The Data Center Labor Gap
in Virginia, Texas,
and the Midwest

Three U.S. regions are absorbing a disproportionate share of the data center construction boom: Northern Virginia, Texas, and a Midwest corridor running from Indiana through Ohio and Michigan. Each has a distinct labor market, a distinct set of projects, and a distinct version of the same problem — not enough engineers and skilled trades to build what's been committed to. Here's the breakdown.

Northern Virginia: The Market That's Already Maxed Out

Loudoun County, Virginia — widely known as "Data Center Alley" — hosts approximately 35% of all U.S. data center capacity. Microsoft, Amazon Web Services, Google, and Meta all have large active construction programs in the region. The combination of established fiber infrastructure, transmission capacity, and proximity to the federal government and federal contracting base has made Northern Virginia the default choice for hyperscaler expansion for over a decade.

The labor market consequence is straightforward: the region has been operating at capacity for years, and the AI investment wave has pushed it well past that point. MEP engineers with data center experience in the Loudoun County market are essentially fully employed. GCs winning new data center contracts in Northern Virginia cannot staff them from the local market — and they know it before they bid.

What's Being Built

Microsoft's Stargate-adjacent programs, Google's campus expansions in Loudoun and Prince William counties, and multiple co-location providers expanding to meet hyperscaler demand are all active simultaneously. The Virginia data center construction market was approximately $15.3 billion in active starts as of early 2026 — a figure that would be remarkable as a single year total, but represents work that is running in parallel across dozens of concurrent projects.

The Roles in Shortest Supply

In Northern Virginia, the acute shortage is in MEP engineering — particularly mechanical engineers with liquid cooling experience, electrical engineers familiar with large-scale power distribution architecture, and commissioning engineers who can verify system performance at hyperscale. Civil and structural engineers for the site and building work are also constrained, but MEP is the binding constraint on project schedules in this market.

GCs active in Northern Virginia who have not established cross-border sourcing relationships are running 90+ day fill times on MEP engineering roles and accepting the schedule consequences. Those who have established those relationships are mobilizing engineers in 4–6 weeks.

Texas: Scale Plus Complexity

Texas is the second-largest U.S. data center market, with $13.4 billion in active or planned construction starts concentrated in the Dallas-Fort Worth Metroplex and, increasingly, in San Antonio and the Austin corridor. The combination of available land, favorable business environment, and the deregulated ERCOT power market makes Texas a natural fit for hyperscaler expansion.

That same deregulated power market creates engineering complexity that other markets don't have. Data center power procurement on ERCOT requires engineers who understand energy market dynamics, demand response obligations, and interconnection requirements that don't exist in ISO-NE or PJM markets. This is a meaningful technical differentiator for MEP and electrical engineers working in Texas data center construction — and it narrows the pool further in a market that is already tight.

What's Being Built

Meta has multiple large campus projects across the Metroplex. Microsoft and Google have substantial Texas construction programs. Multiple co-location providers — Equinix, Digital Realty, CyrusOne — are expanding Texas capacity to meet enterprise demand that is moving to the region alongside the corporate headquarters relocations of the past several years. Balfour Beatty, HITT, and several regional GCs have significant Texas data center backlogs.

The Geographic Advantage for Cross-Border Hiring

Texas's geography is a genuine advantage for cross-border engineering placements under TN visa. The land border crossings at Laredo-Colombia Bridge, El Paso-Zaragoza, and other Texas ports of entry are among the most frequently used for TN adjudications. Mexican engineers traveling to Texas projects can cross at these ports and reach Dallas, Houston, or San Antonio by ground transportation the same day. The proximity reduces both the logistical complexity and the cultural adjustment burden for workers relocating from northern Mexico.

The Midwest: The Market No One Is Watching (But Should Be)

The Midwest data center and industrial construction market has received less attention than Virginia or Texas, but the scale of committed investment makes it one of the most significant labor markets in the country for the next several years.

Indiana and the $10B Campus

Mortenson and Turner are co-building what is being described as one of the largest single data center investments in U.S. history in Indiana — a campus valued at approximately $10 billion. Mortenson's CEO has been publicly vocal about labor being the #1 operational risk on the project. The Indianapolis labor market does not have the engineering depth to staff this project from local sources, and the project's size means it will absorb significant capacity from surrounding markets as well.

Ohio: Microsoft's New Albany Mega-Campus

Microsoft's announced mega-campus in New Albany, Ohio, combined with the Intel semiconductor fabrication plant construction in Licking County, has created a regional engineering talent crisis. Columbus-area GCs report the most severe shortage conditions they have experienced. The Intel fab construction alone requires a workforce of thousands of construction professionals — running simultaneously with Microsoft's campus buildout in the same labor market.

Michigan: OpenAI and the Manufacturing Overlap

OpenAI's announced 500-megawatt infrastructure project in Michigan sits in the backyard of firms like Barton Malow — a firm built on industrial and manufacturing construction that now finds its core markets (automotive, manufacturing, data infrastructure) all demanding engineers simultaneously. Michigan's labor market is further complicated by the EV manufacturing transition, which has pulled significant engineering and skilled trades capacity into EV plant construction across the state.

Why the Midwest Is Underserved

Staffing firms and cross-border placement providers are concentrated on the coasts and in Texas. Midwest GCs — Ryan Companies in Minneapolis, JE Dunn in Kansas City, Clayco in Chicago, Weitz in Des Moines — receive less vendor attention than their peers in Virginia or California, even as their labor markets become equally tight. This creates an opening for firms willing to engage in these markets with a real cross-border program, not just a coastal-market playbook applied to the Midwest.

The Common Thread Across All Three Markets

The pattern across Northern Virginia, Texas, and the Midwest is consistent: hyperscaler investment commitments are real and funded, construction is underway or imminent, and the local engineering labor markets are not equipped to staff the projects being built. This is not a short-term market tightness that will resolve when a few large projects complete — the announced pipeline of hyperscaler investment extends through at least 2028, and domestic engineering training programs cannot produce engineers fast enough to close the gap within that window.

What GCs in These Markets Are Actually Doing

The GCs who have begun building cross-border hiring capacity are doing so quietly — it is a competitive advantage they are not broadcasting. Common patterns:

• Establishing a relationship with a cross-border placement partner before they have an urgent need, so the first placement is not also a learning experience on a live project.
• Starting with engineering roles under TN, which have the fastest timelines and the lowest process complexity.
• Including cross-border staffing capacity as a line item in their bid assumptions for data center work in shortage markets — treating it as a known input, not a contingency.
• Combining domestic and cross-border sourcing simultaneously rather than sequentially — running both tracks from the first day a position is open rather than trying domestic for 60 days and then pivoting.