Data-centre grid capacity is becoming a deliverability test

Power availability is moving into the critical path

Large digital-infrastructure projects have historically been described through land, fibre, construction, customer demand and equipment. Those inputs still matter, but electricity connection evidence is increasingly a gating condition for whether a proposed site can actually be delivered on the claimed timetable. The International Energy Agency reports that data centres used about 415 TWh of electricity in 2024 and projects consumption at roughly 945 TWh by 2030. The agency also notes that local impacts are much more concentrated than the global share suggests [1].

The practical point is not that every data-centre proposal faces the same constraint. It is that a headline capacity figure, a queue position, a memorandum of understanding or an indicative utility conversation does not by itself demonstrate usable power at the required location and date. The IEA identifies long and complex connection queues, multi-year transmission build times and longer equipment lead times as potential sources of delay; it estimates that roughly one fifth of planned data-centre projects could be at risk of delay if grid risks are not addressed [1].

A queue is an allocation process, not an operating asset

The United Kingdom provides a current regulatory example of why the distinction matters. On 29 July 2026, Ofgem launched a consultation focused on speculative data-centre projects in Britain’s electricity connections queue. It reported demand connection applications increasing from 41 GW to 125 GW in under a year, with at least 80 GW attributed to data-centre projects. The proposed response includes a commitment fee and project-progress milestones, rather than treating every application as equally ready to connect [2].

Ofgem’s stated concern is that non-viable applications can occupy scarce capacity, delay projects that are ready to progress and distort network-investment signals. The consultation proposes evidence of financial capability, commercial maturity and procurement activity for projects seeking to retain their place [2]. That is a useful commercial distinction even outside Britain: a project may have a real application, but its connection position, milestones, costs, conditions and dependency on upstream works can still be uncertain.

Oakhampton inference: power diligence belongs before final delivery claims

Oakhampton’s inference is that grid capacity should be treated as a deliverability workstream alongside planning, construction and customer contracting. It is not a forecast of any particular project’s success and it does not replace jurisdiction-specific engineering, utility or legal advice. It is a sequencing point: a project team should establish what has been awarded, what remains conditional, and which party carries the cost and timing exposure before presenting a commercial operating date as firm.

This matters because grid pressure can change both the economics and the risk allocation. A later energisation date can defer revenue, require temporary generation or storage, alter equipment sequencing, affect customer commitments and bring additional network charges or security requirements. European Commission policy material frames faster renewable deployment and necessary grid infrastructure as part of lowering energy costs and strengthening resilience, reinforcing that grid build-out is a system-level issue rather than a site-only utility detail [3].

Evidence that makes a capacity claim more decision-useful

A diligence package does not need to predict the entire electricity market. It does need to reconcile the proposed load profile with the governing connection process and the documents that support it. The following items make the difference between a broad capacity narrative and a testable delivery case.

The evidence should be versioned, tied to the named project entity and reconciled to the site plan, customer-load assumptions and construction programme. Where a developer relies on future network reinforcement, the scope, owner, approvals, funding, procurement status and realistic contingency should be visible rather than embedded in a single target energisation date.

• Connection status: application, offer, acceptance, contracted capacity, expiry dates and unresolved conditions.
• Load definition: committed MW, ramp profile, redundancy design, import/export assumptions and any flexibility obligation.
• Queue position: governing rules, milestones, curtailment or reallocation risk and evidence required to retain priority.
• Network works: on-site and upstream scope, accountable delivery party, cost allocation, security and procurement status.
• Programme linkage: dates reconciled across utility correspondence, civil works, equipment lead times and customer commitments.
• Fallback case: temporary supply, phased energisation, lower initial load or revised commercial milestones, each clearly costed and conditional.

Uncertainty remains material and should stay visible

Public sources establish direction and examples, not a universal connection outcome. The IEA’s outlook includes significant uncertainty around AI adoption, efficiency and the speed of energy-sector build-out; its scenarios span a wide range of data-centre electricity demand by 2035 [1]. Ofgem’s July 2026 measures are consultation proposals, so their final form and effect cannot be assumed [2]. The European Commission’s action plan is also policy context rather than a site-specific connection right [3].

The bounded conclusion is therefore operational. For a data-centre or other large-load proposition, treat available grid capacity as evidenced only when the capacity, conditions, costs, works and timeline can be traced to current primary documents. Until then, label the capacity claim as conditional and keep its commercial consequences in the delivery case. That discipline does not eliminate grid risk, but it prevents an early-stage queue narrative from being mistaken for energised infrastructure.