Some AI data centers now draw more power than entire cities, triggering historic events. This episode explores what’s really powering the AI revolution, and why builders, not coders, are the ones determining how fast it actually arrives.
This episode explores how the AI boom is straining the American power grid, driving a historic natural gas construction wave, pushing up electricity bills, and forcing states like Michigan, Oregon, and Minnesota to rewrite the rules. It's a conversation about growth, infrastructure, and why builders are shaping the next era of intelligence.
Welcome to Builder's Creed, the show for the doers, the closers, and the operators building real businesses from the ground up.
Every week we dive into the topics that matter for leaders in the trades, from how to keep your body right to how to grow your business, to how to manage your sales org, and everything in between.
This isn't theory, it's the playbook straight from the people living it.
If you're building something of your own, you're in the right place.
Let's get into it.
Somewhere in the United States right now, there is a single building that is pulling more electricity off the grid than an entire mid-sized American city, a data center, a building full of racks and chips and blinking lights built to do one thing, think.
And here's the part that should stop you in your tracks.
We built the entire story of artificial intelligence around the idea that it's weightless, that it's code or cloud, that it lives somewhere out there, invisible, infinite, ready whenever you open an app and type a prompt.
But intelligence, it turns out, is not weightless at all.
It runs on turbines, it runs on transmission lines, it runs on natural gas plants, being built faster right now than at any point in American history.
It runs on electricians, welders, pipe fitters, and land developers.
The most futuristic technology on Earth is colliding head on with the oldest constraint there is physics.
And today we're going to talk about what that collision means for AI, for your electricity bill, and for anyone who builds anything for a living.
I'm Justin Brock.
Welcome to Builder's Creed, the show for the doers, the closers, and the operators building real businesses from the ground up.
Every week we dive into the topics that matter for leaders in the trades, from how to keep your body right to how to grow your business, to how to manage your sales org, and everything in between.
This isn't theory, it's the playbook, straight from the people living it.
If you're building something of your own, you're in the right place.
Now everyone is talking about AI as a software story, a model story, a chatbot story.
But underneath all of that noise is a much older, much more physical story.
A story about power plants, power lines, land, water, and the people who actually build the infrastructure that makes any of this possible.
This is a Builders Creed episode about the builders nobody's talking about, the ones keeping the lights on for the entire AI industry.
Let's get into it.
Let's start with the number that should reframe how you think about AI entirely.
Some of the new data centers being built right now, the massive ones, the ones built specifically to train and run AI models, consume more electricity than a mid-sized American city.
Sit with that for a second.
Tens of thousands of homes worth of continuous power funneled into a single facility running 24 hours a day, seven days a week, because AI training and AI inference don't take nights off.
And it's not one building, it's hundreds of them going up simultaneously.
Anywhere land is cheap and a utility thinks it can find the power to feed it.
Here's the thing about electricity demand in America for the last 20 years.
It was basically flat.
Then AI showed up, and in the space of about three years, that flat line turned into a line that looks like it's going straight up a cliff face.
That's not a gradual trend.
A utility can plan a decade around.
That's a demand shock.
And the American grid, built over the better part of a century, expanded slowly, deliberately, region by region, was not designed to absorb a shock like that.
So now we've got a genuinely strange situation.
The most advanced technology humanity has ever built is running straight into the limits of a power grid that in some places still has infrastructure from the 1960s holding it together.
Now here's where it gets interesting, and this is the part that I think most people miss.
For the last decade, the story of American energy has been a green story.
Wind and solar have gotten dramatically cheaper.
States set ambitious clean energy targets, but wind and solar projects take years to permit, finance, and build.
It's a slow, methodical process, and it was built for a slow, methodical world.
AI demand did not arrive at a slow, methodical pace.
It arrived like a tidal wave, and renewable energy, as fast as it's grown, simply cannot be built fast enough to match it.
So what fills the gap?
Natural gas, because natural gas plants can be permitted and built faster than wind and solar farm.
The AI boom has triggered what's being described as the largest construction boom of natural gas power plants in American history.
And it's not just new gas plants.
Utilities and power companies are also keeping aging coal plants running past their scheduled retirement dates.
Think about the irony sitting right at the center of that.
The industry that's supposed to represent the cutting edge of human progress, the frontier of digital intelligence, is in practice extending the life of coal plants and triggering the biggest fossil fuel construction boom in modern history: digital ambition, physical reality.
