The Jobs Question: How Data Centers Compare to the Industries That Built Granite City

Walk through Lincoln Place on a quiet afternoon, past the old churches and gathering halls and the homes built by immigrant steelworkers a century ago, and you're walking through the physical evidence of what industrial employment actually meant for working families.

Those weren't just jobs. They were careers that spanned generations. They were wages that bought homes and sent kids to college. They were pensions that let people retire with dignity. They were the foundation on which an entire community was built.

When data centers are pitched as the new economic engine for communities like ours, the comparison to traditional industry is inevitable. So let's make it honestly: How do data center jobs actually stack up against the industries that built Granite City?

The answer isn't encouraging.

When the Niedringhaus brothers founded Granite City in 1891, they weren't just building a factory - they were building a company town designed to house the workers who would power American industry. The steel mills that followed drew wave after wave of immigrants: Bulgarians, Armenians, Hungarians, Macedonians, Mexicans, and others who came seeking the same thing - good jobs that could support a family.

At its peak, U.S. Steel's Granite City Works employed thousands of workers directly, with thousands more in related industries and services. A single mill could be the economic backbone of an entire region.

More importantly, these were jobs that didn't require a college degree but paid wages that could sustain a middle-class life. A steelworker could buy a house, raise a family, and retire with a pension. The work was hard and sometimes dangerous, but it offered something increasingly rare in the American economy: a path to stability.

That's what Granite City residents think of when they hear "industrial jobs." That's the benchmark against which any economic development pitch will be measured.

So what do data centers offer by comparison?

Let's start with the good news. Data center construction creates real jobs - often hundreds or even thousands of construction positions that can last a year or more. These are skilled trades jobs: electricians, plumbers, HVAC technicians, ironworkers. They pay well and they're local. For the duration of construction, a data center project can be a significant economic boost.

But here's where it gets complicated: construction jobs are temporary. Once the building is up and the servers are installed, most of those workers move on to the next project.

What remains is the operational workforce - the permanent employees who keep the facility running day to day. And that number is far smaller than most people expect.

Industry analysts estimate that a typical data center employs between 50 and 200 people once it's operational. Some larger facilities might employ a few hundred. But we're not talking about thousands of workers the way a factory or mill employs thousands.

When Apple built a billion-dollar data center in North Carolina, the permanent workforce numbered fewer than 100 people. When researchers studied the average job creation from new data centers, they found the number was so low - roughly 46 workers per facility - that it was "correctly interpreted as zero" in statistical terms.

A Wall Street Journal investigation put it bluntly: "Data centers are very labor-intensive to build, not as labor-intensive to operate."

The math is simple: data centers don't manufacture anything.

A steel mill takes raw materials and transforms them into products through processes that require human labor at every step. A car factory assembles thousands of parts into vehicles. A warehouse distribution center sorts and ships physical goods.

A data center is essentially a climate-controlled building full of computers that run themselves. The servers process data automatically. The cooling systems are largely automated. The security is often handled by cameras and card readers rather than guards at every door.

The workers who do remain are specialists: technicians who maintain the hardware, engineers who manage the systems, security personnel, facilities managers. Important jobs, often well-paying jobs - but not many of them.

Here's another way to think about it: A data center might cost a billion dollars to build but employ the same number of people as a mid-sized restaurant or a single floor of an office building.

To be fair, data center jobs aren't bad jobs. Industry surveys suggest that operational positions often pay well above local averages, particularly for technical roles. A data center technician might earn $50,000 to $80,000 depending on experience and location. Engineers and managers can earn significantly more.

These positions often don't require four-year degrees, though they do typically require technical training and certifications. For workers with the right skills, data centers can offer stable employment with decent wages and benefits.

But there's a catch: those workers often don't come from the local community.

One New York Times investigation found that construction jobs at data centers are frequently filled by specialized workers who travel from project to project across the country. These aren't local electricians and plumbers - they're data center specialists who follow the work wherever it goes.

And once the facility is operational, the technical positions may be filled by candidates recruited nationally rather than hired locally. A community might host a data center for decades without most residents ever working there or knowing anyone who does.

Data center proponents often point to indirect job creation to bolster the employment numbers. The argument goes like this: data center workers spend money locally, supporting restaurants and shops. The facility buys goods and services from local vendors. Construction creates demand for hotels and materials.

There's some truth to this. Any economic activity generates ripple effects. But the data center industry's claims about indirect jobs deserve scrutiny.

When OpenAI announced its Stargate AI venture and claimed it would create "hundreds of thousands" of jobs, industry analysts were skeptical. The math only works if you count every conceivable downstream effect - including jobs created by businesses that use the AI services the data center will power.

By that logic, you could credit a power plant with creating every job that uses electricity.

More realistic assessments focus on the local economic impact, which is limited by the small operational workforce and the fact that data centers buy relatively little locally. They need electricity and water, which they get from utilities. They need specialized equipment, which comes from national or international suppliers. The ongoing purchases that support local businesses are modest compared to a traditional manufacturing facility.

None of this is to say data centers are bad or that communities should automatically reject them. But it does mean we should be clear-eyed about what's actually being offered.

A data center in Granite City would not be a replacement for steel. It would not employ thousands of workers for generations. It would not be the foundation on which families build middle-class lives the way the mills once were.

It might bring a few hundred construction jobs for a year or two. It might bring 50 to 200 permanent positions, some of which might go to local residents. It might generate tax revenue, depending on what incentives the company negotiates.

Those aren't nothing. For a community of 27,000 people, 100 good-paying jobs would be meaningful. But it's a different kind of economic development than what built this town, and residents deserve to understand the difference.

There's a larger issue lurking beneath the jobs numbers, one that goes beyond any specific data center proposal.

For a century, American communities like Granite City offered a clear path to the middle class: show up, work hard, earn enough to buy a house and raise a family. The industries that provided that path - steel, auto manufacturing, meatpacking - employed massive workforces because the work required human hands.

The digital economy doesn't work that way. Data centers are capital-intensive, not labor-intensive. They require enormous investment but relatively few workers. The wealth they generate flows primarily to the companies that own them and the shareholders who profit from them.

This isn't a data center problem - it's an economy-wide shift that's been underway for decades. But it does raise questions about what "economic development" means in the 21st century.

If data centers are the future, what does that future look like for working families who don't have specialized technical skills? What happens to communities that can't offer the abundant cheap land, electricity, and water that data centers require? What's the path to middle-class stability when the new industries don't need workers the way the old ones did?

These aren't questions Granite City can answer on its own. But they're worth keeping in mind as we evaluate any specific proposal.

When someone pitches a data center as a job creator, ask for specifics:

The answers may still make the project worthwhile. Tax revenue, infrastructure improvements, and even modest job creation have value. But the pitch should be honest about what's actually being offered.

Data centers are not steel mills. They never will be. Communities that understand that going in will be better positioned to negotiate deals that actually serve their interests - and to avoid being sold a bill of goods dressed up in the language of economic revival.

This is the second article in a five-part series on data centers and what they could mean for Granite City. Next: "Follow the Money: Electric Bills, Tax Breaks, and Who Really Pays."

Have a story tip or want to get involved? Contact me at [email protected].

GraniteCitizen is just getting started. Be sure to sign up to get notified when new stories go live - business profiles, heritage features, local news, and more. No spam, just Granite City.