Construction · Operations · Payroll

How many jobs do data centers actually create?

A peak construction headcount is not permanent employment. A job-year is not a unique worker. A contractor on site is not necessarily a local resident. Count each quantity on its own terms.

Evidence reviewed: 10 September 2026Source periods: Virginia 2024 · Indiana 2024–2026
Evidence-based answerData centers can support very large but temporary construction workforces and smaller, often well-paid operating teams. The balance varies by campus phase and whether contractor roles are counted. There is no single national headcount that answers the headline.
~50Typical full-time workers for a 250,000-sq-ft facility, according to industry representatives cited by JLARC
~1,500Peak workers during a typical building’s construction, same source
12–18 mo.Typical construction duration per building, same source

Dated examples (labeled by claim class)

Attributed estimate · Virginia 2024

JLARC representative facility

Virginia’s Joint Legislative Audit and Review Commission reported that industry representatives described a typical 250,000-square-foot facility as having about 50 full-time workers (roughly half contractors) versus approximately 1,500 workers at peak construction for 12–18 months. These are representative estimates—not a census of Virginia facilities.

Government announcement · Indiana 25 Apr 2024

AWS New Carlisle baseline: at least 1,000 new jobs

IEDC announced AWS plans to create at least 1,000 new jobs with an $11 billion campus. That is a planned permanent-employment baseline, not observed payroll.

Company-linked · Indiana Jul 2026

“Nearly 1,000” / 1,056 full-time employees reported

An AWS-authored 9 July 2026 account said the site had nearly 1,000 operations staff; the South Bend Regional Chamber project page reported 1,056 full-time AWS employees. DataCenterData classifies these as company-reported, not independently verified until the project-specific county CF-1 filing, payroll definition and contractor treatment are reviewed.

What varies by place and project

Operating headcount scales with facility size, automation and whether security/maintenance contractors are counted. Multi-building campuses can show large peak construction even while each completed building adds only tens of permanent roles. Local residency shares are rarely published with the headline job number.

Four measures to keep separate (methodology)
Construction

Peak on-site workers

A snapshot of simultaneous labor demand. Include a date or phase; do not sum across months.

Duration

Job-years

One full-time-equivalent position for one year. Impact models often report job-years—not unique people.

Operations

Direct permanent employees

Operator payroll for the completed facility. Report residency separately when available.

Contractors

Ongoing contracted roles

Security, maintenance and technical contractors can be permanent on-site work without appearing on the operator’s employee count.

Remaining evidence gap and effectWithout the project-specific CF-1 (or equivalent) and a stated payroll definition, readers cannot tell whether company-linked counts meet, miss or redefine the 1,000-job baseline—or how many roles are contractors versus employees. Effect: permanent-job delivery stays “not yet verifiable,” even when company reports look close to the promise.
What to ask for, by nameAsk for four numbers separately: peak construction headcount with its date, job-years from any impact model, direct permanent employees on the operator’s payroll and on-site contractors. Then ask the county which annual compliance filing—Indiana calls it the CF-1—tests the permanent number.

Primary and attributed sources