Can your next US Data Center actually be staffed?
Abilene, Texas has 500 electricians in the entire metro. The Stargate campus has been running 8,000 workers. Our free 13-page report maps that gap across 47 US markets using two federal datasets, and the checker below runs your own site through it.
In brief
What the index found across 47 US markets
- Campus against metro
- Abilene, Texas has 500 electricians in total, while the Stargate campus there has run 6,000 to 8,000 workers. Monroe, the nearest metro to Meta's Richland Parish campus, has 1,590 people across all eight build occupations.
- Local payrolls
- Richland Parish, Louisiana added 149 construction jobs from Q4 2021 to Q4 2025, against 7,500 workers reported at peak. Average construction wages there rose 87.4%, against 18.9% nationally.
- Markets that grow
- Loudoun County, Virginia added 9,081 construction workers (up 53%) and Licking County, Ohio added 2,387 (up 71%), against national growth of 9.3%. Both had an existing metropolitan construction base to grow from.
- Imported labour
- A proposed workforce camp in Laramie County, Wyoming would house up to 5,600 workers, against a local skilled pool of 1,660.
- The professional layer
- Monroe, Louisiana has 220 people in its entire design, management and commissioning workforce. Commissioning managers and MEP design leads have no national referral system like the one that moves union electricians, so this is the layer that delays energisation.
- The case against
- Contractors estimate modular delivery saves 30% to 40% of on-site labour, and construction unemployment rose to 6.2% in June 2026. Specialist pay still climbed: commercial electricians' pay grew 9.9% year on year, against 2.4% for all US pay.
Sources: BLS Occupational Employment and Wage Statistics, May 2025; BLS Quarterly Census of Employment and Wages, Q4 2021 to Q4 2025; peak workforce figures as reported by developers and local government.
The checker
Pick a market, set your capacity, see what it takes to staff it
Every figure below is the total employed population of that occupation across every industry in the metro. None of these people are sitting idle waiting for a Data Center to arrive, which is what makes the ratio meaningful.
Already building here: Stargate (Oracle and OpenAI), reported at around 8,000 workers at peak, which is 4.4 times the entire local skilled population.
Converted at 0.7 to 2.0 construction workers per MW, the planning band published by the Hamm Institute for American Energy at Oklahoma State University.
Abilene, TX has enough depth to absorb this, but you are competing with every other employer in the region for the same people. Lead times matter more than headcount here.
Local skilled population: 1,820 people across all eight occupations, every industry in the metro.
Electricians, pipefitters, HVACR mechanics and site supervisors. Movable through the national dispatch system.
You would need roughly 59 of them. About 17% of everyone in the region doing this work.
Electricians are 1.37x more concentrated here than the national average, usually because another industry got there first.
This is one market. The report covers all 47.
Including the county payroll evidence behind these numbers, which markets grow a workforce and which only reprice, and what it has already cost the ones that could not.
Source: BLS Occupational Employment and Wage Statistics, May 2025 metropolitan release, covering 47 markets. OEWS counts workers at their place of work rather than residence, excludes the self-employed, and pools six semiannual panels over three years, so read these as the pre-existing local industry base rather than a headcount of who is on site today.
What the data shows
Six findings, including the one that contradicts our own headline
electricians in the whole of Abilene
One campus can be bigger than the metro it sits in
This is one of six findings. The full analysis covers all 47 markets.
Download the full report13 pages. Free, no mailing list.
Abilene, Texas has 500 electricians. Not 500 available electricians: 500 in total, across the entire metropolitan area, in every industry, on every job site. The Stargate campus on the north side of town has been running between 6,000 and 8,000 workers.
It is not an isolated case. Meta reports 7,500 or more skilled trade workers at peak for its 5 GW Richland Parish campus. The nearest metropolitan area, roughly 40 miles away, contains 1,590 people across all eight of the occupations that design, build and commission a Data Center.
Local pool is every worker in all eight occupations, across every industry in the metro. BLS OEWS, May 2025.
local jobs, against 7,500 reported at peak
The workers never showed up on local payrolls
If these campuses were staffed locally, county payroll records would show it. We pulled private construction employment for every campus county from Q4 2021 to Q4 2025 and set it against the national series.
Richland Parish is the cleanest experiment in the dataset. Meta says 7,500 or more workers at peak. Local construction employment rose by 149 people over four years. What moved instead was the price: average construction wages rose 87.4%, against a national baseline of 18.9%.
When a labour market cannot supply people, it does not fail. It reprices.
Dotted line is the US national baseline of +18.9%.
construction workers added in Loudoun County
Some markets grow a workforce. Others only reprice
This is the finding that contradicts our own headline, and it is the most useful one for site selection. Loudoun County added 9,081 construction workers in four years, up 53%. Licking County, Ohio added 2,387, up 71%. Both far outpaced national growth of 9.3%.
Where there is an existing metropolitan construction base, the local labour market expands to meet demand. Where there isn't, it doesn't. Loudoun and Licking had something to grow from. Richland Parish and Taylor County did not.
So the question for a site selection team is not how many workers are here. It is whether there is a base here that can grow.
Same counties, same period. The ranking inverts.
beds in one proposed workforce camp
What imported labour looks like when a county plans for it
In April 2026 a developer applied to build an 800-unit workforce housing camp in unincorporated Laramie County, Wyoming, capable of housing up to 5,600 workers, plus 150 RV spaces.
The county's planning director put the logic plainly: they did not want 10,000 to 12,000 people arriving on six figures and saturating the rental, housing and hotel markets. You do not build a 5,600-bed camp for a workforce you expect to recruit locally.
IBEW's own membership publication says much the same thing in a different way. Local 446 in Monroe, Louisiana has roughly 500 members and expects to need 1,000 to 2,000 electricians for the Meta project.
