100kW Commercial Solar Systems for UK Businesses

A 100 kW commercial solar install sits at a genuine engineering step-change rather than just a bigger version of a 50 kW system. You leave G98 connect-and-notify territory a long way behind and land in a full G99 Connection Application — which becomes the critical path for the whole project. The array needs roughly 600 m² of usable roof, a three-phase supply with real headroom, a verified structural assessment, and an inverter topology chosen deliberately rather than by default. Everything below is the specification and delivery detail: what goes on the roof, what the building has to be able to take, and how long each stage actually takes. Figures are drawn from PVSyst modelling against half-hourly meter data, not vendor brochure assumptions.

Who 100kW commercial solar fits

The 100 kW band lands cleanly across a specific group of UK businesses. Mid-format hotels (40–80 rooms) with year-round occupancy and steady daytime conference, kitchen and laundry load. Independent boarding and academy schools with extended-day operations, sports halls, kitchens and ICT load. Care homes and nursing homes above 40 beds where laundry, kitchen, cooling and lighting drive 24/7 daytime baseload. Light-industrial units of 1,000–2,500 m² housing engineering, food processing, plastics or assembly. Mid-format DIY, garden centres, B&M-style retail and supermarkets up to 1,200 m² sales floor. The common threads: roof footprint of 600+ m² usable, three-phase electrical supply with good headroom, a structure verified to take the array load, and daytime baseload of 25 kW plus so the generation is used on site.

What a 100kW system is made of

The kit specification at this scale is more substantial than the 50 kW band: typically 180–185 tier-1 mono panels (Trina Vertex, JA Solar Deep Blue, Longi Hi-MO 6), commercial-grade three-phase string inverters (Sungrow SG-CX, Solis-S6, Huawei SUN2000-100KTL or SMA Tripower CORE) usually configured as 2 x 50 kW or 3 x 33 kW for redundancy, a galvanised-steel mounting system, an MCS-compliant DC and AC switchgear assembly, surge and arc-fault protection, a dedicated PV consumer unit, a half-hourly monitoring stack with sub-string data, full G99 connection paperwork (including Type Test certificates, witness testing arrangements with the DNO and ENA G99 forms), and a chartered structural engineer’s sign-off. Scaffolding and access equipment are itemised. Asbestos work, three-phase headroom upgrades and structural reinforcement are always quoted as separate lines rather than buried in a headline figure.

Generation from a correctly orientated, unshaded 100 kW array in the UK runs 90,000–95,000 kWh in year one, tracking the 900–1,150 kWh per kWp per year band that UK irradiance supports — southern England at the top of it, Scotland at the bottom. For full 2026 pricing and payback see our 100kW solar system cost guide.

Panel count and roof area by module choice

Module wattage is the variable that sets panel count, fixing count and install labour. The kWp total barely moves — what changes is how many units your installer handles and how the array fits around roof furniture. All three rows below land within 0.4 kWp of a nominal 100 kW.

The gap between panel area and roof area needed is the part most owners underestimate. A 545 Wp module is about 2.58 m², so 184 of them is roughly 475 m² of pure panel — but you also need inter-row spacing, edge set-backs from the roof perimeter, walkway access for maintenance, and clearance around rooflights, AHUs, plant and vents. That overhead is why ~600 m² of usable area is the working rule for 100 kW. Usable is the operative word: a 900 m² roof with a large plant deck, three rooflight runs and a shaded northern strip may only offer 550 m² of viable array space.

Roof footprint, mounting and shading

South-facing pitched roofs are ideal but east-west splits work well on flat-roof commercial buildings — east-west arrays generate around 92% of an equivalent south-facing system but produce a flatter, longer daily curve that often matches business load profiles better. We model both layouts and present whichever performs better against your specific consumption pattern. Mounting at this scale is steel-framed ballasted on flat roofs (typically Renusol, Schletter or K2) or rail-mounted on trapezoidal sheets. Asbestos-cement industrial roofs are common on pre-2000 buildings and need licensed removal before install — we work with HSE-licensed contractors and frequently roll roofing and PV into a single project.

Shading is assessed properly rather than eyeballed. A single flue, parapet or neighbouring gable can drag a whole string down if the array is wired without regard for it, so shading analysis drives string layout and MPPT allocation, not just panel placement. Module-level power electronics are specified only where a genuine shading constraint justifies them — on a clean, unobstructed industrial roof they add cost and failure points for no yield gain.

Structural loading and ballast

This is the check that most often changes a design, and it must be verified rather than assumed. The load a 100 kW array imposes depends almost entirely on mounting type:

• Rail-mounted on trapezoidal sheet: roughly 12–15 kg/m² additional distributed load. The array is fixed directly to the purlins through the sheet, so the load path is straightforward and most industrial roofs absorb it without modification.
• Ballasted on flat roof: typically 25–60 kg/m² depending on wind zone, building height, parapet height and array position on the roof. Ballast resists wind uplift rather than holding the panels down by weight alone — perimeter and corner zones need materially more ballast than the field of the roof because uplift pressures there are far higher.
• Penetrating fixings on flat roof: lower dead load than ballast, but every penetration is a warranty and waterproofing question for the existing membrane. Used where ballast would exceed the roof’s capacity.

