The problem
Britain reordered its grid connection queue. The commonly quoted comparison — 700 GW down to 381.5 GW — is not like for like, and the error runs through most coverage. The pre-reform figure of more than 700 GW covers generation and storage only. The 381.5 GW reformed pipeline is 283 GW of generation and storage plus 98.5 GW of transmission-connected demand. On a like-for-like basis the reduction is 700+ GW to 283 GW, with over 300 GW not proceeding.
Offers are running late, and the regulator has intervened twice in opposite directions — a distinction routinely conflated. In February 2026 Ofgem refused relief on protected projects’ connection dates, disclosing that 210 of 340 protected projects were affected and stating it was "frustrated and disappointed". In June 2026 it granted an alternative timetable, setting hard deadlines of 30 September 2026 for Phase 1 offers and 31 January 2027 for Phase 2. Progress against the first of those looks strained: the industry dashboard showed 94 of 633 Phase 1 offers issued in mid-August 2026.
The transmission side of this is a genuine open data success. The Transmission Entry Capacity register is published twice weekly as CSV with an unauthenticated API — 2,205 project rows, retrievable in a single call, no registration.
Distribution improved sharply in 2026, and the standard story about it is now out of date. Ofgem digitalised the Long-Term Development Statement by direction in April 2024, mandating a common information model built on the international grid exchange standard. In the May 2026 cycle, five of six distribution operator groups published standardised capacity heatmaps openly against that model. One group published nothing publicly discoverable.
But the standardisation is thinner than it looks, and the reason is in the specification itself. Ofgem defines the fields and then explicitly delegates the calculation to each operator: available capacity is computed "based on more detailed analysis and internal policies", and a utility "should provide details on how this RAG status is calculated as part of the accompanying documentation." GB has a common schema for grid capacity. It has no common methodology, and the specification does not require one. Two operators can be fully compliant and return numbers that cannot be compared.
The serialisation diverges too. Only two of five emit the specified nested structure; the others flatten it or return a bare array. One operator’s master resource identifiers are sequential integers rather than the globally unique identifiers the standard exists to provide, defeating the entire point of the field. Another uses a transposed field name. Constraint status comes back as "Red", "GREEN" and "green" from three different operators, against a specification that types it as an enumeration. One operator is inconsistent between its own two licence areas.
The joins are broken too. The transmission register carries no geocoding of any kind — the only location field is a free-text substation name. There is no shared identifier between transmission and distribution registers, so a project appearing in both cannot be reconciled except by fuzzy matching. And the reform status column is empty for 63% of transmission projects, because it only populates after agreements are countersigned.
The system
Junction builds the national connection picture that neither government nor the market currently provides. It normalises all six distribution registers plus the transmission register into one schema, resolves entities across them, and maintains the join that does not exist.
It geocodes the transmission queue by building and maintaining the substation gazetteer that the register lacks, turning free-text substation names into mappable locations — the single missing ingredient that prevents anyone from seeing where queue pressure actually sits.
It normalises headroom to a single stated definition. The real barrier between operators is not file format but definitional inconsistency: each defines available capacity differently and publishes it as a picture rather than as data. Junction states its assumptions and applies them uniformly.
And it tracks queue integrity through the reform. With Gate 2 having reordered everything and offers running months late, monitoring which projects hold capacity against which are actually progressing is both newly possible and politically salient — including the protected projects where the regulator has already refused the system operator relief.
Worked examples
Two situations this system answers
A developer asks where they can connect 5MW within two years. Two network operators both report "available capacity" — computed on incompatible bases, because the specification delegates the calculation to each of them.
Junction converts everyone’s figures to one clear definition, and publishes each operator’s own method next to it — so the numbers can finally be compared.
The same project appears in both the transmission and distribution registers under different identifiers, with no shared key. Counting either register alone is wrong; counting both double-counts.
Junction matches the two records to the same project and says how confident it is, rather than quietly guessing.
Data foundation
Every dataset below is open, or its access constraint is stated
| Dataset | Publisher | What it provides |
|---|---|---|
| Transmission Entry Capacity register | NESO | CSV plus unauthenticated CKAN API, twice weekly. 2,205 rows. No geocoding, reform status 63% empty. |
| Embedded Capacity Registers | Six DNO groups | A common 57-column schema whose columns 11 and 12 carry eastings and northings — so distribution needs no gazetteer at all. Three of six serve anonymous users; the other three require free registration. Column names diverge wildly but the order is stable, so parse positionally. |
| Interconnector register | NESO | Open CSV, 33 rows, same schema family as transmission. |
| Connections reform results | NESO | Published as attachments on a web page, not as datasets. The files behind the links are silently replaced with no version history or changelog. The register of existing agreements is an opt-in applicant directory, not a list of reform outcomes — its own header says so. |
| Grid Supply Point boundaries and lookup | NESO | Polygons current to February 2026, plus an under-advertised lookup carrying actual point coordinates for 376 of 380 nodes — though that file has not been refreshed since 2018. |
| Clean Power 2030 Action Plan | DESNZ | Zonal capacity ranges by technology, the basis for strategic alignment decisions. |
| Local Area Energy Plans | Combined authorities | The demand-side counterpart that currently cannot obtain consistent grid data. |
Capabilities
Benefits
For government
- Gives the system operator and the regulator a single view of a queue currently visible only in fragments across seven publishers.
- Supports the digitalisation and data best practice regime that already obliges network operators to publish this data usefully.
- Serves combined authorities producing local area energy plans, who currently cannot obtain consistent grid data at all.
- Because zonal pricing was rejected, locational efficiency must now be delivered informationally rather than through price — which makes transparent capacity data more valuable, not less.
For the public
- Faster connection of clean generation means lower constraint costs, which reach households through bills.
- Communities can see what is queued to connect near them, currently almost impossible below transmission level.
- Smaller developers gain the visibility that only well-resourced firms can currently buy through consultancy.
Delivery
Phasing
Risks & mitigations
The most serious risk, and the one most likely to be missed. Ofgem consulted through mid-2026 on tightening open data triage on security grounds, stating that the changes are "likely to increase the proportion of data being triaged as shared rather than open". An argument resting on openness obligations is running into a headwind. The counter is in Ofgem’s own text: it names impediment to "connection reform, connection planning, and local energy planning" as a consequence of withdrawal, and invites concrete use cases with quantified impacts. Responding to that process with measured evidence is cheap and high-leverage.
Ofgem could tighten the specification to require a common methodology, removing the need. That would be good for the country and fatal for the product — so the work should inform that tightening rather than depend on its absence.
Without a shared key, matching is probabilistic. Confidence is published per match and low-confidence links are never silently merged.
The reform is mid-flight with offers issuing into 2027. Junction is designed to track a changing queue rather than describe a static one.
Sources
All sources checked in August 2026. Figures carry the reference period of their source, which may differ from publication date. Where a figure could not be verified against a primary source it is not used.