2-hour or 4-hour battery: which storage duration pays off for investors

Anyone comparing project documents for battery storage comes across two configurations that look like the same product at first glance: 10 megawatts with 20 megawatt hours, or 10 megawatts with 40 megawatt hours. The grid connection is identical, the container yard is twice the size, the purchase price sits somewhere in between. Whether the second variant is worth it depends on what the third and fourth hour are worth on the market, what they cost to buy, and which items in the calculation count per megawatt rather than per megawatt hour. One step at a time.

What separates a 2-hour battery from a 4-hour battery?

Only the capacity, not the power. Both batteries can feed the same power into the grid; the 4-hour battery sustains it twice as long because twice as many cells sit behind the same inverter. How power and capacity relate, and what else a storage plant consists of, is explained in How does a grid-scale battery storage system work? Design, operation and revenue logic explained; this article is about which of the two configurations pays off economically.

Germany's existing fleet is considerably shorter than either variant. According to the Marktstammdatenregister, large battery storage systems (above 999 kilowatts) had around 2.4 gigawatts of power with a usable capacity of around 3.2 gigawatt hours in operation as at 15 October 2025, an average of roughly 1.3 hours. The projects registered as planned look different: 5.0 gigawatts with 10.4 gigawatt hours, an average of 2.1 hours. The fleet is getting longer, and the question of the fourth hour is part of every new project design.

There is a simple reason the early batteries are so short: they were built for balancing energy, which pays for power and needs little energy. The new projects target power trading, and there what counts is how much energy can be shifted from the cheap midday into the expensive evening.

Why does a 4-hour battery not cost twice as much?

Because only part of the plant grows with capacity. A storage project consists of two cost blocks that scale differently. The power block hangs off the megawatts: grid connection, substation, transformer, inverters, medium-voltage equipment, construction and engineering, land and permits. It costs practically the same at 10 megawatts with 20 megawatt hours as at 10 megawatts with 40 megawatt hours. The energy block hangs off the megawatt hours: cells, modules, containers, battery management, cooling and fire protection. It doubles when capacity doubles.

How the two blocks relate in a specific case is stated in the project's own cost calculation, and that is where it belongs. For the logic, one assumption is enough, and we state it openly: in a 2-hour system, under our model assumption, roughly half of the investment goes to the power block and the other half to the energy block. The third and fourth hour then cost the energy block once more.

That is the first half of the answer: a 4-hour battery costs about one and a half times as much per megawatt as a 2-hour battery, but around a quarter less per megawatt hour. Whether that pays off is decided by the other half: what the additional hours bring in on the market. Where the grid connection and transformer are shared with a solar plant, the power block shrinks further; that configuration is covered in Co-located vs. stand-alone: which gives the better risk structure.

How much more does a 4-hour battery earn on the power market?

Around 1.7 times as much per megawatt, around 15 percent less per megawatt hour. That is the result of our own model with few assumptions: for every calendar day since January 2024 we take the published hourly prices of the German day-ahead market (SMARD, Federal Network Agency), let the battery charge in the two or four cheapest hours and discharge in the two or four most expensive, once a day, at 86 percent efficiency and without fees.

The factor sits between 1.68 and 1.76 in every period, in the summer half of 2025 (April to September, around €44,000 against €75,000 per megawatt) just as in the winter half (around €25,000 against €43,000). It stays below 2 because the third and fourth hour of the day are paid worse than the first two: in 2025 the mean gap between the two cheapest and the two most expensive hours was around €116 per megawatt hour, between the four cheapest and the four most expensive around €100. Per megawatt hour of capacity, the 2-hour battery therefore reaches around €34,600, the 4-hour battery around €29,500. The additional hours bring less per megawatt hour, but they bring considerably more than nothing.

For context on the market: in 2025 the average day-ahead price according to the SMARD annual review was €89.32 per megawatt hour, it was negative in 573 hours and above €100 in 3,494 hours. What negative hours mean for a battery is shown in Negative electricity prices: what they mean for solar and storage investors.

What does that mean per euro invested?

For pure arbitrage, a slight advantage for the fourth hour. Putting the two halves together, you get roughly 1.7 times the market potential per megawatt for roughly 1.5 times the investment. Per euro invested, the 4-hour battery is therefore around 10 to 15 percent ahead of the 2-hour battery, as long as the cost assumption from the section above holds and the revenue comes from power trading alone.

This calculation is what lies behind the observation in Battery storage returns: where the revenue comes from, and what is realistic that 4-hour systems sit at the upper end of the return range in base-case scenarios: not because they earn more per megawatt hour, but because the additional megawatt hours are cheaper than the first ones. How a battery's revenue sources are composed and how they shift towards 2030 is described in Direct marketing explained: day-ahead, intraday and balancing power.

What role do balancing energy, grid fees and cycles play?

They change the denominator, and with it the answer. Three items in a storage project are calculated not per megawatt hour but per megawatt, and each of them shifts the comparison in a different direction.

• Balancing energy pays for power, not energy. For primary and secondary control reserve (FCR, aFRR), a battery receives a capacity price per megawatt and needs little energy capacity for it. A 2-hour battery earns the same here as a 4-hour battery, but cost less. Why these markets are small and fill up faster with every new battery is explained in Direct marketing explained: day-ahead, intraday and balancing power. For sizing this means: the larger the balancing-energy share in the revenue plan, the more strongly it argues for short systems, and the shorter its half-life.
• Grid fees will in future be charged per megawatt. Batteries commissioned by 4 August 2029 are exempt from grid fees for twenty years under §118 (6) EnWG. For projects without that grandfathering, the Federal Network Agency's AgNeS reform provides for a capacity-based grid fee from 2029, that is, a price per megawatt of connection capacity. Per megawatt hour, such a price weighs only half as much on a 4-hour battery as on a 2-hour battery. The details of the reform and of the grandfathering rules are in Grid fees for battery storage: the exemption until 2029 and the AgNeS reform.
• Cycles are the budget, not the years. At the same power, a 4-hour battery discharges at half the rate (0.25 C instead of 0.5 C) and needs half as many full cycles for the same amount of shifted energy. A 2-hour battery that runs two cycles on good days uses up its cycle budget twice as fast and reaches the limit of its capacity guarantee earlier. How cycles, degradation and guarantees relate is explained in The battery as a real asset: lifespan, degradation and warranties of a grid-scale storage system.

