G99 explained: the grid connection process for commercial solar

Every commercial solar system big enough to matter has to clear a hurdle most people ignore until it slows their project down. The local electricity network operator has to agree, in writing, that you can connect it. For anything above the smallest rooftop scale, that hurdle has a name: Engineering Recommendation G99. Understanding how it works is one of the most useful things a facilities manager, developer or finance director can do before committing capital.

What G99 is, and why it applies to your project

G99 is the Energy Networks Association standard for connecting generating equipment “in parallel” with the public electricity distribution network. In plain terms, it covers how a solar system that can push power back onto the grid gets signed off. That sign-off comes from the company that owns the local wires and substations — your Distribution Network Operator (DNO).

The dividing line between the simple and less-simple routes sits at the boundary with G99’s smaller sibling, G98. G98 covers the smallest, fully type-tested installations — up to 16A per phase, around 3.68kW single-phase or 11kW three-phase. Below that threshold, installers can largely notify the DNO after the work is done.

Step above that threshold, and you’re in G99 territory. Almost every genuinely commercial system does. In that case, a formal application has to be approved before you connect, not after.

Retained UK regulation classifies generation by size and connection voltage into four categories — Type A, B, C and D — under the Requirements for Generators framework. For a Great Britain connection point below 110kV, Type A runs from 0.8kW up to 1MW. Type B and above cover everything from 1MW upward.

In practice, that split maps neatly onto how the ENA organises its own guidance. It publishes a dedicated guide for Type B–D generation aimed squarely at “commercial, industry, developers and farms” — the audience this article is written for.

Category	GB capacity band	Typical commercial example	What it means in practice
Type A	0.8kW – under 1MW	Most rooftop systems from around 100kW up to just under 1MW	Standard G99 application; the DNO’s assessment is usually more contained
Type B, C, D	1MW and above	Large rooftop, ground-mount, multi-building estates	A full technical assessment of network impact is standard, not exceptional

The process, stage by stage

Skip the acronyms and the G99 journey breaks down into four stages that repeat, in some form, on every project:

1. Initial enquiry and feasibility check. Before anything is submitted formally, it’s worth asking your DNO an informal question: is there headroom at the nearest substation for a system of this size? Every DNO now publishes some form of network capacity map for exactly this reason. A five-minute check here can save months later.
2. Formal application. The generator (or their adviser) submits the DNO’s standard application form. It includes inverter details, MCS certification where relevant, single-line diagrams and the site’s electrical layout. This is the point the clock formally starts.
3. Technical assessment. For anything beyond the smallest systems, the DNO runs its own study of what your connection would do to the local network. It checks voltage, fault levels, protection, and whether the existing infrastructure can absorb the extra generation without reinforcement. This is the stage that decides whether your connection is straightforward or expensive.
4. Connection offer and acceptance. The DNO issues a formal offer setting out the terms, any required network works, and their cost. You then have a defined window to accept it before it can lapse.

None of this is exotic. The same sequence applies whether the applicant is a 250kW warehouse roof or a 5MW ground-mount array. But the stakes of stage three rise sharply with system size — and that’s exactly where rural sites tend to run into trouble.

Why rural headroom is the quiet risk in this process

The single biggest practical risk in G99 isn’t the paperwork. It’s discovering — after design work, and sometimes after a tender has already gone out — that the local network doesn’t have room for your system. That often means upgrading a piece of infrastructure the developer never sees or controls.

Engineers designed rural and semi-rural networks to move power outward from substations to dispersed, relatively light demand. They never had to absorb large amounts of generation flowing back the other way. As a result, a site a short distance from a well-loaded substation can face a materially different — and more expensive — connection outcome. A near-identical building on a better-connected part of the network might not, simply because of where the cables happen to run.

That’s why we treat the DNO enquiry as a feasibility-stage question, not a procurement-stage one. Check headroom before you sign a lease, issue a tender, or specify a system size for a farm building’s roof. That way, you price any reinforcement cost into the decision from the start. You won’t discover it as an unwelcome addition once you’re already committed.

Realistic timelines

DNOs work to a statutory response window once a complete application lands. But “complete” is doing some work in that sentence: an application missing detail doesn’t start the clock. Instead, the DNO sends it back for more information.

Beyond the formal response period, timelines vary widely. From a first enquiry to a usable, accepted connection offer, the practical range commonly runs from a few months — where headroom is already available — to the best part of a year, where a technical assessment flags reinforcement. Where reinforcement works are required, construction of them adds further time again before energisation.

It’s also a moving target. In April 2025, Ofgem approved a package of grid connection reforms intended to replace the historic “first come, first connected” queue. The new “first ready, first connected” model prioritises projects that are actually deliverable over ones simply holding a place in the queue.

The reform mainly targets the transmission-level backlog rather than routine rooftop G99 connections. However, it signals the direction DNOs are heading. It’s worth checking whether your DNO’s published timescales have moved since you last looked.

Building G99 into the decision, not around it

None of this changes whether solar is worth doing on a given site. But it does change when you find out. A DNO capacity check costs nothing and takes a phone call or a web form. Discovering a six-figure reinforcement bill after a board has approved a project costs a great deal more.

Our feasibility studies build the grid-connection question into the same early assessment as roof structure and shading. That way, the G99 answer arrives alongside everything else you need to make the call — not months into a project that has already spent money assuming the connection would be straightforward.