A chartered structural engineer assesses the existing structure against BS EN 1991 wind and snow loading, confirms the ballast schedule zone by zone, and signs the design off before any kit is ordered. Steel-portal-frame industrial units built after about 1990 usually pass comfortably. The buildings that fail are older asbestos-cement roofs, timber structures, roofs already carrying significant plant, and anything with existing corrosion or deflection. Where the roof cannot take the load, the honest answers are a lighter rail-mounted specification, a smaller array, a re-roof first, or ground mount if land is available — not a heavier ballast schedule signed off on optimism.

Three-phase supply and electrical headroom

A 100 kW array is a three-phase-only proposition. At 400 V across three phases, 100 kW of output is roughly 145 A per phase — single-phase simply is not viable at this scale, and a site without a three-phase supply needs a new connection before solar is even a conversation. At survey we check four things: the main incomer fuse rating, the supply transformer capacity, spare ways in the existing main switchgear, and the daytime baseload the array will actually offset.

Most commercial sites on a 200 A or larger three-phase supply have adequate headroom. Where they do not — typically sites already running close to their agreed capacity — the options are a supply upgrade (a DNO job with its own cost and lead time) or an export limit configured in the inverter, which caps what the system can push back to the grid without restricting self-consumption. Export limiting is frequently the pragmatic answer, because most 100 kW commercial arrays are sized to be consumed on site rather than exported anyway. Both routes are identified before the design is fixed, never discovered on install week.

G99 — why 100kW changes the timeline

At 100 kW you need a full G99 Connection Application (formally ENA Engineering Recommendation G99), approved by your DNO before the system can be commissioned. It is worth being precise about where the thresholds actually sit, because this is routinely misreported. G98 connect-and-notify — install first, tell the DNO afterwards — only covers up to 16 A per phase, roughly 11 kW on a three-phase supply. The G99 fast-track route runs to about 17 kW per phase, roughly 50 kW three-phase. A 100 kW array is comfortably past both, so there is no shortcut available: full application, DNO assessment, connection offer, witness test.

Standard DNO turnaround for the offer letter in 2026 runs 6–12 months, sometimes longer in constrained network areas (parts of London, the Midlands and the south-east). Most 100 kW G99 applications come back with a standard offer and no reinforcement charge; a meaningful minority attract a contribution towards transformer or upstream cable upgrades, and that contribution can be substantial enough to change whether the project makes sense at all. This is why the G99 application is the critical path and why we file it within two weeks of contract signature, run design and procurement in parallel, and time delivery to site for the week the connection offer is accepted. If reinforcement comes back high we re-engineer — export limiting, sometimes a modest downsize — and re-file before any capital is committed. The connection offer is the gate; nothing irreversible happens before it lands.

Worked specification — 100kW on a mid-size hotel

A representative modelled project, to show how the specification decisions above resolve on a real building shape. Site: a 60-room independent hotel in West Yorkshire, year-round occupancy averaging 72%, on-site restaurant and conference rooms. Electrical: three-phase 400 A supply with around 50 kVA of spare headroom — comfortable for a 100 kW array with no supply upgrade and no export limit needed. Roof: 720 m² of flat membrane in serviceable condition, of which 620 m² is usable once the plant deck, rooflights and perimeter set-backs are excluded. Demand: half-hourly meter data showing 380,000 kWh annual consumption with a daytime baseload of 28 kW.

Resulting specification: a 99.6 kW system using 184 x 545 Wp modules, 2 x 50 kW Sungrow SG50CX inverters, and K2 ballasted east-west mounting at 10 degrees. East-west was selected over south-facing tilted rows for two reasons — it packs into the available 620 m² without the inter-row spacing a south array would demand, and its flatter, longer generation curve tracks the hotel’s morning and evening load better. Ballast schedule came in at 34 kg/m² across the field of the roof with heavier corner and perimeter zones, verified against the existing structure by a chartered structural engineer. Modelled year-one yield: 91,200 kWh, which sits mid-band for UK irradiance at this latitude. Self-consumption modelled at 74% — around 282 kWh per day on average, well matched to that 28 kW baseload — with the balance exported. G99 was the critical path at just under eight months from filing to accepted offer; physical install ran 12 working days. The full PVSyst yield model, meter analysis and financial appraisal ship with every proposal. Pricing, tax treatment and payback for a system of this shape are set out on our 100kW solar system cost guide.

How 100kW systems get paid for

Three routes are viable at this scale: outright purchase with capital allowances, asset finance over five to seven years, or a PPA on a no-upfront-cost basis. 100 kW is roughly the smallest scale at which PPA economics start to make sense at all. Which route wins depends on your corporation tax position and how much working capital you want to keep in the business, and that is a genuinely per-company answer rather than a rule of thumb.

Because the numbers are the whole story here, they live on the pages built for them rather than being duplicated across the site. For full 2026 pricing, capital allowance treatment and payback modelling see our 100kW solar system cost guide; for a route-by-route comparison see finance options. We model all three against your specific accounts as part of every proposal.