The three effects do not cancel out; each belongs separately in the project calculation. A project with a high balancing-energy share and commissioning before August 2029 calculates differently from a project that earns mostly on the spot market from 2030 and pays grid fees.

Does storage duration count in the new capacity market?

Yes, by statute. Since 22 July 2026 the Strom-Versorgungssicherheits- und Kapazitätengesetz (StromVKG, BGBl. 2026 I No. 210a) has been in force. It introduces auctions in which firm capacity is procured for the delivery period from November 2031; the European Commission approved the scheme under state aid rules on 2 September 2026, with estimated costs between €15.6 billion and €35.2 billion. For storage, the decisive point is that the Act classifies batteries as an energy-limited technology class and rates their contribution by their maximum delivery duration (Höchsterbringungsdauer), which is exactly the quantity this article is about.

In the first rounds this hardly matters, because batteries can practically not compete there. The auctions for long-term capacity on 8 September and 29 December 2026 (4.5 gigawatts each, §4 StromVKG) require under §12 (5) StromVKG plants that can feed in 80 percent of their installed power for at least ten consecutive hours and restore that ability at any time within three hours at most. Annex 4 of the Act accordingly assigns no de-rating factor (Reduktionsfaktor) to batteries below ten hours of delivery duration; at ten hours it is 0.58, at twelve hours 0.66.

The door opens for storage in 2027. On 18 May 2027, two gigawatts are auctioned under §5 StromVKG without the long-term criterion; on 1 December 2027 and 1 October 2029, technology-neutral auctions for power plants, storage and controllable loads follow under §6 StromVKG, open to existing plants and pooled assets as well. For these rounds §23 (3) applies: energy-limited plants receive de-rating factors depending on their maximum delivery duration, determined by the Federal Network Agency and published at least six weeks before the auction notice. These factors are not yet available. Only the direction is certain: the longer a battery can deliver, the more firm capacity it may bid, and the more capacity remuneration it can receive per megawatt.

For the sizing question this means: whoever plans a 2-hour battery today plans without the capacity market. Whoever builds four hours holds an option on a remuneration whose level will only be known once the Federal Network Agency's de-rating factors and the auction results are in. In a viability calculation that option therefore belongs as a scenario, not as a revenue line.

When is a 2-hour battery still the better choice?

More often than the calculation above suggests. Four situations argue for the short configuration:

• When revenues today come mostly from balancing energy. It pays for power, and the additional megawatt hours earn nothing there. That holds, however, only as long as those markets are not full.
• When the budget is capped. A direct investment is usually tailored to a specific investment sum, for instance because the investment deduction under §7g EStG is tied to limits. For the same money you then get more megawatts with two hours or fewer megawatts with four; which variant fits depends on the revenue plan, not on a rule of thumb.
• When the timeline matters. A 2-hour battery is built and commissioned faster, and commissioning before 4 August 2029 decides twenty years of grid-fee exemption. The dated windows are listed in Germany's battery storage market: the numbers, the drivers, and why "now" has a date.
• When the project is designed for later expansion. Inverters, land, foundations and grid connection can be sized for four hours from the outset while only the cells for two hours are bought initially. The later expansion (augmentation) then benefits from falling cell prices and the experience of the first operating years. Whether a project really has that reserve is written in the grid connection commitment and the permit, not in the presentation.

Conversely, little argues for two hours if the project will only be commissioned from 2030 anyway, earns mostly on the spot market and will pay grid fees per megawatt. Then the fourth hour is the cheaper part of the plant and the capacity market an additional option.

How do we check storage duration in a project?

As a calculation assumption, not as a feature. For a battery direct participation we first check whether power and capacity carry the same figures in the grid connection commitment, the building permit and the purchase agreement; a battery that has four hours in the brochure and two in the grid commitment has a problem that no return solves. Next we split the provider's revenue forecast into items per megawatt and items per megawatt hour and check whether the storage duration is applied consistently in both groups. Finally we run the sizing in reverse: what would the same project earn and cost with the other storage duration? If the answer differs sharply from the provider's recommendation, we want to know why. The other risks we probe are set out in Risks in BESS direct investments, and how they are structurally addressed.

Questions to ask the provider

1. Which power and which usable capacity are stated in the grid connection commitment, the building permit and the purchase agreement, and do the three match?
2. How is the purchase price split between the power block (connection, transformer, inverters, construction) and the energy block (cells, containers, cooling)? What does the additional megawatt hour cost?
3. Which revenue items in the forecast are calculated per megawatt, which per megawatt hour, and how many full cycles per year are assumed?
4. Is a capacity remuneration under the StromVKG included in the calculation? If so, with which de-rating factor, given that the Federal Network Agency has not yet published it?
5. Is the plant designed for a later capacity expansion, and where is that reserve documented: in the inverters, in the land, in the grid connection commitment?
6. Which capacity guarantee does the cell manufacturer give for the planned number of cycles, and does it apply to the optimiser's actual operating pattern?

Whether the storage duration of a specific project fits its revenue plan, grid connection and budget is something we work through against the actual documents in a no-obligation initial conversation, with assumptions named. We give no return guarantees.