Your roof will hold three times more solar than your building can use
Warehouse solar works, and the standard advice to fill the roof is wrong for most sites. A median-profile warehouse uses about a third of what its own roof could generate, so the sizing decision is worth more than the price per kWp you negotiate.
An installer has walked your roof, come back with a number in kilowatt-peak, and it looks enormous next to your electricity bill. That gap is the whole project. Fill the roof and most of what you generate leaves the site at a fraction of what you pay to import. Size to your load and you leave two thirds of the roof empty but pay the investment back faster. This page shows you the arithmetic both ways, with UK prices, grid rules and tax as they stand in August 2026.
- Roof area is not the constraint. Daytime load is. The government benchmark for median warehouse electricity use is 26.59 kWh per square metre a year. A full roof generates roughly three times that.
- Budget £750 to £950 per kWp for a 100 to 500 kW warehouse array in 2026, and £700 to £850 above 500 kW. Under 100 kW you are paying £900 to £1,200.
- A third of commercial roofs need work first. Across 575 structural feasibility surveys run between 2024 and 2026, 35% were unsuitable without remedial work, though only 5% were structurally hopeless.
- Grid connection is the long pole, not the install. The DNO has 45 working days to issue an offer, but end to end most warehouse projects run 8 to 14 months, longer where the network is constrained.
- Solar does not qualify for full expensing. It is a special rate asset. AIA gives 100% up to £1m a year, above which you get a 50% first-year allowance, not 100%.
- The business rates exemption expires on 31 March 2035. Install in 2026 and you get nine years of it, not the twelve the policy started with.
The sizing decision, in numbers
You are here because the quote does not obviously match the bill, and that instinct is right. Work it through on a 5,000 square metre warehouse, which is a common mid-box size.
Gross roof area is 5,000 m². Usable area after rooflight banks, edge exclusion zones, plant and fire-access strips is 70% to 80% of that, so call it 3,750 m². At roughly 8 m² per kWp, the roof holds about 470 kWp. In a typical UK location that generates around 900 kWh per kWp, so roughly 420,000 kWh a year.
Now the building. The government’s median benchmark for warehouse electricity is 26.59 kWh per square metre a year. On 5,000 m² that is about 133,000 kWh. Your full roof generates more than three times everything the building consumes across a whole year, including nights and weekends when the panels produce nothing.
That gap matters because the two kinds of kilowatt-hour are worth completely different money. A unit you consume on site displaces an import at 22p to 28p. A unit you export earns whatever you negotiate with your supplier, commonly somewhere between 3p and 15p. The spread is roughly three to one, and every kilowatt-peak past your daytime load falls on the wrong side of it.
Two arrays on the same roof
Price both and the choice becomes clear. Size to load at 150 kWp and the cost at £950 per kWp is £142,500. It generates 135,000 kWh, of which perhaps 75% is used on site. That is 101,000 kWh saved at 25p, or £25,300, plus 34,000 kWh exported at 8p, or £2,700. Call it £28,000 a year against £142,500. Simple payback is 5.1 years, and 3.8 years after Annual Investment Allowance relief at 25% corporation tax.
Now fill the roof. 470 kWp at £800 per kWp is £376,000, generating 423,000 kWh. Self-consumption cannot rise much, because the building’s total demand has not changed, so you still save around £25,000. The extra 320,000 kWh is exported at 8p for £25,800. Annual benefit £50,800 against £376,000. Payback is 7.4 years, or 5.5 after relief.
Say that out loud in front of a finance director and the conversation changes. It is not a question of whether solar works. It is a question of which of two investments the business is signing up for, and only one of them is a hedge against your own electricity price.
What flips the answer
Three things justify filling the roof, and warehouses that have them are the ones producing the case studies you have read.
- Refrigeration or cold storage. Continuous compressor load through daylight hours, and consumption per square metre several times the median. Cold stores are the best rooftop solar sites in the UK, and it is not close.
