Meta's $600 billion data center pitch counts construction jobs, not the automation it's building

What the $600 billion is actually for

Meta's post commits "over $600 billion in the US by 2028 to support AI technology, infrastructure, and workforce expansion." The framing is explicitly about American economic growth, and the company positions its data centers as the physical prerequisite for its product goal.

At Meta, we're focused on creating the next generation of AI products and building personal superintelligence for everyone. Data centers are crucial to reaching these goals and helping America maintain its technological edge.Montana Labs

So the through-line is clear: the buildings exist to run models. The economic-impact story is the wrapper around that. Reading it that way matters, because the jobs Meta counts sit on one side of a boundary the post never draws.

The labor math describes the buildout, not the operation

Since 2010, Meta says its data center projects have supported "over 30,000 skilled trade jobs and 5,000 operational jobs." That's a roughly six-to-one ratio favoring construction over operations — and the trade jobs are named in detail: steel workers, pipefitters, electricians, fiber technicians.

Trade jobs are project-bound. They exist while a site is being built and taper when it's finished. The 5,000 operational roles are the ones that persist once a facility is live, and they're the smaller number by design. A modern data center is engineered to be run by relatively few people; the whole point of the automation inside is that it doesn't need a large standing workforce.

The post also notes Meta is "currently bringing more than $20 billion in business to subcontractors across the US." That's real spending flowing to real firms — but it's spending tied to the construction cycle, not a recurring headcount the finished data centers will sustain.

Energy and water are the commitments with hard numbers

The most concrete, verifiable claims in the post are physical. Meta says it has enabled "hundreds of millions of new and updated grid infrastructure" and added "15 gigawatts of new energy" to US grids. It also states it pays for the energy costs that benefit its data centers and plans to be "water positive by 2030."

These are the figures that will age well or poorly on their own terms — grid capacity added, water restored to watersheds. Unlike the jobs numbers, they're not sensitive to whether you count a phase or a steady state. Fifteen gigawatts is fifteen gigawatts.

The community spending follows the same pattern of specificity: $58 million in Data Center Community Action Grants for schools, nonprofits, and projects, plus contributions to heating and cooling bill assistance. These are discrete, checkable commitments rather than projections.

The implication: infrastructure for automation is sold as employment

The honest description of this announcement is that Meta is building compute for AI products — including systems meant to do work — and presenting the construction of that compute as a jobs program. Both things are true at once. The steel and fiber work is genuine employment; the facilities it produces are optimized to operate with minimal human staffing.

For anyone assessing the local economic case, the useful question is which side of the line a given number sits on. Trade jobs and subcontractor spending are buildout economics that fade with the project. Operational jobs, grid capacity, and community grants are what remain after the last pipefitter leaves.

Meta's own closing line concedes the trajectory: "As the importance of AI grows, so will the importance of data centers." The buildings will keep getting built. Whether the durable local benefit is 5,000 operational roles per era of construction, or the 15 gigawatts and the grid work, is the distinction this post smooths over — and the one worth holding onto.