I keep hearing about all the jobs AI is going to create to replace those jobs it will eliminate.

Every press release, every economic development announcement, every ribbon-cutting photo op seems to promise a future overflowing with opportunity. Trillions of dollars flowing into AI factories and hyperscale data centers. Regions being “transformed.” A new industrial era powered by intelligence at scale.

Every once in a while, you get a picture behind the black curtain, and you see numbers that actually matter.

Consider the new data center project planned for North Fort Worth, Texas, detailed in this Houston Chronicle report:

On paper, it’s a stunner. Roughly $2.1 billion in projected investment. A next-generation facility optimized for efficiency. Promises of minimal water usage, though what “minimal” really means remains conveniently undefined. The kind of project local leaders love to tout as proof that their city is winning the future. This multi-phase project may take 10 years to complete.

Now here’s the number that stops you cold: 37.

That’s the number of full-time permanent jobs the project is expected to create. Thirty-seven (to be fair, that might be just in phase 1; the information is not clear). The average salary is reportedly around $150,000 per year, which makes them very good jobs. But still — thirty-seven.

A $2.1 billion project. Thirty-seven people.

The project will generate property tax revenue, of course. That’s part of the sales pitch. But even there, the Chronicle notes that tax abatements of up to 70% are being discussed. This is how modern economic development works: Massive capital investment, generous incentives, and a razor-thin human footprint.

This isn’t an outlier. It’s the model.

Modern data centers and AI factories are designed to eliminate friction, inefficiency and people. Software-defined infrastructure. Autonomous operations. Predictive maintenance. Lights-out data halls. If you need hundreds or thousands of workers to run your AI infrastructure, you’re doing it wrong.

And let’s be clear: This isn’t a moral judgment. It’s arithmetic.

We’ve seen this before. Manufacturing went through the same transformation decades ago. Factories that once employed thousands now operate with a fraction of the workforce thanks to robotics, automation, and software-driven optimization. Output went up. Costs went down. Jobs vanished. The system worked exactly as it was designed to.

AI infrastructure is simply the next chapter in that story.

Yes, the jobs that remain are good jobs: Highly skilled, well compensated, and intellectually demanding. But they are few and far between. We are pouring unprecedented amounts of capital into systems explicitly built to minimize human labor. That’s not a bug. That’s the feature.

Market forces reward efficient use of capital. Always have. Always will. And don’t you ever forget that.

This is usually where the conversation veers into wishful thinking. “Technology will create new jobs.” “We’ve always adapted before.” Sometimes that’s true. Sometimes it’s just a comforting reflex.

But here’s the part of the story that deserves more attention.

While AI factories and data centers may not employ many people directly, the capabilities they unlock may enable entirely new industries, new frontiers, and yes — new kinds of work. Not jobs in the traditional sense of tending machines, but roles built on what these systems make possible.

One of the ideas gaining real traction — even among some of the architects of AI themselves — is space.

Space as the next frontier. Space as the next adventure. Space as the next economic expansion zone.

Now, what exactly do those jobs look like? I don’t know. Nobody really does. But history suggests that when foundational technologies reach a certain level of maturity, they enable things that previously sounded like science fiction. AI-driven automation, advanced robotics, and massive compute capacity could make large-scale space exploration, manufacturing, mining, and habitation economically viable in ways that were unthinkable even a decade ago.

And for all my fellow Trekkie friends out there, let’s be honest, this is kind of what we’ve been waiting for.

A future where humanity isn’t just optimizing ad clicks and shaving milliseconds off cloud latency, but actually expanding outward. New careers that don’t yet exist. New disciplines that combine engineering, science, exploration, and imagination. Not guaranteed. Not inevitable. But possible.

Still, even that optimistic view doesn’t change the near-term reality.

The transition period is going to be messy.

AI-driven productivity gains are arriving faster than our social and economic structures can absorb them. This isn’t just a white-collar issue. We haven’t even scratched the surface of what happens when physical AI and robotics mature enough to disrupt skilled and semi-skilled labor — construction, logistics, transportation, maintenance and manufacturing. When machines don’t just think, but move, adapt and act in the physical world.

The same math applies there, too.

Fewer people. Higher output. More capital efficiency.

There’s another temptation we should resist: Inventing busy-work jobs just to keep people employed. That’s not progress; it’s denial. The goal of technology has never been to maximize employment. It’s been to maximize capability, efficiency, and value. Sometimes job creation follows. Increasingly, it doesn’t.

The North Fort Worth data center will likely be profitable. It will generate tax revenue. It will contribute to economic growth. From a capital perspective, it makes perfect sense. From a labor perspective, it forces us to confront a truth we’ve been politely avoiding.

Our economic system still assumes employment is the primary mechanism for distributing income and purpose. But what happens when the most valuable assets in the economy don’t need many people at all?

Maybe some of that productivity eventually funds new social models. Maybe universal basic income moves from academic debate to practical necessity. Maybe we should rethink how value is shared when human labor is no longer the primary input.

Or maybe, true to form, we stumble forward and figure it out after the fact.

What I do know is this: As we head into 2026, this is no longer theoretical. The Fort Worth project isn’t a warning sign — it’s the future unfolding in front of our eyes right now. AI infrastructure is capital-heavy, automation-first, and largely indifferent to traditional job counts.

It will create enormous value. It may even enable entirely new frontiers of work we can barely imagine today.

But it will also force some very uncomfortable conversations about jobs, income, and meaning in an economy where efficiency wins—and headcount loses.

Thirty-seven jobs. Two-point-one billion dollars.

That’s the new math. And whether we like it or not, the future is here. We need to adapt and adjust.