The UK Government had promised that AI Growth Zones would help accelerate the development of the infrastructure needed to support its AI ambitions, but new planning data suggests that progress remains patchy.

According to a Freedom of Information (FOI) request analysed by property data firm Search Acumen, there appears to be limited planning activity across some of the areas designated as AI Growth Zones, despite promises of faster planning decisions and improved access to power.

The clearest example is Culham in Oxfordshire, which became the first AI Growth Zone in January 2025. According to the data, just one relevant planning permission was granted in South Oxfordshire during 2025, compared to five in 2024 and eight in 2023.

Of course, that doesn’t necessarily mean the policy itself has caused development to slow, especially given the relatively short period since Culham received its designation. However, it does highlight the gap that still exists between the Government’s ambitions for AI infrastructure and projects actually making their way through the planning system.

Culham struggles to build momentum

AI Growth Zones were first announced as part of the Government’s AI Opportunities Action Plan in January 2025, with designated areas promised additional support around planning and access to the electricity network.

Five zones have since been announced, covering Culham, Blyth and Cobalt Park in the North East, Anglesey and Trawsfynydd in North Wales, Bridgend in South Wales and Lanarkshire in Scotland.

The idea is relatively straightforward. By directing AI infrastructure towards areas with suitable land and access to large amounts of electricity, while simultaneously removing some of the barriers around planning, the Government hopes to get new data centre capacity built more quickly.

Of course, designation alone can’t simply make those barriers disappear.

Culham was considered particularly attractive because the closure of the Joint European Torus fusion experiment freed up access to significant electrical infrastructure. In fact, access to power is fundamental to the AI Growth Zone model, with designated sites expected to demonstrate that they can support at least 500 MW of AI infrastructure.

That means the relatively slow planning activity can’t simply be blamed on the same grid connection problems facing data centre developments elsewhere in the UK. The bigger question is whether the Government can turn the advantages promised by AI Growth Zone status – including access to power and additional planning support – into projects actually moving through the development process.

Planning therefore becomes particularly important. While the Government has repeatedly promised to make it easier to build data centres, developers continue to face local opposition and lengthy planning processes.

It’s not all doom and gloom, however. Search Acumen’s figures show 10 relevant permissions were granted in North Tyneside during 2025, up from four in 2024, although many of those projects were already moving through the system before Blyth and Cobalt Park received AI Growth Zone status in September.

There are also signs of more tangible progress elsewhere. Nebius has now agreed to deploy AI infrastructure at Vantage Data Centers’ Newport campus, representing the first announced commercial capacity commitment within the South Wales AI Growth Zone.

That makes it difficult to argue that the zones have produced no pipeline whatsoever, but there is still little evidence that designation itself has yet led to the dramatic acceleration promised by the Government.

Are AI Growth Zones actually working?

Part of the problem is that data centre development can’t be solved through planning reform alone – even though it’s a significant hurdle facing many developers across the UK. 

A developer may have access to suitable land and power, but there are still questions around local consent, environmental requirements, construction capacity and whether there is enough certainty to justify committing significant capital to a project.

That is particularly important when assessing the relatively thin pipeline revealed by Search Acumen. AI Growth Zones are still at an early stage, and planning data alone can’t tell us whether developers are actively preparing schemes that have simply not yet entered the formal system.

Andrew Lloyd, Managing Director of Search Acumen, explained, “For occupiers or developers banking on AI Growth Zone status to de-risk a site or speed up a consent, this data offers no reassurance.

“That’s a significant signal for anyone pricing risk on an AI-adjacent development. There are brownfield opportunities, but these remain untested.

“The simple fact that there are so few of these AI-related initiatives coming forward within designated zones, with many classified as minor, and often still undetermined in planning status, tells you just how early-stage the market actually is versus the political narrative around it.”

There is another explanation for the relatively thin pipeline, however. Developers may simply be waiting for greater certainty around policy, occupier demand and the practical delivery of the zones before spending money progressing applications.

Andrew added, “Developers may not commit to planning fees until grid connections are confirmed, policy is clearer, and occupier demand is more certain. For commercial real estate professionals, the question is whether this is a temporary lag or a structural signal that the zones lack the delivery infrastructure to attract capital.

“With almost nothing coming through the system, there’s no evidence yet of the capacity or policy framework to handle major AI infrastructure applications when they do arrive. That’s a risk for any developer who moves early, as they may be finding their way in a policy vacuum.”

That is ultimately the bigger test facing AI Growth Zones. Announcing locations and promising streamlined planning is relatively easy; coordinating planning authorities, developers, investors, infrastructure providers and Government departments is significantly harder. There is also a wider question over how these zones fit into the UK’s existing data centre market, with the industry warning that AI Growth Zones need to strengthen rather than fragment the country’s wider digital infrastructure.

Search Acumen wants more transparency around that process, including standardised digital datasets that would make it easier to track the number and status of projects within each zone.

Andrew said, “If you have a situation where economic headwinds spook markets, delays in progress feel inevitable. But bottlenecks across planning and related property and legal workflows remain in our control and should be eased as a priority. This starts with real-time pipeline visibility: FOIs are a workaround for fragmented, inconsistently structured planning data. Standardised, digital-first datasets and dashboards would help track whether zones are evidentially accelerating delivery.

“Ultimately, practical realities of power, planning, cooling, construction capacity, and government coordination are without doubt slowing delivery. Culham has effectively become a test case for this, telling us whether the UK can build AI infrastructure at hyperscale fast enough to compete internationally. So far, the answer is undecided.”

For now, it is probably too early to declare AI Growth Zones either a success or failure. However, if the Government wants them to underpin its wider AI ambitions, the number of projects progressing from announcement to planning and ultimately construction will need to begin growing significantly faster.