IBEW membership figures from The Electrical Worker. Camp capacity from the Laramie County planning application.
people in Monroe's entire professional layer
There is no Book 2 for commissioning managers
Almost every workforce analysis published on this sector has been about trades. That focus is understandable and it is incomplete. A campus does not stall because it is 40 electricians short. It stalls because the commissioning manager started three months late.
Monroe, Louisiana has 220 people in the entire design, management and commissioning population, across every industry in the region. Forty of them are electrical engineers. Meta is building 5 gigawatts there.
IBEW Book 2 is a functioning national market that moves journeyman electricians across state lines on a referral system built for exactly this. There is no equivalent for commissioning managers, MEP design leads or mission-critical estimators. The trades gap is a percentage problem you close with per diem and a wider catchment. The professional layer is a search problem, and it is the one that moves energisation dates.
Construction managers, electrical and mechanical engineers, and commercial and industrial equipment repairers combined.
on-site labour saved by modular delivery
Why this report might be wrong
Prefabrication is shrinking the problem. Contractors report labour savings of roughly 30% to 40%, and SemiAnalysis estimates modular approaches compress schedules by about 36%. Nearly every quantified estimate in this space comes from parties with a commercial interest, and none of it touches integration or commissioning, which get harder as more arrives pre-assembled.
The construction labour market is also loosening rather than tightening. Construction unemployment rose to 6.2% in June 2026 from 3.7% in April, and residential shed 48,800 jobs over the year.
Against that, nonresidential specialty trades added 63,700, and pay in the specialist end of the market is not behaving like a market with slack in it.
A loosening market does not pay its specialists four times the national rate. Bloomberg, June 2026.
Every client I speak to in the US is modelling power to four decimal places and treating the workforce as a line item they will sort out later. The data says that is the wrong way round.
Dylan runs Clear's US desk from the North America division, placing the people who design, build and commission hyperscale and colocation facilities. If you are planning a build in one of the markets covered here, he can tell you what the workforce picture looks like at role level rather than metro level.
The US Data Center Staffing Readiness Index
13 pages. 47 metropolitan markets. Two federal datasets, benchmarked against the national construction series.
- Which markets can grow a workforce, and which only reprice
- The wage premium already being paid in frontier markets
- Why the commissioning layer behaves differently to the trades
- The case against our own findings, argued in full
Method
How this was built, and what it cannot tell you
Workforce supply comes from the Bureau of Labor Statistics Occupational Employment and Wage Statistics, May 2025 metropolitan release. We used eight occupation codes covering the pre-operational lifecycle: electricians, plumbers and pipefitters and steamfitters, HVACR mechanics and first-line construction supervisors on the trades side, and construction managers, electrical engineers, mechanical engineers and commercial and industrial equipment repairers on the design, management and commissioning side.
County employment and wages come from the Quarterly Census of Employment and Wages, which is not a sample but a near-census drawn from state unemployment insurance records covering more than 95% of US jobs. Every county figure is benchmarked against the national construction series for the same quarters rather than against a single control county.
Limitations we would rather state than have you find
- OEWS measures the local industry base. It counts workers at their place of work rather than residence, excludes the self-employed, and pools six semiannual panels over three years, so a sudden influx is heavily damped.
- Peak workforce figures are self-reported. They come from developers and local government and nobody audits them. Meta revised its own Richland Parish figure up from 5,000 to 7,500 as the project grew. That is why the report tests them against federal payroll data rather than taking them on trust.
- We have not modelled commuting sheds. Workers living one county over appear in the wrong place, though an effect of the size seen in Richland Parish would still have to show up somewhere in state payroll data.
- Mission-critical experience is not in any dataset. Where the report says the qualified population in a frontier market is close to zero, that is a judgement drawn from the occupational data and from our own delivery experience, not a measurement.
Common questions
What is the US Data Center Staffing Readiness Index?
It is an analysis of whether US Data Center markets contain enough skilled people to build and commission the capacity that has been announced there. It covers 47 metropolitan markets and is built from two federal datasets rather than from a survey.
Where does the data come from?
Workforce supply comes from the Bureau of Labor Statistics Occupational Employment and Wage Statistics, May 2025 metropolitan release, using eight occupation codes covering design, build and commissioning. County employment and wages come from the Quarterly Census of Employment and Wages, a near-census drawn from state unemployment insurance records covering more than 95% of US jobs.
Does it cover Data Center operations staffing?
No, and that is deliberate. The index measures whether a facility can be delivered, not whether it can be run. Live operations headcount is a different question with a different answer, and mixing the two is how most workforce analysis in this sector goes wrong.
How many construction workers does a Data Center need?
The planning band published by the Hamm Institute for American Energy at Oklahoma State University is 0.7 to 2.0 construction workers per megawatt at peak. A 200 MW build therefore implies roughly 140 to 400 workers on site at the busiest point, before any allowance for schedule compression or modular delivery.
Why do some markets grow a workforce while others do not?
Where there is an existing metropolitan construction base, contractors, training pipelines and supply chains follow the demand and local employment expands. Loudoun County added 9,081 construction workers in four years. Where that base does not exist, employment barely moves and wages rise instead. Richland Parish, Louisiana added 149 construction jobs while average construction wages rose 87.4%.
Which roles are hardest to hire in Data Center construction?
The design, management and commissioning layer, rather than the trades. Journeyman electricians move across state lines through an established union referral system. Commissioning managers, MEP design leads and mission-critical estimators have no equivalent mechanism, and the qualified population in frontier markets is often close to zero.
Planning a build in one of these markets?
We will tell you what the workforce picture looks like at role level, not metro level.
Talk to our team