Court ruling tests Britain’s £1bn data centre power plans – and energy policy
A High Court challenge to a £1bn data centre beside the M25 has exposed a glaring issue for Britain’s ambitious AI build out. Power hungry projects are effectively being approved before anyone is sure where the power will come from.
The government has accepted that planning permission for a 90MW hyperscale facility at Iver in Buckinghamshire should be quashed, after claiming it was granted on the assumption that low-carbon energy solutions could later be secured.
This assumption, officials have now said, was flawed.
This admission clears the way for the first full legal challenge to a hyperscale data centre in the UK, whilst raising uncomfortable questions around the UK’s AI infrastructure boom.
Power assumptions under scrutiny
The scheme, proposed by Greystoke Land and Altrad UK on a former landfill site, was initially rejected by Buckinghamshire Council over green belt concerns.
It was then waved through on appeal by central government.
Inspectors accepted that the project would have huge landscape and environmental impacts, but concluded these were outweighed by national data centre demand.
They also decided a full environmental impact assessment was not required.
That judgement is now under pressure, with campaigners arguing that ministers failed to properly assess the electricity and cooling demands needed to run modern AI workloads.
That approval, they have claimed, relied on mitigation measures that were never locked in.
In a letter to the court, government lawyers accepted that relying on the future sourcing of low-carbon power without guarantees amounted to a “serious logical error”, exposing some of the tension at the heart of the UK’s energy policy.
And for the tech sector, the concerns lie less in local opposition and more about whether the planning system is keeping pace with the real bottlenecks that AI infrastructure faces.
A wider risk for tech investment
The challenge comes as data centre development accelerates across the UK.
Planning applications hit a record high last year as investors rushed to secure land for compute, particularly in London and the South East.
But power availability is increasingly the binding constraint.
Large, AI-focused data centres require tens of megawatts of continuous electricity, alongside significant cooling capacity.
Grid upgrades can take years, and developers are often forced to make assumptions about future energy solutions to keep projects moving.
This case suggests those assumptions may now face even tougher scrutiny. Developers and investors fear the ruling could aggravate legal challenges, slowing projects and adding uncertainty just as Britain pitches itself as an AI hub.
Meanwhile, it also exposes a tension at the heart of government policy, where ministers want faster approvals for AI infrastructure, but courts know that speed cannot come at the expense of credible energy planning.