A megawatt is a unit of power, not a description of what a data centre can deliver. The same unit can label IT design capacity, incoming electricity supply or the rating of proposed generators. Those figures can all be relevant to one project without being quantities that belong in the same total.
The problem is not arithmetic. It is deciding what the arithmetic would mean. Before adding two numbers, establish that they describe comparable, non-overlapping capacity within the same time and delivery boundary.
Manor Farm illustrates different measurement boundaries and a total that already includes its phases. Wapseys Wood illustrates alternative supply routes and an interim stage. Pinewood Nurseries illustrates a valid equipment-rating sum that still does not establish IT capacity.
Manor Farm: one project, several legitimate numbers
The planning inspector’s report records that the appellant designed the Manor Farm data centre for approximately 72 MW of IT capacity. That is the stated IT design, not a measurement of operating load. [1]
Tritax’s January 2025 announcement describes potential delivery of up to 147 MW of power. It identifies an initial 107 MW phase and a possible further 40 MW phase, both then subject to planning. The 147 MW is the stated potential total, not a third allocation to add to the other two. [2]
The issuer’s H1 2026 results statement subsequently describes 107 MW to be provided in 2027 and a further 40 MW in 2029 through its power joint venture. Those remain attributed supply statements with different delivery dates. They do not establish that all the power was available when the report was published. [3]
There are two separate traps here. Adding 107, 40 and 147 would count the stated phases and their total together. Adding the approximately 72 MW IT design figure to the supply figures would mix a different measurement boundary into that already duplicated sum.
Nor should the difference between 72 and 107 be presented as a capacity cut, measured overhead or available headroom. Dividing one headline by the other does not establish an efficiency metric. That would require matched technical boundaries and appropriate operating or energy evidence, rather than figures drawn from different descriptions of the project.
The useful account preserves each figure with its meaning: an IT design, reported power phases, the issuer’s potential site total and the stated supply timetable. It does not force them into one apparently more definitive number.
Wapseys Wood: alternatives are not extra capacity
The SDC M40 Campus power strategy reviewed on 1 October describes an approximately 35 MW interim connection, followed by one of two main scenarios. One uses an approximately 200 MW onsite energy centre before eventual wider-grid supply. The other proposes a direct connection to existing 400 kV transmission lines, with a new substation, and would not require the energy centre. [4]
These are alternative designs, not two independently established supplies to combine. The 400 kV figure is a voltage, not a power capacity in MW. And the proposed 35 MW interim stage does not, on its own, establish an additional supply that would remain simultaneously available alongside the later route.
The developer describes the generating-centre scenario as supporting up to two of three data-centre buildings. Its expectation of wider-grid capacity by 2040 belongs to that scenario; it is not a date to attach automatically to the direct-transmission alternative. [4]
A capacity table that drops the words “interim”, “either” and “up to two buildings” would lose the project’s most important qualifications. The route, date and building boundary are part of the number’s meaning, not optional footnotes.
Pinewood: a correct sum can still answer the wrong question
The applicant’s Energy Centre Report for Pinewood Nurseries describes eleven engines of approximately 4.5 MW each, giving a proposed total generating capacity of 49.5 MW. Within that equipment description, the arithmetic is explicit and traceable. [5]
But 49.5 MW is not thereby a verified IT-load figure. The sum describes proposed generating ratings; it does not establish installed equipment, permitted output, the amount available to IT equipment or live customer workloads.
The source remains an applicant-consultant report even though it is available through a council planning register. An official document host does not turn the applicant’s design into independently verified operational capacity.
The lesson is not that generation ratings are unhelpful. They answer a different question. Preserve the 49.5 MW as the report’s proposed generating total rather than relabelling it as a 49.5 MW operating data centre.
What has to match before a total is useful?
- Measure and unit
- Is the figure IT capacity, incoming supply or generation? MW is power; kV is voltage. Shared terminology is not enough to establish a shared measurement boundary.
- Physical scope
- Does it cover one hall, one building, one phase or the whole site? Is a component already included in another total?
- Route and design
- Are the figures parts of the same selected scheme, alternatives, or successive versions of a proposal?
- Time
- Do they apply to the same period? Is an interim supply being counted alongside a later replacement without evidence that both remain available?
- Evidence state
- Are you summing designs, attributed agreements, installed ratings or verified operating capacity? The result must retain that distinction.
These questions test whether a total is interpretable. They do not assign a readiness score or certify capacity.
Even matching labels are only a starting point. Two “IT MW” figures could describe a building and the campus containing it, or an earlier proposal and its revision. Counting both would duplicate capacity despite the apparently consistent unit.
At portfolio level, the same discipline applies. A sum of explicitly identified proposed IT designs may be useful as a proposed-design subtotal. It should not be renamed “available capacity”, and it should not absorb supply or generation figures simply because the IT figure for another project is unknown.
Unknown capacity should remain unknown, not zero. If only part of a project list has comparable figures, report the scope and exclusions alongside any subtotal. A subtotal for the records with usable evidence is not a census of the entire market.
Keep the numbers; improve the labels
The solution is not to discard inconvenient figures or choose the largest number as the definitive size of a project. It is to retain the source, date, unit, capacity boundary, phase and evidence status for each claim.
A smaller, clearly defined subtotal is more useful than a larger total assembled from incompatible claims. Where the boundaries cannot be reconciled, explain the uncertainty rather than using arithmetic to conceal it.
Read What does “power secured” actually mean for a data centre? for the evidence behind supply claims, and A completed shell is not an operating data centre for the distinction between delivery milestones. The research guide introduces the wider series.
Sources and scope
This analysis uses evidence retained by 1 October 2026. No fresh external source review or project observation was performed for this article. No comparable cross-project capacity total or independently verified operating IT capacity is asserted.
- Manor Farm appeal decision and inspector’s report (PDF), published 10 June 2026, inspector’s report section 5.5: the appellant’s approximately 72 MW IT design.
- Tritax Big Box, Manor Farm development announcement, 21 January 2025: potential 147 MW power total and initial 107 MW / possible 40 MW phases. Historical issuer proposal, not present-day availability.
- Tritax Big Box, H1 2026 results statement (PDF), page 17: stated 107 MW in 2027 and further 40 MW in 2029. Cover/approval dated 5 August 2026; exact first-publication day not established. Underlying agreements not independently inspected.
- SDC M40 Campus, Our power strategy, opening, Scenarios 1 and 2 and Common elements, as reviewed on 1 October 2026. Developer proposals; exact publication/revision date not established. The 2040 expectation is specific to Scenario 1.
- Buckinghamshire planning register, PL/26/06414/OA, Documents: Energy Centre Report, v1.0, August 2026, ELD on behalf of ACRE, page 2. Council document-list date 10 August; exact authorship date not independently established. Applicant-consultant evidence, subject to reliance/reuse limitations.
Report files are not reproduced here. Council portals may require their ordinary access acknowledgement. The interpretation framework is not an engineering assessment, contractual opinion or investment recommendation.