There's a growing movement of people fighting data center construction in their communities. I've been watching it, and honestly? I sympathize.
Not for the reasons you'd expect. I'm not anti-AI. I work with it daily. I think there's genuine promise in it, and when I explain what I do, I can articulate exactly how it's valuable without scaring the hell out of everyone in the room or overselling what it's capable of.
But I also work in the steel industry. I sit at the intersection of the tech world and the physical economy. And from where I'm sitting, the anti-data center movement makes perfect sense — because it's not really about data centers.
It's about contempt.
The Robber Barons Understood Something
Say what you want about the Carnegies and Rockefellers of the world. They were ruthless. They exploited labor. They accumulated wealth at a scale that was genuinely obscene.
But they understood a social contract that today's tech titans seem to have completely forgotten: if you get that rich, people are going to hate you. So you build hospitals. You fund libraries. You build infrastructure. Not because you're a good person — because you understand that the society you extracted from needs something back, or it's going to turn on you.
Now look at today's tech billionaires and ask yourself: where are the hospitals? Where are the libraries?
What you'll find instead is money dumped into politics, vanity space programs that fly celebrities around for press coverage, "philanthropy" that looks a lot more like experimentation than generosity, and AI companies run by people who seem to extract the life force from everyone around them. The lawsuits between these guys read like soap operas. The public statements range from tone-deaf to genuinely alarming.
I have a few soft spots here and there — some of these people are clearly trying, in their own way. But the general public's instinct to give the whole group a middle finger? I get it. When your idea of giving back is pouring money into the things that benefit you and calling it charity, people notice.
The Margins Tell the Story
I sit here watching customers who need a specific size of steel sheet. To get it, a plant has to operate a gigantic shearing machine. Four hours of work. Two people running it. A facility with cranes that can lift ten tons. And we're scrounging to get that sheet out the door for under a hundred bucks.
Meanwhile some tech company is skimming 25-30% off the top of restaurants by organizing people to drive food from point A to point B. That's the business. That's the innovation. And it's rewarded at three to four times the margin of the industry that puts steel in buildings and bridges.
Metal distribution tries to hold 10% margins when it can. Often accepts 6-8%. The capital investment is enormous. The labor is skilled and physical. The logistics are genuinely complex — we're moving tons of material, not chicken sandwiches.
And the economy has decided that organizing the chicken sandwich delivery is worth three times more.
That margin gap isn't just economics. It's a values statement. When your work makes that much money that easily, of course you start thinking you're smarter than everyone else. When your entire professional life is people in Patagonia vests congratulating each other on Series B rounds, of course you develop contempt for the guy operating a shear press.
I talk to people from that world. They're dicks. Not all of them — but enough that it's a culture, not a coincidence.
You Automated the Digital World Because You Built It
Here's the part that really gets me.
The reason AI can write code is because millions of developers threw their work on the internet for three decades. Stack Overflow. GitHub. Reddit. Forums. Documentation. Blog posts. The training data existed because the industry created it as a byproduct of doing the work. We built the digital world, and then we automated it using the record of how we built it.
Tech seems to have memory-holed this. They're out here talking about humanoid robots that'll do your laundry, wash your dishes, answer your front door, and replace any physical job a human can do. With the same confidence they had when they said "we'll teach AI to code."
Except they had the receipts for the coding claim. They had thirty years of data.
Where's your thirty years of data on roofing? Where's your Stack Overflow for plumbing? Where's your GitHub repository of every way a pipe fitting can go wrong in a house built in 1974?
It doesn't exist. The physical world's expertise lives in the hands and heads of people who've been doing the same job for twenty years. Guys who've seen every deviation from the manual. Who know that just because the spec says a certain screw size, that doesn't mean that screw size was actually used. Who understand tolerances not as an engineering concept but as Tuesday.
The Physical World Doesn't Compile
I've done enough DIY around my house to understand this viscerally.
A couple years ago I needed to replace the fan on my A/C unit. I watched probably thirty YouTube videos. Studied the wiring diagrams. Felt pretty confident.
