The argument against data centers often begins and ends with one statistic: they don’t create many permanent jobs.

But that statistic tells only one facet of the story. And when politicians use it to argue against building America’s AI infrastructure, my fellow citizens deserve to know what the politicians and media are leaving out.

Imagine a company proposes building a multibillion-dollar data center in your county. The construction will employ hundreds, perhaps thousands, of workers during the buildout. Once the facility opens, however, its permanent workforce may be relatively small. It takes surprisingly few people to operate a highly automated building full of computers. And here sits a massively, large monolith left behind.

Opponents to data centers seize on that fact. They point to the enormous building, the electricity it consumes, the water it may require, and the comparatively small number of permanent employees. They conclude that data centers are a bad deal for the community.

It’s an appealing political argument. It’s also an incomplete one.

The problem isn’t necessarily the statistics. It’s what those statistics leave out.

A Building Is Not an Economy

A recent study of 1,584 U.S. data centers, published in The Conversation, examined the relationship between these facilities and local employment. The researchers found that the direct employment footprint is modest, while the broader local employment effects develop over time.

There is no argument here on that snippet of useful information, but measuring the jobs inside a data center, or even the employment changes in the surrounding county, is not the same as measuring the economic activity that the computing infrastructure makes possible.

A data center isn’t a factory producing a finished product. A person does not walk into a data center pushing a shopping cart down isles of finished goods. It is infrastructure. It provides computing power that other businesses use to develop products, deliver services, conduct research, and build entirely new industries.

Think of it this way… An electric power station. We wouldn’t judge a power station’s economic value solely by counting the people who work inside its perimeter fence. We would take into account the factories, hospitals, laboratories, shops, and businesses that depend on the electricity it produces.

The same principle applies to data centers.

The Jobs You Don’t See

A data center supports an ecosystem of businesses and workers that may never set foot inside the building.

What is not brought up are the countless, employed, individuals who maintain the computing infrastructure, manage networks, secure systems, develop cloud platforms, and replace and upgrade the physical hardware as well as the software. There are the electrical engineers, equipment manufacturers, cooling-system specialists, contractors, and technical service providers.

But the real, much larger unseen benefit: What happens when thousands of businesses gain access to computing power that was previously too expensive or too difficult to obtain?

A small software company–– even a single individual––can now rent or gain access to the computing capacity needed instead of investing millions of their own hardware and software. An entrepreneur can develop an AI-powered service with a handful of people. Scientists can analyze enormous datasets. Manufacturers can deploy predictive maintenance. Small law firms can automate routine document processing. Businesses can develop specialized products for markets that previously couldn’t support them economically.

Some of these ventures will fail. Others will grow and hire employees, and still others will become suppliers to entirely new industries that did not exist.

These are the downstream economic effects that a simple count of permanent data center employees cannot capture, id one is simply looking at te four walls of a data center.

Not every new business will owe its existence to a particular data center, and AI can eliminate jobs as well as create them. But as I have stated ad nauseum:

There has never been a distruptive technology since the taming of fire that has not resulted in exponential growth of humanity.

The point is not that every facility guarantees an employment boom. It is that the economic consequences extend well beyond the facility’s payroll.

The Political Question Nobody Should Ignore

Why, then, are we so quick to reduce the debate to a building’s permanent head count?

Opponents of data center development have legitimate issues to raise. Electricity demand, water use, environmental effects, tax concessions, and the costs of expanding local infrastructure all deserve serious scrutiny.

But there is a difference between demanding that developers pay their fair share and arguing that the infrastructure itself has little economic value.

When politicians present a narrow employment statistic as though it settles the entire economic question, voters should ask what has been left out. And when opposition to data centers becomes a campaign talking point, voters should ask what alternative those politicians propose for maintaining America’s ability to compete in AI.

Here is the irony. Now a voter can use A.I. residing in these data centers to get answers to these questions, cutting through the political word-salad.

The political incentives are straightforward: the visible costs of a data center are immediate, while many of its potential economic benefits are dispersed, delayed, and difficult to attribute to a single facility. That makes the costs easy to campaign against and the benefits easy to overlook.

We should judge these projects on the evidence, not on which political narrative is most convenient.

If We Don’t Build It Here, Someone Else Will

This is the strategic question that deserves far more attention.

