Different numbers describe different things
AEP Ohio’s 13 February 2026 update reports 5,642 MW of contracts signed under its Data Center Tariff as of 12 February. It reports 17,861 MW of contracted projects in total, scheduled progressively through 2035. These are utility-reported contract figures at a historical cutoff. They are not a measurement of operational IT capacity in November, and should not be presented as one.
A contract still has dependencies
The utility’s tariff information describes a study and agreement process and notes that transmission-dependent in-service estimates are contingent. A signed agreement is stronger evidence of commitment than an unqualified announcement, but it does not remove construction, equipment, approval or commissioning dependencies. The precise contract and project programme matter more than the presence of one milestone in isolation.
Source: AEP Ohio — Data Center Tariff
Keep the measurement attached
A pipeline register should preserve whether a value is IT MW, facility demand, an electrical-service quantity or a phase total. MVA cannot be relabelled MW without the relevant assumptions. A long-term campus ambition should not be added to its first phase as if both were independent projects. Where those definitions are absent, the honest entry is “basis unresolved”.
Follow the chain without inventing a conversion rate
An announcement, controlled site, operative permission, supply agreement, construction milestone and operational claim can each have a different evidence date and source. We recommend recording each separately. This is an analytical framework, not an empirical model of how frequently projects succeed. There is no defensible automatic discount factor in this note that converts announced capacity into a deliverable total.
Monitor the change that changes the decision
A changed phase date, a withdrawn permission or a revised supply assumption can be more useful than another headline total. A project watch should explain which source changed, whether the claim is attributable and what follow-up is needed. An absence of new public information is not proof that a project has stopped; a repeated announcement is not necessarily a new project.
What a decision brief should say
Present the source’s claim, the analyst’s interpretation and the unresolved dependency as separate statements. State a cutoff and the coverage limits. Government siting research is useful context for understanding why workload and location requirements differ; it does not certify an individual project. This original commentary was checked on 13 November 2026. No forecast, investment recommendation or professionally reviewed capacity assessment is offered.
Source: Government Office for Science — Data centre siting in the UK
Original sources
Checked 13 November 2026. No source files, imagery or restricted plans are republished. These are summaries and original commentary; linked content retains its publisher’s terms.
- AEP Ohio — February 2026 load update to PUCO
13 February 2026; figures as at 12 February. Utility-reported 5,642 MW under the new tariff and 17,861 MW contracted in total, progressively through 2035; not operational IT capacity or independent verification.
- AEP Ohio — Data Center Tariff
Live page; reviewed 13 November 2026. The utility describes its load-study and agreement process, cost obligations and qualified energisation estimates. Applicable tariff and territory require exact case review.
- Government Office for Science — Data centre siting in the UK
12 November 2026. Siting research considers workload, latency, infrastructure and community issues; meeting notes are research, not site consent.