Ground-mount or rooftop? The 2026 decision tree for UK farms

Most UK farms with serious capital to deploy on solar face a fundamental choice: rooftop PV on the farm buildings, or ground-mount on land — or some combination. The answer isn’t always the same, and the question keeps changing as DNO capacity tightens, lease offers from third-party developers move, and SFI 2026 brings new agrivoltaics actions. Here’s the 2026 decision framework.

When rooftop wins

Rooftop PV is typically the right answer when:

1. High on-site electrical baseload — dairy parlours, intensive livestock, grain stores during drying season, polytunnels with supplementary lighting. Rooftop PV self-consumes at grid retail tariffs (24–28p/kWh) which is far more valuable than the 8–15p/kWh of SEG export.

2. Suitable roof condition — post-1995 farm buildings with profiled steel, membrane, or solar-suitable roof types in good condition. Asbestos cement roofs need re-cladding first (see our combined re-roof + PV piece).

3. DNO capacity for self-consumption design — even on capacity-constrained networks, rooftop PV sized for self-consumption typically connects faster than ground-mount, which usually needs significant export capacity.

4. Limited or no surplus land — most farms have buildings; not all farms have marginal land suitable for ground-mount lease.

5. No appetite for long-term lease commitments — third-party ground-mount developers typically want 25–40 year leases, which some family farms aren’t willing to commit on the back of generational uncertainty.

When ground-mount wins

Ground-mount makes sense when:

1. Marginal land available — pasture below productive grade, awkward field corners, former quarry land, low-fertility ground. Rule of thumb: if the land yields less than half-average for the farm’s main crop, it’s a ground-mount candidate.

2. Limited on-farm electrical baseload — sheep-finishing farms, beef-finishing, some arable holdings without grain drying. With low baseload, rooftop PV self-consumes only a small fraction of generation; ground-mount lease offers better income because the developer assumes the export risk.

3. Third-party developer lease available — current 2026 market rates for ground-mount lease (where DNO capacity exists) are £900–£1,300 per acre per year on 25–40 year terms. That’s typically 2–4× the arable rental income of the land in question. Many farms in capacity-constrained rural areas can’t get a developer to commit at any rate because DNO grid is the binding constraint.

4. SFI biodiversity stacking opportunity — SFI 2025 introduced more agrivoltaic-compatible actions. Sheep grazing under raised panels is well-established and stacks with SFI biodiversity bundles. Pollinator margins around the array stack further.

5. Strong supermarket Scope 3 motivation — large arable suppliers to retailers like Tesco and Morrisons increasingly reference ground-mount as material in supplier sustainability disclosure.

The hybrid case

For mid-to-large farms (above 800 acres), the hybrid case — rooftop on buildings with strong baseload, ground-mount on marginal land — is becoming the standard. Typical hybrid project: 200–400 kW rooftop on the grain store and workshop (self-consumed), 250–800 kW ground-mount on 5–15 acres of marginal pasture (third-party lease or owner-developer). A single G99 application can cover both. A single monitoring portal manages both.

Hybrid economics are typically better than either standalone option: the rooftop captures the high-tariff self-consumed kWh, while the ground-mount captures the lease income on land that wasn’t productively used. The land continues to graze sheep year-round (under raised panels with 3.5m+ clearance), and the farm gets a long-term diversified income stream.

The planning picture

Rooftop on agricultural buildings: typically Permitted Development under Class A Part 14 GPDO 2015 — no planning permission required. Listed buildings, AONB locations, and very large installations may need consent.

Ground-mount under 9m × 9m × 4m: Permitted Development on agricultural curtilage.

Ground-mount above 9m × 9m × 4m: full planning permission required. AONBs, National Parks, Conservation Areas, and proximity to listed buildings all add complexity. Above 1 MW, Environmental Impact Assessment typically required.

For most farms considering serious ground-mount (above 100 kW), the planning timeline is 6–12 months with full consultation, ecological assessment, and conservation officer engagement.

The DNO picture

Rural DNO networks are increasingly capacity-constrained, particularly for export. The 2026 picture:

• Western Power Distribution / National Grid Electricity Distribution: Generally capacity available but with regional hotspots constrained. Export capacity quoted 6–18 months for ground-mount above 100 kW.
• UK Power Networks: Generally better capacity in East Anglia and London/Home Counties than further east in Essex.
• Northern Powergrid: Yorkshire and the North East generally available; some Wear Valley and Tees Valley constraints.
• Scottish Power Energy Networks: Cheshire and the North West face capacity pressure in specific areas.
• Electricity North West: Manchester and Lancashire have export constraints in industrial corridors.
• SSEN (Scottish and Southern Electricity Networks): Mixed picture — south Scotland and the south of England have variable capacity, with some Highland feeders entirely closed to new generation above 50 kW.

For ground-mount projects, DNO connection is usually the binding constraint. Submit the G99 application early — sometimes years before the actual install if developing a major project.

SFI 2025 and agrivoltaics

SFI 2025 introduced several actions compatible with ground-mount agrivoltaics:

• Pollinator margins: £589/ha/year for unmown margins around the array
• Skylark plots: £124/ha/year for nesting plots
• Hedge management: scheme rates per metre for the surrounding hedge maintenance
• Soil health bundles: various rates for soil-improving actions under the array

These stack with the ground-mount lease income. A typical 10-acre ground-mount installation can deliver £12,000–£14,000/year in lease income plus £3,000–£5,000/year in stacked SFI actions — total £15,000–£19,000/year for marginal land that previously yielded perhaps £400/acre in arable rental income.

Our recommendation

For most UK farms with serious capital to deploy:

1. Map every farm building’s roof — area, orientation, condition, on-site load nearby
2. Map every parcel of underused land — yield record, DNO grid proximity, AONB/landscape status
3. Pull half-hourly meter data for every farm building’s main supply
4. Model rooftop, ground-mount, and hybrid scenarios in parallel
5. Compare returns side-by-side — payback, IRR, lease income, SFI stacking
6. Plan a 2–5 year capital programme rather than a single-project decision

The right answer depends on the specifics of the farm. We deliver feasibility studies covering all three scenarios for every working farm we work with — and the hybrid case is winning more often than not in 2026.