TL;DR

Research by the environmental thinktank Verdant puts direct permanent employment across every UK datacentre now planned at 10,400 people, against the 40,000-plus that trade body TechUK has projected. Ministers have not challenged the underlying headcount; they have challenged the yardstick. That argument is arriving exactly as the committed build queue passes 14.6GW, so the volume of capacity being consented is climbing while the evidence for its employment case is still unsettled.

What Verdant counted

The thinktank worked from 20 schemes where staffing numbers are already in the public domain — planning submissions, filed accounts and company announcements — and checked them against comparable sites abroad. On that basis, facilities already operating in Britain sustain fewer than 4,400 permanent roles between them, which works out at 8.6 people for every megawatt drawn. TechUK’s projections imply roughly five times that density, at 43.7.

Verdant’s more awkward finding concerns what comes next. Because the consented pipeline skews towards very large single-site campuses, and those are less labour-intensive per unit of power than the smaller facilities built to date, the figure could fall as low as 1.6 permanent posts per megawatt. The count deliberately excludes construction work, which is numerous but temporary.

Co-director James Meadway argued the sector is being expanded “on very flimsy economic grounds”, and that the case for each site ought to be established before it is built rather than afterwards.

The government contests the ruler, not the reading

Whitehall’s response went after methodology. A spokesperson called the jobs-per-megawatt measure “deeply misleading”, objected to benchmarking datacentres against hospitals and schools, and said the analysis “appears to confuse electricity capacity with actual consumption” — while adding that it ignores the sovereignty case for hosting sensitive workloads domestically. Verdant’s answer is that it used the capacity operators themselves say they require, since that is what gets reserved on the grid whether or not it is drawn.

TechUK defended its figures as resting on a published methodology and standard economic impact techniques.

Looking forward

Notice what is not in dispute: nobody has produced a competing headcount. The exchange is about whether employment per megawatt is a fair test at all, which leaves councils weighing applications without an agreed measure of the benefit. That vacuum is now doing real work in Scottish politics, where protesters massed at Holyrood and 24 hyperscale proposals sit in the planning system. Expect jobs-per-megawatt to surface in objections long before the statisticians settle it.