Summary
- Verdant estimates around 10,400 direct jobs across the current planned UK data centre pipeline using public information from individual developments.
- TechUK separately forecast 40,200 additional operational jobs by 2035 under a higher-growth data centre capacity scenario.
- The competing methodologies increase pressure on developers and policymakers to separate permanent employment, construction work, and wider economic benefits.
The employment case used to support UK data centre development is facing renewed scrutiny after environmental think tank Verdant estimated that the current pipeline of planned facilities would directly employ around 10,400 people.
The analysis challenges assumptions that permanent operational employment will increase broadly in line with the amount of new data centre capacity being built.
Verdant compiled public information from planned developments, including planning applications, government statements, and company announcements, and used that material to estimate employment associated with individual projects.
Its briefing argues that newer and larger data centres require fewer operational workers per megawatt than implied by industry-wide employment assumptions derived from the existing sector.
Verdant estimates that existing UK data centres directly support roughly 4,400 jobs and that the current planned pipeline would support around 10,400 direct roles nationally.
The figure has been compared with research commissioned by techUK, but the two estimates are not measures of exactly the same thing.
TechUK’s 2024 report forecast that faster data centre capacity growth between 2025 and 2035 could create 40,200 additional direct operational jobs by 2035. The report assumed capacity growth accelerated from around 10% annually to 15%.
It also projected 18,200 additional direct construction jobs over the period, £44 billion of additional gross value added, and £9.7 billion of additional tax revenue.
Verdant is instead testing employment intensity against the current project pipeline and argues that the jobs-per-megawatt relationship weakens significantly at larger facilities.
Megawatts and jobs do not scale together
The central issue is whether electrical capacity is a reliable basis for forecasting operational employment.
A 500MW campus does not necessarily need five times as many employees as a 100MW site. Large operators can share security, facilities management, network operations, monitoring, and administrative functions across multiple buildings.
Automation and remote operations can reduce the number of staff required on site as capacity increases.
That creates economies of scale in labour even while electrical demand, cooling plant, backup power, and network infrastructure expand substantially.
Verdant argues that this effect makes historical jobs-per-megawatt ratios increasingly unreliable for the large AI-oriented campuses now entering the UK planning system.
The government and techUK have disputed the think tank’s conclusions.
TechUK maintains that its figures are based on a transparent economic model and a scenario in which the industry grows faster than its recent historical trend. The government has also argued that employment per megawatt is too narrow a measure of data centre economic value.
That criticism has merit when data centres are viewed as infrastructure. Their economic purpose is not principally to employ large numbers of people inside the buildings. They provide computing, storage, connectivity, and digital resilience used by businesses and public services elsewhere.
It does not follow that permanent employment claims should be ignored.
Jobs are frequently used in planning applications and political statements to demonstrate the local benefit of facilities that consume significant electricity and occupy substantial industrial sites.
When those figures form part of the development case, authorities need to distinguish between temporary construction work, permanent on-site employment, indirect supply-chain roles, tax receipts, and broader economic activity enabled by the computing infrastructure.
The distinction is becoming more important as project sizes increase.
The UK Compute Roadmap forecast a need for at least 6GW of AI-capable data centre capacity by 2030, while the AI Growth Zone programme was designed around sites capable of supporting at least 500MW of demand.
Electricity at that scale carries an opportunity cost. Connections, transmission reinforcement, land, water infrastructure, and construction capacity can also be used by manufacturing, housing, energy storage, transport electrification, and other industrial projects.
Verdant uses its findings to support a wider call for a pause in data centre construction until the UK has a national framework governing need and resource allocation. That recommendation reflects the organisation’s broader environmental position and goes beyond the employment calculation itself.
The narrower challenge remains useful regardless of whether a moratorium is adopted.
The case for data centres may ultimately be stronger when made around digital resilience, sovereign compute, tax revenue, investment, and the economic activity hosted by the infrastructure rather than around large permanent workforces.
If that is the argument, planning and policy documents need to distinguish those benefits clearly instead of allowing megawatts and jobs to be treated as though they expand at the same rate.
