How big should a farm battery be?

The short answer

Size the battery to the daily surplus you cannot otherwise use, not to annual consumption. On most farms that lands well below what is proposed.

Stay under 1 MWh unless the duty genuinely requires more, because at a megawatt hour the fire guidance changes and demands a 30 metre standoff from occupied buildings and a substantial water supply. Both are very hard in a working steading.

And model the case on increased self-consumption alone. Everything else, the flexibility markets and the capacity payments, is upside at best and in one case is worth almost nothing at all.

Why a farm battery is not a big house battery

A house has one occupancy pattern, a peak of four to six kilowatts, and uses under 3,500 kWh a year. A farm differs on five axes at once, and they push sizing in different directions.

The single most important difference is the one people get backwards. On a house, generation and consumption are badly matched, and a battery is the fix. On many farms the mismatch is smaller than assumed, and cheaper fixes exist first.

The state of published guidance, stated plainly

We looked for a farm battery sizing methodology with numbers from NFU Energy, AHDB, the Andersons Centre and Defra. None of them publishes one.

The only farm-specific worked example we could find from a farming body uses a 12 kWh battery at about £7,000, which is a domestic-scale unit being used to illustrate farm storage. AHDB’s energy storage page is undated and cites a 2019 projection about what costs would be by 2023.

So when a client arrives with a payback figure for a farm battery, ask where it came from. There is very little in the public domain for it to have come from.

The sizing method

Not annual consumption. Not a rule of thumb per cow or per bird. This:

1. Take a clear summer day’s generation from the array, hour by hour.
2. Subtract the load happening at the same time, from the holding’s own half-hourly data.
3. What is left is the daily surplus. That is your sizing target.
4. Then discount it, because a battery sized to the best day of the year is oversized for the other three hundred and sixty.

Two sanity checks on the result:

• If the farm already achieves 70% or more self-consumption, be sceptical. There is very little surplus left to capture, and the battery is competing for a shrinking residue. This is the most commonly missed disqualifier.
• If the site is export constrained, the answer changes materially. See the section below.

Realistic scale bands

The 1 MWh line, and why it should shape the design

The National Fire Chiefs Council’s guidance, revised December 2025, applies to lithium battery systems of one megawatt hour or greater. At and above that threshold it sets out:

• 30 metres minimum between battery units and occupied buildings
• 0.914 metres between units where testing shows fire will not propagate between them
• 3.7 metre road width, 3.1 metre gateways and a 16.8 metre turning circle for fire appliances
• At least 25 litres per second at any hydrant on site, or roughly 180,000 litres of static supply lasting two hours

The 30 metre standoff and the water requirement are the two that stop farm schemes. Thirty metres of clear separation from any occupied building is very difficult in a working steading, and 180,000 litres is a serious piece of infrastructure.

So design at 800 kWh rather than 1 MWh wherever the duty allows. That single decision can be the difference between a straightforward installation and a scheme that needs a new water supply.

Below the threshold there is no statutory fire consultee. That does not mean there is no duty. It means nobody is going to catch a siting mistake for you, which is a good argument for engaging the local fire and rescue service voluntarily. It is cheap insurance.

What the revenue stack actually pays

This is where farm battery appraisals most often go wrong, because they are built from commercial battery marketing where the revenue stack is real.

Three of those deserve expanding.

The Capacity Market has stopped paying. The T-1 auction for 2026/27 cleared at £5.00 per kW per year and the T-4 for 2029/30 at £27.10. On a 100 kWh battery with 50 kW of power, heavily de-rated for its short duration, the T-1 figure is perhaps £50 to £100 a year before aggregator fees and metering costs. Reject any farm battery appraisal that leans on Capacity Market income. It is not a revenue stream, it is a rounding error.

Triad avoidance no longer exists. Ofgem’s charging review moved residual network charges to fixed charges, and a battery cannot reduce a fixed charge. Older models and some installer spreadsheets still include it.

Flexibility services are real but geographically specific. Network operators buy flexibility from assets as small as 10 kW, and they are not getting enough of it: Electricity North West contracted 42 MW against 157.9 MW tendered in one year. That shortfall is a bargaining position for a well-sited asset. But the value exists only where the network operator has a constraint in that zone. Check the published requirement zones before assuming any flexibility income at all. Outside one, it is zero.

There is one genuine improvement worth knowing. From 14 April 2026 the national Demand Flexibility Service cut its minimum entry from 1 MW to 0.1 MW and added paid turn-up, meaning payment for increasing consumption when there is surplus generation on the system. That opens the door to farm-scale assets for the first time, though the payment basis is not something to bank on.

Model the case on self-consumption alone. Treat the rest as upside.

Where a battery clearly does pay

On an export-constrained site with existing generation. This is the strongest case and it is routinely left out of installer models, because it requires knowing the client’s actual export limit.

Here is the logic. Under an export limit, surplus generation is not sold cheaply. It is thrown away, because the inverter is turned down. So a unit recovered by a battery is worth the full import price of about 29p, not the 24p difference between import and export on an unconstrained site. That roughly doubles the value of every unit stored.

Where the alternative is a large connection quote. A battery that defers or avoids network reinforcement is competing against a five or six figure quotation, not against a per-unit spread. Framed that way the numbers work far more often than a simple payback comparison suggests. Get the reinforcement quote before dismissing storage.

