Commercial solar at 1 MW and above.

The fact that defines this band is that the grid stops being a formality and becomes the project's pacing item. Below a few hundred kilowatts the DNO connection is usually a managed step. At a megawatt it can be the longest lead time and the largest single cost variable on the job, and it often dictates the design rather than following it.

Three things activate together at this scale:

• Connection and capacity. Every commercial array runs under ENA Engineering Recommendation G99. At this scale you are typically a Type B generator and frequently connecting at high voltage (11 kV), which means a formal application, network studies, and mandatory witness testing before the DNO will let you energise. A larger array can need new switchgear, a cable run or a substation, and in a constrained part of the network it can need reinforcement. Recent reform of the connections process has shifted the queue toward a readiness basis, so projects that can show planning progress and land control tend to be prioritised. The practical implication is simple: start the grid conversation first, because it sets the calendar.
• Export limits shape the design. Where full export capacity is not available, a limited-export or zero-export connection under ENA G100 lets the project proceed at an agreed maximum export limit instead of waiting for capacity. That pushes the design toward self-consumption and storage, because any surplus the site cannot use or store is curtailed. Designing for what you consume, not what the roof could theoretically generate, is the right instinct at a megawatt.
• Network charges become material. Transmission (TNUoS) and distribution (DUoS) charges are a real line on a large site's energy bill, and the way they fall on half-hourly sites has been changing in recent charging reviews. Solar offsets daytime import directly; pairing it with battery storage lets a site shift load away from the peak DUoS red band and the winter demand peaks that drive these charges. At this scale the charging position is part of the business case, which is exactly why a megawatt project should never be costed as a panel count alone.

Because all of this is live at once, the genuine question at a megawatt is often not only what to buy but whether to buy at all. A power purchase agreement (PPA), where a developer funds, owns and maintains the array and you pay for the electricity it produces, is a real alternative to outright purchase here. The trade-off between owning the asset and buying the power is a board decision, and we model both so you can take it on evidence.

At a megawatt and above we do not publish a price per kilowatt-peak, and you should be wary of any installer who does. The reason is structural: at this scale the cost is driven less by the modules and more by the things around them, and those vary site by site.

What actually moves a megawatt-scale budget:

• The grid connection. The single largest variable. A straightforward connection is modest against the whole; a high-voltage connection needing new switchgear, cabling or a substation, or reinforcement in a constrained area, can run to a substantial share of the project. This is why we open with the DNO, not the roof.
• Rooftop versus ground-mount. A roof array reuses an existing structure but depends on its load capacity, age and layout. A ground array needs racking, foundations, fencing, cable trenching and civils, and the land itself. The two routes have genuinely different cost shapes, and we cost the one that fits your site.
• Modules, inverters and the DC and AC system. Large commercial modules and string or central inverters, plus the DC strings, AC distribution, switchgear, metering and protection. Volume helps here, which is where scale economics show up.
• Mounting, access and structure. Structural survey and any strengthening on a roof; ballast or penetration design; scaffold, edge protection and access for a site of this size. Larger jobs carry more access and safety provision, not less.
• Storage, where it earns its place. Battery storage is frequently bundled at this scale, both to lift self-consumption against an export limit and to manage network-charge peaks. It is a design decision with its own return, modelled separately.

The scale economics are real and run in your favour, but only directionally: as the system grows, the cost for each kilowatt-peak tends to fall, because fixed elements such as design, mobilisation and some grid and access works spread across more capacity. That is a true statement of direction, not a number, and the connection cost can pull against it on any given site. We will not put a figure on it before we have seen yours.

For the indicative engineering picture at 1 MW, using commercial rules of thumb: roughly 1,665 to 1,850 modules at 540 to 600 W panels (about 1.7 to 1.85 modules per kWp); roughly 5,000 to 7,000 m² of usable flat roof per MW once row spacing is allowed, against roughly 2 to 2.5 hectares for the same megawatt as a ground array; and a modelled annual output around 850,000 to 950,000 kWh per MW on a Yorkshire yield of roughly 850 to 950 kWh per kWp installed. Treat every one of these as indicative and bracketing only. The figures that go in a board paper come from your measured roof or site, your half-hourly demand and a PV*SOL study, never from a table.

For payback, internal rate of return and bill-reduction modelling at this scale, we keep those figures on the pages built to carry them properly, with their assumptions stated: see our commercial solar finance options and our commercial solar ROI and payback guide. The only fixed third-party figure worth noting here is the Annual Investment Allowance: capital expenditure on solar qualifies for 100% relief in the year of purchase up to the £1m AIA cap (shared across a group), with the balance entering the special-rate pool. Tax treatment is for your accountant; we point to the mechanism, not the outcome.

A megawatt-plus system suits organisations with either very large daytime loads or very large amounts of space, and ideally both. The economics turn on self-consumption: every kilowatt-hour you use on site displaces grid electricity at your full import rate, while exported surplus earns far less and, against an export limit, may not leave the site at all. So the strongest candidates are the ones that consume what they generate during the day.

• Very large logistics and distribution. Big-shed estates and major distribution parks have the roof area to host a megawatt and, increasingly, the chiller, conveyor, charging and handling loads to absorb it. The roof is already there, which makes the rooftop route the default to test first.
• Cold and chilled storage, and food production. Refrigeration and process loads run hard through daylight hours, giving high self-consumption and a strong fit with a large array plus storage to ride network-charge peaks.
• Heavy manufacturing and process sites. Steady, high daytime demand is close to the ideal load shape for solar. At this scale storage and demand management add a second layer of value through the charging position.
• Ground-mount on owned land. Where roofs cannot host a megawatt but you hold the land, a ground array on roughly 2 to 2.5 hectares per MW is the route. It carries more civils and a more involved connection, and it depends on grid capacity at the point of connection, which is why we model the site before committing to it.
• Multi-site portfolios. Estates, retail and industrial portfolios and public-sector campuses often reach a megawatt across several roofs rather than one. One coordinated programme, designed and procured together, beats a scatter of separate jobs.

If your site is large but your daytime demand is not, that is not a dead end; it changes the route. A storage-led design, an export-limited connection, or a PPA where a developer monetises the roof can still make the space work. The job at this scale is to match the system to the load and the grid, and we model your half-hourly data to find which of those routes fits before any commitment.