We hire out powered access to the contractors who build data centres. So we counted how many of them there are, and where the work is going.
The two don’t line up.
Why a Powered Access Company Counted Britain’s Data Centres
We hire out scissor lifts and boom lifts. Our customers are the electrical contractors, the mechanical and cooling engineers, and the fit-out crews. They’re the people who turn a big empty shed into a working data centre, and they’re the ones hiring the machines that get them working safely at height.
Britain is in the middle of a data centre boom. Almost everything written about it is about planning, energy or water. All of that matters. But nobody had counted the people who physically build them.
So we did.
We went back through the PlanIt register to find every data centre planning application lodged with a UK council since February 2000. That’s 678 schemes across 210 councils, from a single unit conversion in Aberdeen to a 540,000 sq m campus in Northumberland that’s still waiting for a decision.
Then we worked out which construction trades do the work on site, by industry code, and counted how many of them exist in each region.
Two things came out of it.
Planning is not the blocker people assume. Councils approve 97% of these schemes. What they don’t do is decide quickly.
We looked into the latest ONS BRES data to find the figures on what trades are involved in the construction of data centres, how many there are and where they’re based.
It’s important to remember here: there are no specific data centre tradespeople. A data centre is a shed with a very large mechanical and electrical fit-out inside it. The work lands on the same electricians and cooling engineers who build everything else, just far more of them per square metre than any warehouse would need, and every hospital, school and office fit-out in the country is competing for the same people.
The pipeline isn’t asking for a workforce that doesn’t exist. It’s asking for a bigger share of the one everyone already uses.
One thing we should say plainly. We hire out access equipment to the trades in this report, so we have an interest in people taking their availability seriously. That’s why the whole dataset and the method are on this page. Check every number, and disagree with us if you want to.
The Headline Numbers
- 678 applications lodged with UK councils since 2000, across 210 authorities.
- 44 a year is the rate now. It was 14 a year in the 2000s and 25 in the 2010s.
- 97% approved. Of the 580 that reached a decision, 563 were approved and 17 refused. The rest were withdrawn, referred elsewhere or have no recorded outcome.
- 63 are waiting with no decision at all, at a median of 315 days.
- 8,760 construction workers per scheme in the current pipeline, on Great Britain figures. In Wales it’s 2,409.
The Planning System Says Yes. It Just Takes its Time.
Of the 678 schemes lodged since 2000, 580 have reached a decision: 563 approved and 17 refused. That’s 97% approved and under 3% refused. Another 29 never got a council decision (26 withdrawn, two referred to other bodies, one unresolved), and six have no recorded outcome on the register. The remaining 63 are still waiting.
In short, councils are not actively standing in the way of data centres.
What has changed is the wait. Of the applications lodged in 2019, 92% had an answer within twelve months. In 2021 it was 90%. Of those lodged in 2025, 54%.
Sixty-three schemes are currently sitting with no decision. Twenty-eight have been waiting over a year, thirteen over two. Among them is a proposed 300MW AI data centre in Falkirk, one of the largest schemes anywhere in the dataset.
The queue isn’t evenly spread. The South East is holding 18 of the 63. The East of England has 15, waiting a median of 469 days. Thirteen sit inside a single corridor west of London, across Chiltern South Bucks, Slough, Buckinghamshire and Wycombe, and they’ll want the same contractors in the same months when they come through.
Outline applications, the early-stage ones where the details come later, are 21% of all applications and 46% of the queue.
The pipeline hasn’t stopped growing while this happens. In the first five months of 2026, 30 applications were lodged. At the same point in 2025, it was 22.
Where the Work Has Gone
We compared each region’s share of the 119 schemes lodged since the start of 2024 with its share of the trades that build them, then turned the difference into people.
The implied gap is what a region would need in a trade to match its share of the pipeline. It’s a proportional comparison, not a hiring forecast.
Wales is the sharpest case in Britain. It has lodged 24 data centre applications in its history, and 16 of them have come since the start of 2024. Two-thirds of everything Wales has ever proposed has arrived in the last two and a half years, and its implied gap across the sixteen trades is just over 100,000 people, including 22,000 more electrical installers than the 10,000 it has.
The West Midlands is the other end of it. Three schemes, 77,200 construction workers, and 50,900 more of them than its share of the pipeline needs. If there’s capacity anywhere, it’s there.
