Research pillar
Grid and Infrastructure
Whether Britain's transmission and distribution networks can physically deliver the power that AI compute is being planned around.
What this pillar covers
We have not yet published original analysis under this pillar. Rather than fill the page with placeholders, here is exactly what it will cover — and the verified material already on the site that relates to it.
- Connection queues, connection reform and what the published timetables imply
- Network reinforcement programmes and who pays for them
- Regional constraint evidence and its confidence level
- Generation options put forward for AI load, including their delivery risk
From the UK AI Energy Index
Every indicator carries its source, the period it covers and a confidence level. Where the evidence is insufficient, the Index says so rather than estimating.
Data-centre evidence
The project register is compiled from planning records and company announcements. It is not a complete national census and does not claim to be.
Tools for this subject
All outputs are modelled estimates based on the inputs you provide, not measured values.
Recent news in this pillar
All news →
Nano Nuclear partners with Tillman Digital to deploy micro nuclear reactors at US AI data centers - DCD
datacenterdynamics.com
Data centres drive Ireland to reopen nuclear power debate
ft.com
Digital cities: the future of Britain’s data centres
thetimes.com
Two planned datacentres will have higher UK carbon emissions than ExxonMobil, analysis finds | Datacentres - UK | The Guardian
theguardian.com
Where are authorities restricting data centres amid AI boom? | Reuters
reuters.com
Proposed London datacentre ‘incompatible’ with net zero with 1m-tonne carbon footprint | Datacentres - UK | The Guardian
theguardian.com
How we research this
We work from primary UK sources — DESNZ, NESO, Ofgem, planning authorities and published company documents. Where a figure is modelled by us it is labelled an AI Energy Intelligence modelled estimate, with its assumptions shown. Where the evidence does not exist, we say so.