- Automation. Conveyors, sorters and goods-to-person systems create a flat weekday daytime load that maps well onto a generation curve.
- Fleet electrification. Electric forklift charging is already a growing share of warehouse demand, and HGV or van charging changes the picture entirely. If depot charging is on your five-year plan, size for that load rather than today’s, and read our guide on solar for EV charging.
Battery storage is the fourth lever and the weakest of the four on its own. It converts export into self-use, but only where you have surplus in the day and demand at night, and warehouse night demand is usually thin outside cold storage. Price it separately rather than letting it ride along in a package. Our battery cost guide covers the per-kWh numbers.
What warehouse solar costs
Cost per kWp installed is the only figure that lets you compare quotes across different system sizes, and it falls sharply with scale because design, scaffolding, mobilisation and the grid application are broadly fixed whatever the array size.
| System size | Cost per kWp | Typical project value | Grid route |
|---|---|---|---|
| Under 100 kW | £900-£1,200 | £20k-£110k | G99, often fast-tracked |
| 100-500 kW | £750-£950 | £85k-£475k | G99 with connection study |
| 500 kW-1 MW | £700-£850 | £350k-£850k | G99, reinforcement possible |
| Over 1 MW | £650-£800 | £650k+ | G99, study and possible HV works |
Most warehouse projects land in the highlighted band, which is also where quotes vary most. A modern portal-frame unit with good roof access and spare switchgear capacity prices near the bottom of its range. A 1990s unit needing purlin strengthening, a new sub-main and a cherry picker prices near the top or above it.
Two line items are worth pulling out of any quote before you compare. Ongoing operations and maintenance runs roughly £7 to £15 per kWp a year, so a 470 kWp array costs £3,300 to £7,000 annually to keep under warranty and inspected. And VAT on commercial solar is charged at 20%, recoverable as input tax if you are registered, which is different from the 0% domestic rate and catches people who have read a household guide first. Our commercial solar cost guide breaks the components down further.
The roof: structure, age and fire
Assume the roof needs assessing before you assume it needs panels. Across 575 PV structural feasibility surveys conducted between 2024 and 2026, more than a third of commercial rooftops were found unsuitable for installation without some form of remedial work. The reassuring half of that finding is that only 5% were structurally inadequate outright, so most problems are fixable rather than fatal.
Commission the structural survey before you sign, not after, and make the quote conditional on it. A ballasted or clip-fix array adds dead load and changes wind uplift behaviour across the whole roof, and purlin capacity on 1980s and 1990s sheds was designed with very little spare. Finding this out after mobilisation is how a project acquires a five-figure variation.
Roof age is the other question, and it is the one installers ask least. Profiled steel cladding has a service life measured in decades, but a panel array is a 25-year asset bolted on top of it. If your roof has ten years left, you are choosing between replacing the cladding first or paying to strip and refit the array halfway through its life. Get the remaining life in writing from a roofing surveyor, and if the answer is under fifteen years, price the recover into the project. Metal roof mounting using clip-fix systems avoids penetrations entirely, which helps but does not extend the cladding’s life.
Fire and your insurer
This has moved from a footnote to a live underwriting issue, and warehouses are the buildings insurers worry about most. UK fire services attended 212 fires involving solar panels in 2025, up from 91 in 2022, according to freedom of information responses gathered from 46 of 49 fire services by QBE. DC cabling and connectors were the leading cause in 2025 at 49 incidents, nearly double the 26 recorded the year before. The 2024 fire at a Lidl distribution warehouse in Peterborough is the one the industry cites.
Tell your insurer before the contract is signed, not after commissioning. Underwriter guidance, including the RISC Authority joint code of practice RC62, recommends PV on roofs achieving A1 or A2-s1,d0 to BS EN 13501-1 where practicable, and AXA has warned specifically about arrays over combustible roof build-ups. Many older warehouses have composite panel roofs with combustible cores. Installing without disclosure is a material non-disclosure that can affect claims well beyond the array itself.
