A megawatt figure can make a data-centre project sound easy to describe. It says little about whether the relevant planning permission has been issued, which phase has a power commitment, whether a customer has signed, or what the advertised completion date actually measures.
The difficulty is not simply finding more information. It is working out which statements can be compared, which have been superseded and which remain claims rather than independently verified outcomes. Four examples from our research show why a single project-status label can conceal more than it reveals.
Manor Farm: compare the same milestone
In January 2025, Tritax described a target of H2 2027 for Phase 1 practical completion and income at Manor Farm, conditional on planning and a pre-let. Its later half-year results statement gives a practical-completion forecast of Q2 2028. [1] [2]
That is a meaningful forecast revision because the comparison concerns the same broad milestone: practical completion within a powered-shell development model. It is not a comparison between grid energisation and a fully operational customer installation.
It would nevertheless be wrong to describe this as a measured delay against a binding construction contract. The earlier target spans half a year, and neither forecast establishes full IT commissioning. The results statement acknowledges planning-process delay, but that does not establish that planning alone explains the entire programme change.
The distinction matters commercially. Power availability, shell handover, occupier fit-out and the start of computing workloads are different events. A project can make progress on one while another remains conditional. Finding the revised forecast in an older document also does not mean the forecast changed on the day we read it.
Akzo/Wexham Road: an old status can outlive the decision
For the former AkzoNobel site on Wexham Road, an earlier council summary left application P/00072/152 under consideration. The later application record and the first page of its signed decision notice establish conditional outline permission on 30 July 2026. [3]
A directory built from the earlier summary could continue to describe a pending application after the decision had been made. Even the direct portal needs careful reading: at our review, its generic status label still said “Registered application”, alongside the specific approval decision and date.
The better conclusion is narrow but useful: this application received conditional outline permission. It is not evidence that every condition or reserved matter has been approved, that a power connection is secured, or that construction or operation has begun. Permission changes the planning position; it does not settle every other dimension of the project.
188–216 Bath Road: withdrawal is an application event
An earlier summary described a resolution to grant permission for P/20367/001, subject to a Section 106 agreement. The direct council register subsequently records “Withdrawn by Applicant”, dated 20 August 2026. [4]
That supersedes the old pending-agreement description for this application. It does not prove that the entire site has been abandoned, that no replacement proposal exists, or that a development could never proceed under another permission. Our reviewed evidence did not establish the reason for withdrawal.
There is a second consequence: proposed substations described in the withdrawn application should not be carried forward as if they were a current committed power route. But the withdrawal is not proof that no power arrangement exists elsewhere. The defensible position is to identify what has changed and leave the rest unresolved.
Langley: a named applicant is not a verified delivery outcome
A Ramboll scoping report dated 13 May 2026, available through application P/00437/096, identifies MSFT MCIO LTD as the applicant for a proposed Langley data-centre development. It describes an evolved scheme requiring new full applications for the campus and its substation, reports substantial site clearance and anticipates construction during 2027–2031. [5]
That is valuable identity, design and programme context. But it remains applicant-consultant evidence. Publication on a council portal does not convert every statement in the report into an authority-verified fact.
The construction window is a forecast, not an observed delivery record. The proposed sequencing of a dedicated substation does not establish an accepted grid connection or binding energisation date. The named applicant is not, by itself, evidence of an independently checked land title or a signed customer lease. Nor does a scoping submission establish that replacement full planning applications have been granted.
One project, several evidence states
A project can have a planning status, a power claim, a customer claim and a delivery claim based on different sources, different dates and different levels of verification. Compressing them into “planned”, “under construction” or “operational” can erase the qualifications that make the information useful.
- Planning
- Which application, decision and conditions apply to this phase?
- Power
- Is this proposed infrastructure, a reported agreement or independently established supply?
- Customer
- Is a tenant being discussed, a lease reported as signed, or the agreement itself verified?
- Delivery
- Does the date refer to energisation, shell completion, fit-out or actual operation?
The number itself needs a boundary too. IT load, incoming electrical supply and generation capacity are not interchangeable measures. Floorspace cannot be converted into delivered IT megawatts without assumptions that may not hold for the scheme.
Dates deserve the same discipline. The date of a decision, the publication date of a document and the day an analyst reads it are separate clocks. A recently accessed source may still describe an old position. A newly discovered historical event improves the research without being a new event in the market.
Better evidence is not the same as more certainty
The useful response is not to force every project into a confidence score. It is to make the supported statement precise, retain its attribution and show what remains unverified. A decision notice may settle a planning question while leaving the power and customer questions open.
Equally, “no change found” only describes the result of checking particular sources for particular information. It does not establish that a project is on track. If the checked sources could not reveal a lease falling through or a connection date moving, their silence cannot resolve those questions.
Headline capacity is a starting point. For anyone assessing future data-centre supply, the more useful question is: what, exactly, does the evidence establish about this phase, as of this date?
For the wider infrastructure context, read Can the UK compete for the next wave of AI infrastructure?
Sources and method
This analysis uses public documents reviewed by 1 October 2026. It is a selective explanation, not a live status service, a complete consent or contract audit, or a comprehensive account of any site. Conditional permissions, forecasts and applicant statements retain those qualifications. The findings are historical research discoveries, not newly occurring October project events.
- Tritax Big Box, January 2025 development announcement, 21 January 2025: conditional Phase 1 practical-completion and income target.
- Tritax Big Box, H1 2026 results statement (PDF), pages 6 and 17: revised completion forecast and powered-shell model. The document’s cover/approval date is 5 August; its exact first-publication day was not established. Also available through the issuer’s results library.
- Slough planning register, P/00072/152: decision dated 30 July 2026 and signed decision notice, first page. The full conditions and discharge record were not comprehensively audited.
- Slough planning register, P/20367/001: withdrawal entry dated 20 August 2026. The generic status label and specific decision were not treated as interchangeable.
- Slough planning register, P/00437/096, Documents: Ramboll, Former Langley Business Centre EIA Scoping Report, v1.0, dated 13 May 2026; introduction and sections 2.3–3.3. The register describes scoping while its application-type field says screening. The report was read as applicant-consultant evidence, not independent delivery verification.
Planning portals may require their ordinary access acknowledgement. Report files, detailed research records and internal monitoring data are not reproduced here.