Data center job growth study: what 2.7 jobs per MW means for a site's community case

1. What the data center job growth study found

The data center job growth study behind this week’s headlines is by Simon Buechler, an assistant professor of finance at Miami University, and Lyndsey Rolheiser, a professor of urban and real estate economics at York University in Canada, with four colleagues. They summarised it in The Conversation on 27 October 2026. Phys.org and Fortune republished it.

The authors recorded the location, opening date and power capacity of 1,584 US data centers that opened from 2000 to 2023, then matched them with federal county data on employment, wages and business counts. They compared host counties with similar counties that had no data center.

Their main results, as the authors report them:

Finding	Figure
Extra nonconstruction jobs in the host county, three to five years after opening	roughly 1,000 to 1,300
Change in the first two years after opening	little
Jobs per MW installed, five years later	about 7.4
Later expansion linked with each MW	about 1.7 MW
Jobs per MW of eventual capacity	about 2.7
Example: a 100 MW facility	roughly 275 jobs outside the data center

The gains were concentrated in transportation and warehousing, professional and technical services, and real estate. The authors found little evidence of growth in restaurants, stores and other local services. Further investment tended to stay in the same county, mostly as expansions by operators already there.

“Data center job growth study” is a rising related search under “data center” in Google Trends this week, up 14,650%.

2. How to use 2.7 jobs per MW in a planning case

Planning committees ask developers how many jobs a data center will bring. On-site staff numbers are small, so most of the case rests on jobs elsewhere in the county. The study gives a dated, sourced figure for those.

Use the right per MW figure. The 7.4 jobs per MW figure counts the first megawatt plus the 1.7 MW of expansion that tends to follow it. If your application covers the full build-out, the authors’ 2.7 jobs per MW of eventual capacity is the figure that matches. A 100 MW campus then comes to roughly 275 county jobs, which is the authors’ own example.

Keep the on-site number separate. The authors write that a data center typically employs no more than a few dozen people inside the building. Put on-site operations staff, construction jobs and county-wide indirect jobs in separate lines, so a planning officer can check each one.

Give the timing. The study found little change in the first two years, with gains three to five years after opening. A community case that promises jobs on opening day is promising something this evidence does not show.

Name the sectors. Transportation and warehousing, professional services and real estate are where the study found growth. If the council expects new shops and restaurants, the study found little growth there.

Present it as an average across US host counties. The authors describe their per MW results as associations, and an application that says “associated with” claims no more than the study does.

3. What the study does not prove

The authors list the limits themselves.

The data stops before the large AI wave. It covers facilities opened through 2023, before the current wave of very large AI data centers. The authors add that larger data centers do not necessarily create proportionally more jobs, so the results may not apply to the biggest projects now in planning. Scaling 2.7 jobs per MW to a 1 GW campus is a guess the study does not support.

The study measures jobs only. In the authors’ words: “We also looked only at job gains and did not explore whether data centers are worth their cost to local communities.” Electricity and water use, noise and tax breaks are outside the study, and opponents at a hearing will raise them. For the grid side, see data center power line disputes.

All 1,584 sites are in the United States, and the comparison is between US counties. None of the sources cited here gives a European figure, so an application in Europe should present it as US evidence.

The figures are averages across host counties. They do not forecast the result for one site.

4. Data center job postings in 2026

Hiring for data centers is rising, which is a separate question from county employment. Indeed Hiring Lab’s US Labor Market Snapshot of 22 October 2026 says data center related job postings grew another 20% in the three months through September and are now 147% above their June 2024 level. In July, Hiring Lab had thought postings were reaching a plateau.

Its earlier post, Hiring for the Data Center Build-Out of 14 July 2026, found six in every 1,000 US job postings were data center related. About a quarter of data center openings were for installation and maintenance workers, and the largest 10 tech firms accounted for 71% of data center postings in 2026. The authors left open whether those jobs will be permanent once the centers are running.

Postings count openings anywhere in the country, including construction and roles at headquarters. They do not tell a council how many people its county will employ. For a planning case, the county study is the closer fit, and the postings data shows demand for the trades a build-out needs.

5. A checklist for the jobs section of an application

1. State the eventual capacity in MW and apply about 2.7 nonconstruction jobs per MW, citing Buechler and Rolheiser’s study.
2. List on-site operations staff from your own staffing plan, separately from county-wide jobs.
3. List construction jobs separately. The study’s figures exclude them.
4. Say when the jobs arrive: little change in years one and two, gains in years three to five in the study.
5. Name the sectors the study found: transportation and warehousing, professional and technical services, real estate.
6. Note the limits: US counties, facilities opened through 2023, jobs only.
7. Answer the power, water, noise and tax questions with their own evidence.