The practical mitigations are cheap relative to the exposure: an IEC 62446-1 commissioning report, surge protection on inverter inputs, a documented annual inspection regime, and arc fault detection on the DC side. Insurers treat rooftop PV as an ageing electrical installation rather than a fit-and-forget asset, and premiums and claims both reflect whether you do too. Our guides on fire safety labelling and solar insurance claims cover the paperwork.
Grid connection is the long pole
Every warehouse array needs G99 approval, since the threshold is 16 A per phase, roughly 11 kW on single-phase and 17 kW on three-phase. Nothing at warehouse scale escapes it.
Published timescales look contradictory until you see what each one measures. The DNO’s target is 45 working days to issue a connection offer, and that figure is real. Installers quoting 8 to 14 months are measuring something different: pre-application enquiry, full application, technical review, the offer, your acceptance, any reinforcement works, then commissioning with DNO witness testing. On constrained networks the same end-to-end figure runs 14 to 24 months. Both numbers are honest. Ask any installer which one they are quoting you.
The DNO can approve your system and still refuse the export you priced your business case on. Export limits are imposed where local network capacity is tight, sometimes at zero, and industrial estates in the Black Country, parts of Greater Manchester and parts of Yorkshire are known constrained areas. A 500 kWp array on a 250 kW export limit throws away generation on weekends and shutdowns rather than exporting it. Get an indicative capacity answer from a pre-application enquiry before you commit to a system size.
The 2025 and 2026 queue reforms cut the national connection queue from 722 GW to 283 GW by removing speculative applications, replacing first-come-first-served with a readiness test. For a viable warehouse project on an unconstrained feeder that is good news. It does not create capacity where there is none. Read our guides to G98 against G99, the DNO approval process and export limitation before you fix a capacity in a board paper.
Twelve months of HH meter data sets the system size. Structural feasibility runs in parallel.
Indicative capacity answer, often free. Do this before you approve capital.
45 working day target, commonly 12 to 20 weeks in practice. The critical path.
Landlord consent, if applicable, needs to be closed out here rather than later.
Two to six weeks on the roof for 300 kWp. Rarely the reason a project is late.
The array cannot be energised or export until this is signed off.
The tax position, done properly
Three reliefs apply and one of them is widely reported wrong, in a way that matters most at warehouse scale.
Capital allowances. Solar PV counts as an integral feature, which puts it in the special rate pool. That means it is excluded from 100% full expensing, and excluded from the new 40% first-year allowance introduced on 1 January 2026, both of which apply to main rate assets only. The route to 100% relief in year one is the Annual Investment Allowance, which is capped at £1m a year. Above that cap, special rate expenditure attracts a 50% first-year allowance, with the balance entering the special rate pool at 6% writing-down allowance.
For a £250,000 array that distinction is academic, because AIA covers it. For a 1.5 MW array at £750 per kWp, or a multi-site rollout, it is not. On £1.5m of spend you get £1m under AIA and £250,000 under the 50% FYA, so £1.25m of first-year relief rather than the £1.5m a full-expensing assumption would have put in the model. At 25% corporation tax that is a £62,500 difference in year-one cash. Pages still referring to Enhanced Capital Allowances are describing a scheme that closed in 2020.
Business rates. Eligible plant and machinery used in onsite renewable generation and storage is excluded from rateable value in England from 1 April 2022 to 31 March 2035, with an equivalent scheme in Wales. No application is needed, the Valuation Office Agency applies it. The catch is the end date. Installing in 2026 buys you about nine years of exemption, and any payback model running to year 25 should show the array becoming rateable from April 2035 unless the relief is extended.
The exemption is aimed at generation consumed on the same hereditament. That is another quiet argument for sizing to load: an array built predominantly to export sits less comfortably inside the exemption than one built to supply the building underneath it. If you are going large, take advice rather than assuming. Our capital allowances guide goes into the mechanics, and none of this is tax advice.
