The First Labor-Market Signal from AI Data Centers: Skilled-Trade Payroll Before Permanent Jobs
The AI infrastructure boom is young, but it is not too young to measure its first labor-market channel. This paper studies eight high-confidence U.S. counties with AI-specific or AI-boom hyperscaler data-center milestones in 2024 and links them to BLS Quarterly Census of Employment and Wages county-industry records through 2025 Q3. The main result is clear: the first measurable local labor signal from AI data centers is in power infrastructure work, not permanent data-center operation. In NAICS 237130, power and communication line construction, average weekly wages rise 18.6 percent relative to matched controls over event quarters 0-4, and all four contributing treated counties have positive wage effects. The result survives the main identification checks: the leave-one-county-out minimum is 13.8 percent, the estimate remains positive under treatment timing shifts of plus or minus two quarters, augmented matching on pre-boom power-line levels and wage slopes leaves a minimum effect of 11.9 percent, six placebo sectors have wage effects no larger than 2.1 percent, and a within-match randomization test gives a one-sided p=0.002. Payroll in the same industry rises 23.8 percent on average but is more dispersed. By contrast, the combined skilled-trades index is not positive, and directly AI-specific data-processing employment has only one disclosed treated-county observation. AI data centers are already visible in local labor data, but the visible channel is concentrated in grid-adjacent construction trades before permanent operating jobs appear in public employment records.
Reviews
The paper asks a well-posed early-signal question and leverages a strong administrative source (QCEW) with a transparent event-time framing and a plausible mechanism: grid-adjacent construction should respond before steady-state data-center operations staffing appears in public employment records. The headline finding—an ~18.6% relative increase in average weekly wages in NAICS 237130 over event quarters 0–4 across eight treated counties—appears internally consistent with the narrative, and the author provides several sensitivity checks (leave-one-out, timing shifts, augmented matching, placebo sectors, and a within-match randomization test) that, if implemented correctly, are directionally reassuring. The focus on wages (rather than only employment) is also sensible for tight skilled-trade labor markets where hours/overtime and pay premia can move before headcount. The main uncertainty is identification and measurement: “treatment” is based on project milestones rather than verified construction start dates, and the treated set is hand-curated and small, which increases sensitivity to selection and to coincident regional shocks (utility capex cycles, storm rebuilds, large non-AI infrastructure projects, union contract renewals, or prevailing-wage policy changes) that could differentially move 237130 wages. Matching on a limited set of pre-2022 levels (and some slopes) may not ensure parallel trends in 237130 wages specifically, and the paper excerpt does not specify whether inference accounts for the small number of treated clusters (8) and repeated-quarter correlation. The conclusion that “AI data centers are already visible” is plausible but should be stated more narrowly as evidence consistent with AI-data-center-associated infrastructure buildouts correlating with local power-line construction wage increases, rather than a clean causal estimate, until the event-study pre-trends, standard errors/CI construction, and alternative explanations are more fully ruled