Solar carport vs rooftop solar: which to install first

Once an organisation has accepted that on-site solar pays for itself, the next decision is where to put the panels: on the roof of the building, or on a steel canopy built over the car park. The two are not interchangeable, and they are not rivals. Rooftop solar reuses a surface you already own, so it is cheaper per kilowatt and pays back faster. A solar carport builds the surface first, which costs more, but it does jobs a roof cannot: it shelters vehicles, it carries EV charging, and it turns dead tarmac into generation. This guide compares the two honestly on the terms that actually decide it, and sets out which to install first.

The honest headline: rooftop is cheaper, a carport does more

There is no getting around the cost gap. Rooftop commercial solar lands at roughly £700 to £1,050 per kWp installed, because the roof already exists and already bears load. You pay for panels, inverters, mounting rails and labour, and very little structural work. A solar carport runs about £1,100 to £1,600 per kWp at scale, and £1,200 to £3,000 per kWp on smaller schemes, or somewhere around £6,000 to £12,000 per parking space. That makes a carport roughly two to three times the cost per kWp of an equivalent rooftop array.

Those figures are quoted excluding VAT, which is how a commercial buyer should read them: the 20% is recoverable for a VAT-registered business. The 0% rate you may have seen advertised applies only to residential installations, and only until 31 March 2027, after which it reverts to 5%. It does not apply to a commercial car park.

Where the cost gap comes from

The whole difference is the structure. On a carport, the steel — columns, beams, purlins and the frame — plus the foundations account for roughly 40 to 50% of the total project cost. Rooftop solar avoids that entirely, because the roof is already there. That single fact explains the price gap, and it also explains why carports only make financial sense at scale. The steel cost is broadly fixed per bay, so a large scheme spreads it across far more generating capacity than a small one. A ten-bay canopy is poor value. A fifty-bay-plus canopy is genuinely competitive once you count everything it returns.

Generation is roughly the same per kilowatt

Per kWp, the two technologies generate very similarly. A UK yield of around 900 to 950 kWh per kWp per year is a reasonable national planning figure for both, before you adjust for your latitude, orientation and shading. On a carport, plan around 2 kWp per standard bay, which produces somewhere in the region of 1,700 to 2,000 kWh per bay each year. There is no single guaranteed number, because output depends on your site, but the point stands: the roof does not generate more per panel than the canopy. What differs is how much capacity you can physically fit. A roof gives you whatever the roof gives you once rooflights, plant and walkways are deducted; a carport is sized from the parking footprint.

The install is a different job

This is where the two routes diverge most, and it is worth understanding before you compare quotes, because you are not buying the same thing twice.

Rooftop: reuse what is already there

A rooftop install is comparatively light. The structural survey confirms the roof can take the additional dead load and wind uplift, the mounting system clamps or ballasts to the existing structure, and the array ties into the building’s supply. On most commercial roofs the biggest constraints are age and condition — you do not want a 25-year array on a roof with ten years of life left — plus the usable area that survives once access zones and existing plant are removed. But there is no groundwork and no new structure to build.

Carport: a structure first, solar second

A carport is a construction project with a solar array on top. It needs foundations, usually piled or mass-concrete bases sized to the ground conditions; a fabricated steel frame engineered to Eurocode wind and snow loads; and drainage for what is now a new impermeable surface. Only then does the PV go on. It has to be built over a live or phased-out car park, which brings traffic management and reinstatement. This is why lead times are longer, and why you want a single installer holding structural steel, groundworks and MCS-certified PV together rather than three trades pointing at each other. The full sequence is set out in how a solar carport is installed.

The DNO connection is the longest item on either route

Both routes usually meet the grid the same way, and it is the slowest step. Anything up to 3.68 kW per phase can use the G98 fit-and-inform process. Above that, which covers almost every commercial system of either type, you need G99 pre-approval from your Distribution Network Operator before you energise, and that commonly takes 8 to 16 weeks. On a well-run project the DNO application is the critical path, not the steelwork, so it goes in early on both a roof and a carport.

Planning: prior approval vs permitted development

Rooftop commercial solar is usually permitted development, the simplest route. A carport in England sits under Class OA of the permitted development order, in force since 21 December 2023. That is a prior-approval process, not full planning: the council can only consider siting, design and glare, and it must determine the application within 56 days. It is a standing right, not a time-limited offer, so ignore anyone who tells you to rush a decision before it closes.