The Trades That Do The Work
Sixteen construction trades build a data centre on site. These are the industry codes the ONS counts them under, smallest first.
Just over a million people. That sounds like plenty until you look at the top of the list.
Two trades account for 41% of it on their own. Electrical installation employs 238,000 people and plumbing, heat and air conditioning another 187,000. Those are large, mobile trades that work across multiple sites such as offices, retail, hospitals, schools and warehouses.
Those two trades are the data centre workforce. What separates a data centre from the warehouse next door isn’t a special trade, it’s intensity: far more electrical and cooling work per square metre, to tighter tolerances, than any shed normally needs.
Every scheme in the pipeline is asking for a bigger slice of the same 425,000 people that every hospital, school and office fit-out is already drawing on.
The most specialised job on site is installing the plant itself. 22,550 people across Great Britain install industrial machinery and equipment, and on a data centre that means getting the generators, the uninterruptible power supplies and the chillers into position and working. It’s the smallest specialist pool a data centre calls on, and it’s not a data centre trade either. The same firms are fitting out factories and process plants.
We’re not saying those firms are all fully booked. We haven’t measured that, and frankly, we couldn’t if we wanted to. The point is how small the pool is relative to the number of projects now asking it for capacity.
What the pipeline does is stack that demand up in the same places at the same time. Thirteen schemes are waiting in a single corridor west of London, and when they come through they’ll want the same M&E crews, in the same months, as everything else being built around them.
Where the Biggest Gaps Are
The same comparison, broken down by trade. These are the ten biggest gaps in Britain, and they’re nearly all M&E.
Every Region, Every Trade
What This Means
Three things come out of the numbers.
- Planning isn’t the constraint. 97% of these schemes get approved. If you’re waiting on one, it will almost certainly come through. The question is when, not whether.
- The queue is going to release in a lump. Sixty-three schemes are sitting in it now, twenty-eight of them for more than a year. They won’t come out evenly. Thirteen are in a single corridor west of London and will want the same contractors, the same plant and the same access equipment in the same months.
- There is no reserve workforce for this. The work lands on the same M&E trades as everything else being built in Britain, and the pipeline isn’t landing where those trades are. Wales has 13% of the national pipeline and 4% of its electrical installers, a gap of 22,000 people in one trade alone. London has plenty of firms that put up buildings and very little of everything else. And the most specialised pool on site, the 22,550 who install the plant, is shared with every factory and process plant in the country.
If you’re planning capacity for next year, pricing work, or deciding where to put a team, the shape of it is worth knowing. This work is real, most of it is either approved or will be, and it is not spread evenly across the country or across the trades.
The buildings will get their planning permission. The harder question is who’s free to build them.
How We Did This
Applications come from PlanIt, which aggregates council planning registers and covers around 250 of the UK’s roughly 400 planning authorities, so our counts are on the low side. A small number of schemes (19 of the 678 total since the year 2000, 1 scheme in the decade of 2020) are from Northern Ireland and the Crown Dependencies, whose registers PlanIt also covers, are included in the totals.
We matched on “data centre”, “datacentre” and “data center”. Full and outline applications only, with condition discharges, reserved matters, amendments and neighbouring-authority consultations stripped out so one scheme isn’t counted twice. Applications that only mention a data centre in passing are excluded.
Approval rate is approvals as a share of decisions, meaning approvals plus refusals, which is the same convention government planning statistics use. Withdrawn, referred and unresolved applications aren’t council decisions, so they don’t count towards it, and six applications with no recorded outcome on the register are left out of both sides.
Decision speed is the share of applications that got any decision within 365 days, counting only those where a full year has passed. We don’t use average decision time, because schemes still waiting aren’t in that average, which makes recent years look faster than they are.
Workforce figures are ONS Business Register and Employment Survey (BRES) 2024, by region and industry code. Architects, surveyors, developers, manufacturers and merchants are excluded because they don’t work on site. ONS covers Great Britain, so Northern Ireland has planning data here, but no workforce figure, and ONS rounds its counts, so the smallest regions show a direction rather than a precise number.
We are not forecasting. We haven’t tried to work out how many people any individual building needs. We’re comparing each region’s share of the work with its share of the workforce, and showing where the two have come apart.
Everything on this page is free to use with credit. The charts carry a link back.
