The rapid expansion of artificial intelligence datacentres across the UK has become a political and economic flashpoint. The government has championed their development as vital for national competitiveness and job creation. But fresh analysis suggests the economic case for this buildout may be considerably weaker than ministers and the tech industry have claimed.
Environmental thinktank Verdant has completed a detailed study of planned UK datacentres, examining staffing projections from 20 projects with publicly available data. Their findings paint a starkly different picture from the employment numbers circulating in government and industry circles. Where TechUK, the sector's main lobby group, estimates datacentres will create over 40,000 permanent jobs, Verdant's research suggests the actual figure will be closer to 10,400. That's roughly a quarter of what's been promised.
The jobs mismatch: what the numbers reveal
The discrepancy isn't trivial. Verdant based its analysis on datacentres already operational in the UK, which have directly created fewer than 4,400 jobs to date. When you divide this by the energy these facilities consume, it works out to approximately 8.6 jobs per megawatt. That's dramatically lower than TechUK's claim of 43.7 jobs per megawatt.
More troubling for future development: the datacentres currently at planning stage tend to be significantly larger than those already built. This scale works against job creation. Verdant estimates these newer facilities will generate just 1.6 permanent positions per megawatt of capacity.
To illustrate why this matters, consider comparative figures. A car manufacturing plant typically creates around 400 jobs per megawatt of energy used. A hospital or major health facility generates more than 3,600. By this measure, datacentres are exceptionally capital-intensive but labour-light operations.
What this means for property and communities
For homeowners and communities near planned datacentre locations, the implications warrant careful consideration. While permanent job creation appears limited, these facilities do demand substantial resources. Water consumption is significant. Energy demand is enormous. Local grid capacity often requires upgrading. Yet if job benefits are overstated, the case for accepting these trade-offs becomes harder to justify.
Communities evaluating planning applications for new datacentres should scrutinise actual staffing commitments rather than accepting industry projections at face value. Verdant's co-director James Meadway makes the point sharply: "We're being rushed into supporting a frantic pace of datacentre buildout on very flimsy economic grounds."
The government has dismissed Verdant's methodology, arguing the thinktank uses a "deeply misleading jobs-per-megawatt metric" and conflates electricity capacity with consumption. That's a fair technical point. But it doesn't change the underlying reality: datacentres are simply not labour-intensive industries, regardless of how you measure it.
The construction jobs question
One distinction worth noting is that Verdant's analysis focuses on permanent operational positions. The building phase involves many more workers, though these are temporary roles. If you're living in an area about to host a major datacentre project, you'll likely see construction activity for 18-24 months, bringing jobs and traffic disruption alike. Once operational, however, the facility will require only a skeleton crew.
This temporal mismatch creates a particular challenge for local planning decisions. Short-term disruption is tangible and visible. Long-term economic benefits, it appears, will be thinner than advertised.
Where mortgage rates and housing fit in
For homebuyers and sellers in areas targeted for datacentre development, it's worth understanding the likely trajectory. Infrastructure investment can sometimes boost property values in surrounding regions, particularly if it triggers secondary economic activity. However, when those promised jobs don't materialise in expected numbers, that multiplier effect may not emerge.
With UK house prices currently averaging £272,188 and annual growth at just 2.0%, every bit of local economic strength matters. Don't assume that a major datacentre project necessarily strengthens your property's long-term prospects. Monitor planning documents carefully, and factor in both the construction phase and the post-completion reality when evaluating a property's location relative to these facilities.
What you should do
If you're considering buying or selling near a planned datacentre, request the detailed planning application and staffing schedules from the local authority. Compare the developer's job claims against actual data from similar facilities elsewhere. Ask whether your local council has modelled the energy and water demands realistically.
The technology itself isn't the issue. Artificial intelligence and cloud computing are genuine infrastructure needs. But the economic narrative around datacentres deserves scrutiny. Base promises about community benefit on transparent, auditable figures rather than industry optimism.