And then I opened the unit and it looked nothing like any of the videos. There was an extra wire nobody mentioned. I spent hours figuring out what it was, where it went, and whether connecting it wrong would blow the capacitor.
Every single DIY project I've ever done has been like this. The YouTube video shows you the clean version. The version where the house was built to code, where the previous owner didn't improvise, where the parts match the manual. Your house is never that house.
The physical world doesn't compile clean. There's no syntax error that tells you which wire is wrong. There's no stack trace. You connect it and pray you didn't cross something that's going to start a fire.
AI is smart. It can explain parts of this really well. But explaining and doing are different things, and the gap between them in physical work is enormous. The value of a twenty-year HVAC veteran isn't that he knows the manual. It's that he's seen twenty years of things that didn't match the manual, and he's still here to tell you about them.
None of that experience is in any training set. And it's not going to be, because it doesn't get posted on the internet. It gets passed down on job sites, in apprenticeships, in the ten seconds of explanation a senior guy gives a junior guy while they're both looking at the same broken unit.
The Threat Plus the Ask
So let's put this together from the perspective of a normal person in a normal town.
A tech industry that openly looks down on you. That operates at four times your margins for a fraction of the physical effort. That's run by a rotating cast of billionaires who range from "eccentric but trying" to "genuine vampire." That's telling you, publicly, repeatedly, that it's building machines to take your job.
And now it wants to build a 40,000-acre facility in your community.
A facility that, by the way, won't employ many people once it's running. A data center isn't an Amazon fulfillment center. There's no bustling workforce inside. It's a giant building full of machines, maintained by a skeleton crew. It won't become a hub of local employment. It'll consume your land, your water, your power grid — and give you almost nothing back.
An Amazon warehouse, whatever you think of Amazon, is at least a hundred-thousand-square-foot argument that the building benefits the community. A data center is a hundred-thousand-square-foot argument that it doesn't.
And they justify it with "we need to beat China." As if that's your problem. As if the geopolitical AI race is something a town in Virginia or Ohio should sacrifice its water table for.
Of course people are protesting. The real question is why it took this long.
The Closed Loop
Here's the other thing that nobody talks about.
The money in tech doesn't leave tech.
A startup raises venture capital from a VC firm run by tech people. It spends that money on cloud computing from a tech company, advertising on a tech platform, office space in a tech hub, and salaries for tech workers who spend their money at tech-adjacent businesses.
The money circulates in a closed loop. Tech VCs fund tech bros who burn cash on tech services to build tech valuations so they can raise more from tech VCs.
The physical economy — the one that actually feeds people, houses people, builds the infrastructure everyone uses — operates on razor margins with massive capital requirements. And it gets to watch this closed loop of money congratulate itself from the outside.
Is this a sustainable culture? Is this an industry that has earned the trust of the communities it's now asking to host its infrastructure?
The Actual Point
I'm not anti-AI. I'm not anti-data center. I'm not anti-technology.
I'm a software engineer who's been in this industry for 17 years and who also works in the steel business. I see both sides of this every day. I talk to tech people. I talk to trades people. I watch the margins. I watch the attitudes.
The anti-data center movement is a symptom. The disease is a tech culture that has spent two decades extracting value from the physical economy while treating the people who run it as beneath them. That's now promising to automate a world it didn't build, using a playbook that only worked because it built the digital world itself. That needs massive physical infrastructure to deliver on that promise. And that can't even be bothered to make the case for why any of this benefits the communities it's asking to sacrifice for it.
The technology is real. The potential is real. I believe that because I work with it and I see what it can do.
But the culture around it is broken. And until the people building this stuff can stop long enough to listen to the people they claim they're going to serve — until they can explain the value of what they're building without threatening everyone's livelihood or waving the China flag — they're going to keep getting protested.
And they'll deserve it.
Tech's biggest vulnerability right now isn't a competitor or a regulation. It's that it forgot it needs the physical world more than the physical world needs it. And the physical world is starting to figure that out.