The connection problem
Why the grid can't carry your electric fleet on its own, who could fix it, and the timetable for reform.
The grid is the reason the financier deal exists. This page sources the scale of the problem, how the connection system works, and why the current reform programme doesn't yet reach the depot infrastructure this report is about.
Six months. 460% growth.
The scale of the problem
A single HGV charging at 350 kW draws what 350 average homes draw simultaneously. A depot of ten trucks charging at peak is 4 to 10 MW. That requires a dedicated HV connection the local DNO must design, install, and energise — often with upstream reinforcement.
Three things broke the system in quick succession. Electricity demand surged as data centres, EV charging, heat pumps and industrial electrification all hit the queue at once. Application costs are low relative to project scale — cheap enough to lodge speculatively — which flooded the queue with projects that had no intention of building. And the queue collapsed under its own weight.
The transmission demand queue grew 460% in six months to June 2025, from 17 GW to 96 GW. Total demand queue across transmission and distribution reached 125 GW against a GB peak of roughly 45 GW. Most of that queue is data centres — but the queue doesn't separate. A logistics operator wanting an 11 kV upgrade joins the same line. Typical wait for a new HV connection: 18 to 36 months in a functioning market. SMMT documented operator waits of up to 15 years.
How a connection works
Most fleet depots connect at 11 kV or 33 kV — the distribution network. Data centres connect at transmission. This distinction matters because Phase 1 of the current reform programme covers transmission only.
A connection splits into two types of work. Non-contestable works touch the live network; only the DNO can do them, at regulated prices. Contestable works — cables, switchgear, transformers within the customer's site — can be tendered to an Independent Connection Provider. Operator frustration concentrates on the non-contestable works. The DNO is a monopoly there and timelines are entirely within their control.
Where the substation lacks spare capacity, the DNO must reinforce upstream before connecting. That's expensive, slow, and its costs are shared between the triggering customer and future users. Disputes over cost allocation are common. The IDNO model (connection wrap, at /report/deals) exists in significant part because the regulated DNO process cannot deliver reinforcement on operator-relevant timelines.
Four tensions shaping the depot decision
Years to wait vs months to deliver
What the grid asks vs what the balance sheet allows
Who owns the wire, who runs the depot
125 GW queue vs 45 GW peak
The institutions
Six bodies do most of the work. The six DNOs (NGED, Northern Powergrid, SP Energy Networks, UK Power Networks, SSEN Distribution, Electricity North West) hold the monopoly on non-contestable distribution works. ICPs compete on contestable works; IDNOs within that market own the connection asset long-term. NESO manages the transmission queue. Ofgem leads Curate and Connect. DESNZ leads Plan, using Section 18 of the Planning and Infrastructure Act 2025 to designate strategically important projects.
The reform programme
The joint Ofgem-DESNZ-NESO reform programme has three pillars. Curate cleans up the queue — higher financial requirements, readiness milestones, removal of speculative applications. Plan designates strategically important projects for priority access. Connect speeds up physical delivery through independent contractors, phased connections, and performance regimes.
Phase 1 launched in 2026. It applies to transmission-connected projects only. Phase 2 would extend the tools to distribution, but has no published timetable.
On 11 March 2026, the government named EV charging hubs alongside data centres as strategically important. That recognition hasn't been operationalised. Distribution-connected depot infrastructure is still waiting for Phase 2.
The four live tensions
The reform is real and delivering at transmission scale. Four structural problems remain unresolved for operators.
Distribution versus transmission. Phase 1 reaches data centres because they connect at transmission. EV charging hubs connect at distribution. The government named both as strategic. The system still treats them differently.
Designation versus market. The Plan mechanism requires a published government strategy document to trigger. That works for AI Growth Zones. It doesn't work for fleet electrification, which is delivered by hundreds of separate commercial decisions. There is no list to designate.
Small versus large. The financial mechanisms designed to deter speculative data centre applications are calibrated to 200 MW projects. A deposit that's a rounding error to a hyperscaler can be existential to a 5 to 10 MVA depot project. The reform hasn't shown how it distinguishes these cases.
Speed versus quality. The reform timetable is being driven by AI commitments. Mechanisms designed for data centre deployment may get applied to fleet electrification without adequate design scrutiny. The two demand types are structurally different in load profile, capital structure, and locational pattern.
The constraint payment context
NESO's Annual Balancing Costs Report 2025 puts constraint payments at £1.7 billion for 2024/25. The projected peak for 2030 is £7.2 billion. These are payments to generators not to generate, because the network cannot move the energy to where it is needed.
The political case for grid reform doesn't depend only on transport decarbonisation. Not reforming the grid is itself expensive, at a scale that dwarfs the cost of the reform.
Structural framing in the four live tensions section adapts the analysis in Andy Harper's "Grid Connections: State of Play" primer prepared for 12 Pillars Workshop 2, May 2026. Credit line to be confirmed with Andy Harper before publication.