It is no secret that Andy Burnham has pronounced devolution as central to his strategy for the country. Burnham kickstarted his tenure as Prime Minister by reorganising Whitehall to support that devolution agenda. Local-growth policy and devolution strategy were transferred to No. 10 North, while regional mayors were brought into a revived National Economic Council.
Simultaneously, AI policy has moved closer to the centre. The new AI Taskforce sits in the Office for the Prime Minister and Cabinet, alongside No. 10 North. Its chair, Lord Vallance, reports directly to the Prime Minister.
AI Growth Zones therefore sit at the intersection of two governing ambitions – stronger central direction of AI policy and greater devolution of economic decision-making.
The Programme Burnham Inherited
AI Growth Zones were first announced in the AI Opportunities Action Plan in January 2025, followed by a delivery paper in November of the same year. Five locations have since been announced: Oxfordshire, the North East, North Wales, South Wales and Lanarkshire.
The programme’s immediate purpose is largely physical. AI Growth Zones foster and utilise faster grid connections, electricity-price support in eligible areas, and a more supportive planning regime. They also enable central coordination between the Government, developers, and infrastructure providers across the AI stack. Further to this, Starmer’s Government provided funding for AI adoption and skills, and in England, local authorities can also retain additional business-rate revenue for 25 years.
By utilising these advantages, the programme aims to accelerate the development of large-scale AI infrastructure and expand the UK’s domestic compute capacity. In doing so, the Government intends to attract private investment and support AI research and adoption. Alongside these national objectives, place-based provisions seek to ensure that host areas share in the benefits through employment, skills development, business adoption and additional local revenue.
But large-scale AI infrastructure developments are not guaranteed to deliver broader benefits to the places that host them. Data centres are highly automated facilities and their production is capital-intensive. Their permanent employment can be modest relative to the scale of investment, evidenced by a proposed £10bn campus in Northumberland, for example, which is expected to create 400 full-time on-site jobs.
This is not to deny the economic value of data centres or AI Growth Zones. Rather, it is to recognise that their contribution to local prosperity is not automatic and should be assessed as a distinct outcome.
Three Questions for Analysing Place-based Growth
Three questions help assess whether AI Growth Zones deliver local prosperity. First, do national objectives align with the capabilities of respective regions? Second, does infrastructure investment produce additional local benefits? Third, how is authority distributed across levels of government? These questions are not exhaustive, but instead are drawn from common themes found in analysis on place-based industrial policy.
Firstly, on strategic alignment, the primary objective of AI Growth Zones is to develop AI infrastructure. The aim of increasing domestic compute capacity by reducing barriers to infrastructure development is a leading priority within the Government’s delivery plan. Beyond this central purpose, the plan also expects zones to support skills, AI adoption, and wider regional growth. In turn, concerns around strategic alignment should scrutinise how the programmes primary objectives link with secondary objectives such as regional growth.
The OECD’s work on place-based industrial policy frames this in terms of alignment between national sectoral priorities and the distinct capabilities and development paths of individual regions. This is important, because the UK’s AI Growth Zones extend across several regions in England, Scotland and Wales. These differences are visible across the designated zones. Culham sits within Oxfordshire’s research-intensive economy, which has established strengths in AI, quantum computing, life sciences and clean energy. The North East zone, based at Blyth and Cobalt Park, draws on a different combination of assets, including renewable energy, regional universities and industries such as advanced manufacturing. Planning arrangements also differ as the Government’s delivery plan sets out specific reforms for English planning authorities, while committing to work with the devolved administrations on sites in Scotland and Wales. Whether zones deliver local benefits may therefore depend on how well national objectives match local capabilities.
The second question is whether infrastructure investment creates additional local benefits. Announced investment and activity is not in itself an identification of local benefits. The Government’s delivery plan identifies employment, skills, local investment and AI adoption as channels through which host areas may benefit. It provides up to an initial £5m per zone for locally designed adoption, research and start-up support. In England, local authorities will also retain 100 percent of additional business-rate revenue for 25 years. These measures establish potential routes through which benefits may reach host economies, however, they do not establish their eventual additionality or distribution. The Government’s Green Book defines additionality as “outcomes that take place as a result of the intervention”. The Magenta Book likewise recommends explaining how place-based interventions are expected to produce outcomes and examining “spillovers” beyond their formal boundaries.
The What Works Centre for Local Economic Growth found mixed evidence that “local residents benefit (in terms of employment, unemployment or poverty)” from enterprise zones. The review described the evidence on wages and incomes as mixed and reported “limited evidence of positive spillovers”. Hence, analysis of AI Growth Zones should therefore determine the effects attributable to the programme, while examining whether those effects accrue to local residents, firms, institutions and public authorities or spillover to surrounding areas.
Thirdly, the extent to which AI Growth Zones are “place-led” has important implications for the delivery of local benefits. Responsibility for local infrastructure decisions is shared across several levels of government. The OECD argues that coordination across these levels of government has real implications in placed-based policy and therefore the delivery of local benefits. In turn, Burnham’s devolution agenda may significantly shape the extent to which AI Growth Zones produce such benefits.
Delivery So Far
Applying the three questions to delivery so far produces a mixed initial assessment.
By January 2026, the Government had designated five AI Growth Zones and reported £28.2bn of prospective investment and more than 15,000 expected jobs. These figures demonstrate the scale of the investment pipeline assembled around the programme. However, they principally describe commitments and forecasts rather than operational capacity or realised economic effects.
Projects associated with several zones have moved beyond designation. In Lanarkshire, a £300m financing package is supporting the expansion of an existing data centre and construction of a second facility. South Wales has secured its first announced commercial capacity commitment. There is also evidence of local tailoring as the North East Mayor and the Government have agreed a regional skills target, alongside programmes for students, teachers and work placements.
Nonetheless, delivery remains uneven. An August 2026 analysis of planning data found patchy activity across the designated areas, while recognising that the programme remains at an early stage. Published employment figures also provide limited evidence about local distribution. When asked to divide the projected jobs between construction, permanent operation and indirect employment, the Government stated it did not make specific assumptions about the nature or geographical location of indirectly created jobs.
So far AI Growth Zones have produced designations, investment commitments, financing, enabling works and some locally designed programmes. However, there is not yet sufficient evidence to show how much of this activity would not otherwise have occurred or how widely its benefits are distributed.
Looking Forward
Burnham therefore inherits a programme that has begun to mobilise investment, but evidence of its local effects and governance remains limited. Responsibility for AI Growth Zones now falls within Kanishka Narayan’s joint ministerial portfolio, although the departmental home of the dedicated Growth Zone Delivery Unit has not yet been publicly confirmed.
Burnham’s devolution agenda makes the distribution of authority within the programme the more consequential question. The creation of No. 10 North is intended to connect regional empowerment with national growth policy. Applied to AI Growth Zones, evidence of that approach could include mayors and devolved governments influencing zone objectives.
The £300m financing package announced for Lanarkshire represents continued progress under Burnham. Its relevance to his devolution agenda, however, will depend on the influence Scottish and local institutions exercise over the project’s development and local-benefit arrangements. Looking ahead, the central question is not only how much investment AI Growth Zones mobilise, but whether Burnham’s devolution agenda changes who shapes that investment, and whether any such change affects the additional benefits realised and retained in host places.