Assisted intelligence (A.I, S.I., call it what you want) is becoming a foundational technology, much as electricity, telecommunications, and the internet did before it. It will influence scientific research, medicine, manufacturing, transportation, financial services, national security, and countless businesses we cannot yet anticipate.

That requires computing infrastructure.

If the United States chooses not to build enough capacity to support its AI industry, the rest of the world will not necessarily make the same choice. Other countries can build the facilities, attract the investment, develop the technical workforce, and support the businesses that emerge around them.

And if American companies find the computing capacity they need elsewhere, some of the associated investment, expertise, supplier relationships, and entrepreneurial activity may develop abroad rather than at home.

That outcome isn’t inevitable. Businesses make location decisions for many reasons, and American companies can use computing resources located overseas. But the competitive risk is real:

The United States cannot assume it will lead the development of a technology while declining to build the infrastructure needed to support it.

There is a further irony here. A community that rejects a data center may avoid some local costs, but it also gives up the opportunity to negotiate for local benefits. The investment may go elsewhere, where another community can seek the construction activity, tax revenue, workforce training, and supporting businesses.

The question should not be whether every data center is worth having. It should be whether we are making informed decisions about which projects to accept, what they will cost, and what we can secure in return.

Don’t Just Accept the Developer’s Promises. Make Them Invest in the Community.

Here is where local government can do something constructive.

If a data center developer wants favorable zoning, public infrastructure support, or tax incentives, the county should negotiate a binding agreement that gives residents a stake in the opportunities the project creates.

For example, the county could require the developer to fund a local technology and entrepreneurship program.

That program could train residents in electrical systems, cooling, networking, cybersecurity, cloud computing, and data center operations. It could also provide pathways into AI software development, data analytics, automation, and other digital careers.

Local community colleges and technical schools could deliver the training. Employers could help design the curriculum. Apprenticeships and industry-recognized credentials could connect graduates with actual job opportunities.

And why stop at training people to work inside the facility?

The county could support small-business development, cloud-computing education, technical mentorship, and programs that help local entrepreneurs build products and services using AI.

The developer would help finance the program. The county would establish measurable outcomes, publish annual results, and require accountability for commitments made.

The goal would be to create a local workforce capable of participating in the broader technology economy, not merely competing for the handful of jobs inside the data center.

That investment could continue benefiting the community long after construction ends.

Stop Counting Buildings. Start Building Capabilities.

The argument over data centers shouldn’t be reduced to a choice between accepting every proposal and rejecting the entire industry.

Some projects will be poorly located. Some will demand unreasonable public subsidies. Some may impose infrastructure or environmental costs that outweigh their benefits. Those projects should face tough scrutiny.

But a blanket rejection based on the number of permanent employees misses the larger economic question.

We should be asking what infrastructure the country needs, what costs developers should bear, how communities can protect their resources, and how local residents can share in the resulting economic opportunities.

The United States has a choice. It can develop the infrastructure needed to compete in AI while negotiating better outcomes for its communities, or it can allow investment and some of the associated opportunities to flow to places more willing to build.

The real issue isn’t whether a data center employs thousands of people inside its walls. It’s whether the computing power inside those walls helps create businesses, careers, and economic opportunities—and whether the community hosting it has negotiated a fair share of those benefits.

Before accepting a politician’s claim that data centers are a bad deal, ask a few questions:

• Are we measuring only direct employment, or the wider economic effects?
• What are the actual costs to local taxpayers, ratepayers, and natural resources?
• What tax revenue and other benefits will the community receive?
• Can the developer be required to fund workforce training and local entrepreneurship?
• If this investment goes elsewhere, what opportunities might our community lose?

Those are the questions voters deserve answered.

Don’t confuse a small workforce inside a building with a small economic footprint. And don’t let an argument about the visible costs of new infrastructure substitute for a serious examination of the opportunities—and risks—that come with it.

Build wisely. Negotiate firmly. Train locally. Compete nationally.

That’s how a community can turn a data center from a controversial development proposal into an opportunity worth evaluating on its actual merits.

The last thing we want is for a foreign entity to control a US industry. Since 2020, we have seen real chokepoints: COVID supply chain, Hormuz cut-offs. Take a country like China who one day decides to cut off access to everything we now take for granted.

But if after reading this article you still feel data centers are a bad thing and you want to put a stop to them, you can achieve your goal with one simple change: stop taking pictures of yourself in the mirror. It takes approximately three and a half Super data centers to be built per year, to keep those images around.