There are conditions and limits worth knowing. The canopy must be no more than 4 metres high and more than 10 metres from a dwelling, and the right excludes listed buildings and scheduled monuments. Managing surface-water run-off through a sustainable drainage (SuDS) strategy is a mandatory condition of the right, not one of the matters the council weighs at prior approval, but a requirement you must satisfy regardless. Class OA is an England measure: in Wales and Northern Ireland a car-park canopy generally still needs full planning permission, decided case by case, while Scotland has its own permitted-development route. One more England point — a Class OA canopy must be completed within three years of approval. The planning permission guide covers the detail.

Where the carport earns its premium

If a carport costs two to three times as much per kWp, why build one at all. Three reasons a roof cannot answer.

First, EV charging. Retail, employer, healthcare, leisure and education car parks face growing pressure to provide chargepoints, and that infrastructure is funded whatever you do with the solar. A canopy carries the cable routes and a behind-the-meter solar supply for those chargers, so you dig the car park once and feed the sockets with daytime generation at your own cost rather than full grid price. The sockets themselves can be part-funded through the Workplace Charging Scheme: £500 per socket, up to £2,000 per socket for state-funded education, capped at 75% of the cost and at 40 sockets, through an authorised installer, until 31 March 2027.

Second, shelter. A carport keeps vehicles out of frost, hail and glare, a genuine benefit for staff, customers and fleet vehicles that a roof array simply does not provide.

Third, it works when the roof does not. Plenty of commercial buildings have a roof that is too old, too cluttered, too shaded or too small, and on those sites the car park — covering, in some cases, the entire site frontage — is the only solar surface left.

Tax and funding treatment

On the tax side the two are treated the same, and it is worth being precise because the sales patter is often wrong. Solar PV is a special-rate (integral-features) asset, so it does not qualify for full expensing. It does qualify for the Annual Investment Allowance, which gives 100% relief on up to £1m of qualifying spend, and companies can claim the 50% first-year allowance on special-rate additions above that. Business rates on the solar plant are exempt in England until 31 March 2035, though that relief does not extend to Wales or Scotland. None of this is advice for your specific accounts, so take your own tax advice before you budget on it. The grants and funding guide sets out what stacks with each route.

So which do you install first?

The rule is simple: fill the roof first where you can.

The roof wins when you have a sound, well-oriented, unshaded roof going spare. It is cheaper per kWp and pays back in four to seven years, against eight to ten for a solar-only carport. It also wins when the car park is small, because the fixed steel cost cannot be spread across enough capacity to justify itself.

A carport wins when the roof is full, unsuitable, shaded or too small, and the car park is large, say fifty bays and up. It wins when you have to provide EV charging anyway, because the canopy funds infrastructure you were building regardless. And it wins on a site with strong daytime demand, where most generation is self-consumed at full price instead of exported cheaply. That self-consumption, plus any EV revenue, is what pulls a carport’s payback in from ten years towards eight.

For a lot of sites the honest answer is both: fill the roof for the cheapest kilowatt-hours, then add a carport over the car park for the capacity the roof cannot hold and the EV charging you have to deliver. They are complementary, not competing. Where a car park shares a grid connection with an existing rooftop array, the only extra step is confirming spare DNO capacity before you size the canopy.

An illustrative comparison

To make it concrete — and these figures are illustrative, not a quote — take a distribution site with a large, sound warehouse roof and a 120-bay staff and visitor car park. The roof takes a 300 kWp array at roughly £700 to £1,050 per kWp, pays back inside about five to seven years, and should be the first move. The car park could later carry a roughly 240 kWp canopy at £1,100 to £1,600 per kWp with twenty EV sockets beneath it. Solar-only, that canopy pays back nearer eight to ten years, but the roof was always going to be full, the EV charging was coming anyway, and the sockets draw footfall and self-consume the generation. Roof first, carport second. Your own numbers turn entirely on roof condition, bay count, load profile and tariff.

If you want a straight answer for your own building, we will model the roof against the car park on your real numbers — request a feasibility and we will tell you which to install first, and whether it is worth doing both.