If you lease the building
Most UK warehouse space is leased, and the split incentive is the single most common reason a sound project never happens. The tenant pays the electricity bill and gets the benefit. The landlord owns the roof and takes the capital cost, the structural risk and the reinstatement question at lease end. Neither party alone has the full business case.
Three routes work in practice. The tenant funds and installs under a licence with an agreed reinstatement position at lease end, which suits tenants with long unexpired terms. The landlord funds and recovers through rent or a service charge, which suits institutional owners with portfolio decarbonisation targets. Or a third party funds it and sells power to the occupier under a PPA, which sidesteps the split entirely at the cost of a lower saving.
The lease term is the deciding variable. A tenant with four years left should not be buying a 25-year asset. A tenant with twelve years and a break at year eight should be negotiating the reinstatement clause before anything else, because an obligation to strip the array and make good at lease end can remove a third of the project’s value.
MEES, and what changed in June 2026
Landlords have a new deadline that pulls in the same direction. On 18 June 2026 DESNZ published its interim response to the 2019 and 2021 non-domestic MEES consultations. From 2031, privately rented non-domestic buildings over 1,000 square metres in England and Wales will need EPC B, where cost effective. Buildings under 1,000 m² stay on the existing EPC E minimum with no deadline. The previously proposed EPC C milestone for 2027 has been dropped.
Two things follow for warehouses. Almost every let warehouse is over 1,000 m², so almost every let warehouse is in scope. And because non-domestic EPCs are assessed on a carbon-based rating rather than a cost-based one, onsite generation improves the rating directly, which makes solar one of the more efficient routes to a B on a large shed with limited fabric options.
This is an interim response, not law. The EPC B requirement takes effect only once secondary legislation passes, and the cost-effectiveness qualifier and existing exemptions remain. Treat 2031 as a planning date rather than a settled obligation, and check the position again before committing capital on the strength of it.
Buy, finance or PPA
Buying outright produces the best lifetime return and needs the capital and the tax capacity to use the allowances. If the business is loss-making or the capital is better deployed elsewhere, the allowances are worth less than the headline suggests and the case weakens.
Asset finance or a lease spreads the cost and usually still allows the array to be treated as your asset, depending on structure. Check with your accountant which party gets the capital allowances before you sign, because that answer changes the economics materially and is not always what you assume.
A PPA puts zero capital in and typically sells you power at 30% to 50% below your standard commercial rate. Providers usually want at least 500 m² of roof or an equivalent car park. The trade is straightforward: you keep none of the export revenue, none of the capital allowances and none of the asset, and you sign a long contract with terms about roof access, reinstatement and what happens if you vacate. It is the right answer for occupiers who cannot deploy capital and the wrong one for owner-occupiers who can. Our financing options guide compares the routes.
When it does not work
Cold storage, food processing, or any continuous refrigeration load.
Automated fulfilment with flat weekday daytime demand.
Owner-occupied, or a lease with 15 years or more unexpired.
Roof with 20 years of cladding life left and confirmed spare structural capacity.
Depot EV or forklift charging on the plan within five years.
Unheated, unlit ambient storage with minimal daytime load.
Cladding with under ten years left and no recover budgeted.
Lease with under five years unexpired and a hard reinstatement clause.
A constrained feeder where the DNO offers a low or zero export limit.
Composite panel roof with a combustible core and no insurer engagement.
The unheated ambient store is worth naming plainly, because it is the case the sector’s marketing skips. A dark racking warehouse with LED lighting, a couple of dock levellers and a small office runs on very little. Its bill is small, so the saving is small, and the roof is enormous, so the temptation to fill it is large. That combination produces the worst payback numbers in warehouse solar, and it is common.
- kWp
- Kilowatt-peak. The array’s rated output under standard test conditions. UK sites typically yield 850 to 950 kWh a year per kWp installed.
