh. Middle East and Africa storage at $120 is about 10% below the rest-of-Asia-Pacific average, although regional models can differ in method and the release does not say they are directly comparable. The US range is the starkest: if storage is 65% to 75% cheaper than a peaker, then the peaker costs between about 2.9 and 4 times as much, a gap that depends on carbon pricing and is not in the published report. A single wide range from a spokesperson is a pointer, not a measurement.

Why gas got more expensive

The release attributes part of the shift to gas rather than batteries. Abdullah said gas turbine shortages and rising fuel volatility are driving up peaking costs. According to Utility Dive, Wood Mackenzie projects turbine prices will reach $600 per kilowatt by the end of 2027, up 195% since 2019, a figure from an April 2026 report by the firm rather than the October analysis. A 195% rise means the price has nearly tripled; working backwards, the 2019 level was roughly $200 per kilowatt. Utility Dive also describes order backlogs at GE Vernova, Siemens Energy and Mitsubishi of 35 to 116 gigawatts, but the attribution of figures to individual manufacturers is ambiguous, so we cite only the range.

This feeds directly into peaker economics. Because a peaker's LCOE is mostly capital cost, a near-tripling of turbine prices pushes its cost up almost in proportion to the equipment share of the build, while a battery's cost trend runs the other way in most regions. Wood Mackenzie also says North American gas investment is entering a supply deficit cycle through the late 2030s, driven by data-centre load growth, which it expects to keep thermal capital costs high. In that sense the story is about two things at once: batteries getting cheaper in most markets, and turbines getting scarcer and more expensive because the same data-centre demand is competing for them.

What cuts the other way

The release is not uniformly optimistic about storage. In the United States, tariffs, anti-dumping and countervailing duties and new Section 232 import restrictions are pressuring near-term solar costs, and storage benefits from tax credits that partly offset restrictions on foreign entities of concern and supply constraints. Wood Mackenzie expects US storage costs to fall 10% by 2060 and warns of a cost spike after the investment tax credit phases out from 2038. Europe's rise in battery capex shows the downward trend is not guaranteed in any year. These points are why a one-line headline about 43 markets can carry a different message from the report's regional detail.

Abdullah also framed the result as something broader than a contest between technologies: “It is about storage and solar together redefining what the economics of a power system look like.” That framing is more defensible than a claim that batteries replace gas plants. Pairing solar with storage addresses the evening peak that peakers were built to serve, but the evidence offered here does not address weeks of low renewable output or the cost of running a system with very little firm capacity.

Do batteries replace gas power plants?

On the evidence presented, batteries are cheaper than new open-cycle peakers for the specific job of covering a few hours of peak demand, in the markets modelled and under the firm's assumptions. That is a meaningful result for developers deciding what to build, and for data-centre operators facing turbine waiting lists. It is not evidence that storage can substitute for the full range of services gas plants provide, and the release does not claim it.

What to watch

• The paywalled report's market list, capital cost assumptions, hours of operation, charging-energy pricing and financing rates, none of which are public.
• Whether turbine prices reach the $600 per kilowatt projection and how long delivery backlogs last.
• US tariff and import-restriction decisions affecting storage and solar equipment.
• The path to the 2038 end of the investment tax credit and what happens to storage costs after it.

What the evidence establishes is a firm's modelled result for one technology pair on one metric, with supporting regional forecasts and an explanation of why gas peaking costs have risen. It does not establish that batteries can replace gas for reliability, that the result holds in the markets not named, or that the US range will survive publication of the underlying assumptions.