On a dairy with substantial solar and evening milking. Milk cooling is around a third of dairy electricity and milking another fifth, the evening peak sits squarely in the expensive network charging band, and the cycling is daily, predictable and year-round, which suits an appraisal.

On cold stores and packhouses. Continuous load, high absolute consumption, and thermal mass that allows genuine load shifting.

Where a battery does not pay

Arable with low seasonal load. Cereals farms sit predominantly below 50 kWh per hectare and 5% use no mains electricity at all. A battery sized for harvest drying sits idle for ten months and cannot accumulate the cycles to repay itself.

Farms already at 70% or more self-consumption. Very little surplus left to capture.

Broiler units treated as an electricity problem. A 27,000-bird house uses about 35 MWh of electricity against 240 to 270 MWh of heat. Ventilation load already correlates well with sunshine, so solar matches it without help. The money is in the heat, and a battery does not touch it.

Any farm that has not done the efficiency work first. AHDB’s work at Rhual Dairy identified £5,400 a year from heat recovery, variable speed drives and efficient compressors, against a £20,000 annual bill. That is a 27% reduction for a fraction of the capital, with none of the fire, insurance or replacement risk. Do this first, then re-measure, then consider storage.

Anywhere a generation-side fix is cheaper. The dairy at Kilmeaden in County Waterford installed a 27 kW bifacial vertical solar fence facing east and west. It matches the 7am and 5pm milking peaks directly, achieving what a battery would have been bought to do, at lower capital cost and with no degradation or replacement liability. Always test that alternative before recommending storage on a dairy.

What it costs, and the gap that disappoints people

Global turnkey battery system prices averaged around US$117 per kWh in 2025, and about US$177 in Europe, falling roughly 31% in a year.

Do not apply those figures to a farm. They are utility-scale, and the reason they are low is a product format effect rather than a volume discount: systems using large cells were around 50% cheaper than those using small ones, and 4 MWh containers were 39% cheaper per unit than smaller configurations. A 100 kWh farm cabinet accesses none of that. The gap between grid-scale pricing and farm-scale pricing is the single biggest source of client disappointment on this subject.

The most useful UK heuristic remains that the battery itself is only about half the installed cost. Budget roughly double the cell or pack price for a delivered, energised, connected system. On top of the battery you are paying for the power conversion equipment, G99 protection, export limitation hardware, switchgear, civils, fire measures and insurance.

Insurance runs 0.3 to 1.2% of total project value a year, with lithium iron phosphate priced more favourably than nickel manganese cobalt.

Specify lithium iron phosphate. Guidance produced for DESNZ states plainly that it is inherently more thermally stable and less prone to thermal runaway. Its lower energy density is irrelevant on a farm, where space is not scarce. Insurers price the difference, and so should you.

Appraisal mechanics

Four things that separate a real appraisal from an optimistic one.

1. Read which limit the warranty binds on. Commercial warranties state a term, a total throughput and a cycle count, whichever comes first. One solar cycle a day is about 365 cycles a year, so the ten-year term binds. Two cycles a day for arbitrage hits the cycle cap in five or six years. The revenue strategy determines which limit bites, and aggressive trading can void a warranty written for solar shifting.
2. Warranties typically guarantee 60 to 70% retained capacity at term. A battery sized to today’s surplus is undersized by year ten.
3. Appraise over ten years, not fifteen or twenty-five. Solar can carry a twenty-five year appraisal. A battery cannot. A twenty-year battery valuation with no replacement line in it is a red flag.
4. Budget replacement at roughly half the original system cost, because the inverter, switchgear, civils, enclosure and connection all persist. Against a falling cost curve, a year-twelve replacement is cheaper in real terms than the original purchase.

Before anyone signs

Notify the insurer in writing before installation. This is the one that causes real damage when it is missed. A battery installed in or against an existing farm building can invalidate cover on the building and its contents, not just on the battery. On a grain store or a livestock building, the contents value dwarfs the battery many times over.

Then:

• Confirm the chemistry is lithium iron phosphate and get it stated in the specification
• Ask whether the insurer requires a standalone enclosure rather than installation inside a building. Many now do
• Check the maintenance agreement, because both the premium and the insurability depend on it
• Ask for the fire test data for the specific product, not the manufacturer’s range
• Confirm business interruption cover extends to loss caused by the battery, not merely loss of the battery

And on a tenanted or mortgaged holding, settle before installation whether the battery is a landlord’s fixture, whether any third-party finance conflicts with the lender’s security, and what happens to it at the end of the term.

The order to do things in

1. Meter first. Twelve months of half-hourly data, per enterprise.
2. Do the efficiency work. It usually beats the battery return for a fraction of the capital.
3. Establish the export limit. If the site is constrained, battery value roughly doubles. If it is not, be sceptical.
4. Get the reinforcement quote, because that is what storage is really competing against.
5. Test the generation-side alternative, such as east-west or vertical orientation.
6. Check the network operator’s flexibility zones.
7. Size to daily surplus, then discount.
8. Stay under 1 MWh unless the duty genuinely requires more.
9. Model on self-consumption alone, and reject anything containing Capacity Market income.

Figures current as at 1 September 2026. Capacity Market clearing prices are from the March 2026 auctions and change annually. Flexibility market thresholds and values change more often than that, and are specific to each network operator’s constrained zones. Reviewed twice yearly.