A report from economic think tank Verdant has found that new jobs generated by AI data centres could be only a quarter of the scale the industry predicts. The research estimates that data centres across the UK will create only around 10,000 permanent positions by 2035, a figure that stands in stark contrast to the vision of high-paying, new career paths for young people that the technology sector has been promoting.

On 12 August 2026, construction workers were spotted building a new data centre in Stone Ridge, Virginia, USA. While the AI boom has led to widespread expectations of significant employment growth, the economic analysis suggests that the actual number of permanent roles these facilities bring will be far lower than advertised.

In a report published on Wednesday, Verdant stated that the UK government and industry leaders have approved a wave of new data centre projects based on widely cited figures from the technology industry association techUK. However, the think tank argues that this data exaggerates the number of permanent new jobs the projects will genuinely support. Notably, Verdant itself advocates for stricter controls on data centre development.

Where the estimates diverge

While techUK predicts that UK data centres will support 24,300 direct jobs, Verdant calculates that the existing stock of data centres actually supports only around 4,400 direct roles. For projects that are planned or currently under construction, Verdant forecasts that roughly 10,400 permanent operational positions will be created by 2035. This figure is approximately a quarter of techUK's prediction of 40,200 jobs. techUK bases its assessment on typical staffing and compute capacity, arriving at a figure of 8.5 jobs per megawatt. Verdant believes this ratio is too high, suggesting it is based on older, smaller facilities that require more maintenance personnel.

By examining planning approvals and public staffing data for 20 data centres proposed in the UK, Verdant calculated a capacity-weighted average of just 1.2 operational jobs per megawatt. The think tank notes that newer, larger data centres are far more automated and require fewer permanent on-site employees. Furthermore, techUK's methodology includes off-site roles such as contractors and those in the supply chain, which Verdant argues inflates the overall number.

The report also highlights that data centres generate far fewer jobs per unit of electricity consumed compared to other large facilities like steel plants, car factories, hospitals, and schools, which can support hundreds or even thousands of roles per megawatt of energy use.

In response to a request for comment, techUK defended its figures, stating that its estimations use public methodology, industry data, and standard economic impact assessment models. “The estimate was based on the best available industry data at the time, the process was transparent, and we are confident the number still reasonably reflects the industry's contribution to employment,” techUK said.

“Verdant’s employment estimate is based on a sample size that is too small to be reliable. Furthermore, this report only counts staff who work physically inside data centres, ignoring the vast number of jobs indirectly supported by the centres and the services they provide,” techUK added.

A UK government spokesperson rejected Verdant's conclusions, saying: “The 'jobs per megawatt' metric used in this report is highly misleading, comparing data centres to hospitals and schools is inappropriate, and it confuses installed power capacity with actual electricity consumption.”

“More importantly, the report ignores the broad economic value of the compute industry and the national security and digital sovereignty significance of building key digital infrastructure here at home,” the spokesperson added.

The context of the AI boom

The UK government has made expanding AI infrastructure a priority. It has labelled data centres as “vital to almost all economic activity and public services” and designated them as Critical National Infrastructure in 2024, underscoring their strategic importance and their vulnerability to cyber attacks. In June, the government released a paper suggesting that future data centres need to be more aesthetically pleasing and better integrated into their local communities.

Meanwhile, major technology companies are pouring tens of billions of dollars into building dedicated AI data centres. Google, Microsoft, Meta, and Amazon have all significantly increased their investment in this area this year.

The data centre construction boom is already changing the labour market. Sander van't Noordende, the CEO of human resources giant Randstad, said in March that the traditional path of “going to university and then directly into an office job” is becoming a thing of the past. He suggested that young people should consider skilled trades, which offer comparable career prospects.

An analysis of 50 million job postings by Randstad found that between 2022 and 2026, demand for advanced technical roles such as robotics technicians within data centres has risen by 107%. Demand for cooling and HVAC engineers has grown by 67%, and demand for industrial automation technicians has increased by 51%.