Where 100kW lands hardest by sub-vertical

Across the 2025 UK commercial solar market, the 100 kW band overweighted heavily across five sub-verticals: mid-size hotels and conference venues (year-round daytime demand, kitchen and laundry load), independent and academy schools (extended-day operations, ICT load, kitchens), care homes above 40 beds (24/7 demand including laundry and cooling), light-industrial units 1,000–2,500 m² (single-shift manufacturing, food processing, packaging), and mid-format retail and supermarkets (chiller load, lighting, tills). Common threads: 600+ m² of usable roof, a three-phase 200 A or larger supply with headroom, a structure that takes the array load without modification, and daytime baseload comfortably above 20 kW so the generation is consumed on site rather than exported.

Install timeline — contract to commissioning

The single most useful thing to understand about a 100 kW project timeline is that the installation is not the long part. Physical work on site is around a fortnight. The G99 connection process is what sets the calendar, and it runs for months — which is why everything else is deliberately sequenced to happen inside that window rather than after it.

Contract to commissioning typically lands at 5–9 months end to end. Stages overlap: procurement, design freeze and site preparation all run inside the G99 window, so the elapsed time is close to the G99 duration rather than the sum of every stage. Generation begins the day the witness test is signed off.

The 100kW survey process

Survey runs in two passes, same as smaller installs but with a heavier feasibility step. Desk feasibility uses your half-hourly DCP228 export, satellite and Lidar roof modelling, draft G99 grid feasibility against the local DNO's heat map, and a full PVSyst yield run — turnaround five working days. If the desk feasibility shows worthwhile IRR (we don’t take projects forward where 25-year returns fall below a genuine threshold), we visit on-site for structural survey, asbestos register check, electrical infrastructure assessment and roof condition inspection. Final fixed-price proposal lands within seven working days of the site visit, with the full meter analysis, yield model, financial DCF and grid feasibility narrative attached.

Inverter strategy at 100kW — string versus central

Inverter selection at 100 kW is more nuanced than at smaller scales, and it is one of the few specification decisions on this page that a building owner should actively have an opinion about. We typically specify two or three commercial-grade three-phase string inverters rather than a single central inverter for three reasons. First, redundancy: a single 100 kW inverter failing takes the entire system offline; two 50 kW or three 33 kW inverters mean a single failure costs you a half or a third of generation while the failed unit is replaced. Second, MPPT granularity: split arrays facing east-west, or arrays with awkward shading, benefit from multiple maximum power point trackers — more MPPTs across multiple inverters typically delivers 3–5% more annual yield than a couple of MPPTs across a single central unit. Third, lead times: 50 kW inverters from Sungrow, Solis and SMA typically have stock availability of 4–8 weeks in 2026, whereas larger central units routinely run 16–20 weeks, slowing project delivery.

We model the options and present the comparison in the proposal. Sometimes a single central inverter is the right answer — a large, uniform, single-orientation array with no shading and a client who values the simpler maintenance footprint. Usually a multi-inverter string topology wins on availability, resilience and yield.

Monitoring and what happens after commissioning

A 100 kW array generates enough that quiet underperformance is expensive and invisible without proper instrumentation. Monitoring is half-hourly at string level as standard — not a single whole-array total, which is where faults hide. If one string of 20 modules drops out on a 184-module array, whole-array monitoring shows an 11% dip that looks like weather; string-level monitoring shows exactly which string stopped and when.

The stack logs generation, self-consumption, export and per-MPPT yield, with alerting on string dropout and inverter fault codes. Both your facilities team and we can see it. That matters for a reason beyond maintenance: it is what makes the original yield model auditable. The PVSyst figure in your proposal is a prediction, and string-level half-hourly data is how you hold it to account in year three. Practical upkeep at this scale is modest — inverters are the component most likely to need replacement within a 25-year system life, panels degrade at roughly 0.5% per year under standard warranties, and cleaning is worth doing on low-pitch arrays or dusty sites but is rarely economic on a steeper roof in UK rainfall.

Cabling, switchgear and electrical infrastructure

The electrical work on a 100 kW project is more substantial than most owners expect. DC cabling runs from each panel string through MC4 connectors, into combiner boxes, to the inverter — typically 80–120 metres of DC cable per string with surge protection devices and arc-fault detection on every string. AC cabling from inverter output to the customer's switchgear typically runs 30–80 metres depending on inverter location, sized to accommodate the full inverter output current with adequate voltage drop margin (we target sub-2% voltage drop end-to-end). At the point of connection we install a dedicated PV consumer unit with main isolator, AC isolator, kWh meter for SEG export reporting, and protective devices. If your existing main switchgear has limited spare way, we add a tap-off enclosure, itemised separately on the quote rather than absorbed into a headline figure. Earthing and bonding is upgraded to suit the additional DC array, with TT or TN-C-S earthing systems handled per BS 7671. SPD (Surge Protection Device) coordination across DC, AC and signal circuits is standard — protects the inverter and the wider building electrical system against lightning-induced transients. None of this is exotic; it's just thorough.