- Self-consumption
- The share of generation used on site rather than exported. The single biggest driver of warehouse solar returns.
- G99
- The Energy Networks Association recommendation governing connection of generation above 16 A per phase. Approval is required before commissioning.
- Export limit
- A cap the DNO can impose on how much you may push onto the network. The inverter curtails generation above it, and that output is lost.
- Special rate pool
- The capital allowances pool for integral features, written down at 6% a year. Solar sits here, which is why full expensing does not apply.
- MEES
- Minimum Energy Efficiency Standards. The rules setting the lowest EPC rating at which a property may be let.
Common questions
How much roof do I need for 100 kW of solar?
Around 700 to 1,000 m² of usable roof, at roughly 7 to 10 m² per kWp. Usable area is typically 70% to 80% of gross roof area once rooflights, plant, edge zones and fire-access strips are excluded, so a 1,200 m² roof is about right for 100 kW.
What payback should I expect on warehouse solar?
Four to six years for a site with strong daytime load, such as cold storage or automated fulfilment, and seven to nine years for a weekday-only ambient warehouse with around 50% self-consumption. Annual Investment Allowance relief takes roughly a quarter off the net cost for a profitable company and shortens payback accordingly.
Do solar panels increase my business rates?
No, not until 2035. Eligible onsite renewable generation and storage is excluded from rateable value in England from 1 April 2022 to 31 March 2035, applied automatically by the Valuation Office Agency, with an equivalent Welsh scheme. The exemption is aimed at generation consumed on the same site.
How long does a warehouse solar project take?
Eight to fourteen months end to end on an unconstrained network, and fourteen to twenty-four where reinforcement is needed. The roof work is usually two to six weeks of that. The G99 connection process is the critical path in almost every case.
Can I install solar on a warehouse I lease?
Yes, with landlord consent, and the reinstatement clause matters more than the consent. Settle in writing whether you must remove the array and make good at lease end before you commit capital. With under five years unexpired, a PPA or a landlord-funded scheme usually makes more sense than buying.
Will my insurer cover a warehouse with rooftop solar?
Generally yes, but disclose it before installation rather than after. Underwriters are increasingly focused on roof fire rating, DC-side installation quality and evidence of ongoing inspection. Expect to provide an IEC 62446-1 commissioning report and a maintenance regime, and expect closer questions if the roof has a combustible core.
Size to the load, then decide about the rest of the roof
Warehouse solar is a good investment and the sector’s default advice oversells it, because roof area is the thing installers can see and daytime load is the thing that pays. Pull twelve months of half-hourly meter data before you accept a system size, and make the installer show self-consumption at their proposed capacity rather than gross generation.
Then decide deliberately about the rest of the roof. Filling it is a defensible choice if you have a DNO who will take the export and a supplier who will pay a decent rate for it, but understand you are buying a second, lower-return asset on top of the first one, not more of the same thing.
Three conditions precedent before capital approval: a structural feasibility survey, a DNO pre-application enquiry, and written confirmation from your insurer. Any of the three can change the project, and all three are cheap compared with finding out afterwards.
METHODOLOGY: Cost bands compiled from published UK commercial installer pricing across August 2026 and cross-checked for consistency. Consumption benchmark of 26.59 kWh/m² is the government median for warehouse electricity. Structural findings from 575 PV feasibility surveys conducted 2024-2026. Fire incident figures from QBE freedom of information requests answered by 46 of 49 UK fire services by April 2026. Worked examples assume 900 kWh/kWp annual yield, 25p import, 8p export and 25% corporation tax.
DISCLAIMER: Prices exclude VAT, which is charged at 20% on commercial installations. Figures are indicative and vary by site, roof condition, access and network position. Nothing here is tax, legal or financial advice. Capital allowances and MEES obligations depend on your circumstances and on legislation that is still changing.