They don't always point the same direction.
Here's where this stops being an abstract energy policy conversation and starts being something that actually shows up in your mailbox.
Electricity bills are climbing in a lot of utility territories across the country right now.
And a meaningful part of that increase traces back to the strain that data centers are putting on regional grids.
Factories in the Rust Belt, actual manufacturing plants, actual blue-collar employers are seeing their power costs pushed up because they're now sharing a grid with AI facilities that need enormous, constant, guaranteed power.
This is the part of the AI story that doesn't make it into the keynote presentations.
When a tech company announces a new billion-dollar data center campus, it sounds like pure economic development, jobs, investment, tax revenue, and there is real value there, but there's also a cost that gets socialized across everyone else on that grid.
Households, small businesses, manufacturers, all of them competing for capacity with a customer that can outbid nearly anyone.
And this is exactly the kind of tension that builders understand better than almost anyone else in this economy.
Every builder knows that growth isn't free.
Every builder knows that when demand outpaces capacity, somebody pays the difference in cost, in delay, in strain.
The only question is who?
Right now, in a lot of these communities, the answer is everyone but the data center shareholders, unless the policy is written to prevent that.
Which brings us to what a handful of states are actually trying to do about it.
Michigan, Oregon, and Minnesota moved first on this.
All three passed laws in roughly the last year and a half aimed at making sure that new data center demand doesn't quietly bulldoze the clean energy commitments those states had already made for 2040.
Michigan's approach is the one I find most interesting because it's not a moral appeal, it's a financial lever.
If a hyperscale data center wants to keep a lucrative sales tax exemption in Michigan, it now has to hit 90% clean energy within six years.
No clean energy, no tax break.
That's not a suggestion.
That's a builder's incentive, structured the way a builder would actually respond to it, through the numbers, not the messaging.
Oregon and Minnesota took a slightly different route, ordering their utility regulators to make sure that whatever power gets built to serve data centers stays aligned with the state's existing emissions targets rather than letting that new demand quietly become the excuse to blow past them.
And this isn't staying contained to three states.
Similar bills have shown up in California, Illinois, New Jersey, Pennsylvania, and Virginia, states that combined host a huge share of the country's existing and planned data center capacity.
There's a quote from this fight that I think captures the whole tension perfectly.
An Oregon utility watchdog put it bluntly: the state's 2040 clean energy target was already a stretch before data centers showed up, and now it's hard to hit with them and hard to hit without them.
Because these companies bring real investment, real construction jobs, real tax revenue.
States don't want to turn that away.
They just don't want to torch their climate commitments to get it.
That's the tightrope.
Growth versus discipline, which, if you've built anything, a company, a building, a career, you already know is not a new tightrope at all.
It's just the AI era version of an argument builders have always had with growth.
Now, here's a twist that I think is actually the most builder-coded part of this entire story.
Some of the biggest tech companies aren't waiting around for utilities to catch up.
They're going out and building or directly funding their own power generation.
Google, for example, has been striking direct deals to bring new clean power online specifically to match its data center demand, geothermal projects, wind and solar paired with battery storage, agreements structured so the company effectively pays to expand the grid rather than simply drawing down capacity that already exists.
And there's a real logic behind why utilities are willing to play ball on these deals.
A utility that partners with a hyperscale data center on new generation gains a massive, reliable, long-term customer who's footing the bill to expand the grid.
The alternative, a company that gets frustrated with slow utility timelines and just builds its own standalone power plant off to the side, leaves the utility with nothing.
So increasingly, the negotiation isn't about whether new power gets built.
It's about who controls the queue, who owns the interconnection, and who captures the value of getting hooked up to the grid first.
That right there is the real fight happening beneath the surface of the AI boom.
Not is AI good or bad, not will the models get smarter.
The fight is over grid access, over interconnection cues, over permitting timelines, over who gets to build and how fast and on whose terms.
Which means, whether we like it or not, the next chapter of the AI story is going to be written as much by utility regulators and construction crews as it is by machine learning researchers.
Here's the pattern I want to leave you with because I think it's bigger than AI.
Every technology revolution eventually collides with infrastructure.
The railroad boom collided with the need for steel and track and land rights.
The automobile boom collided with the need for roads, refineries, and fuel distribution.
The internet boom collided with the need for fiber optic cable laid physically into the ground, ocean, and building walls.
Electrification itself, the original one over a century ago, collided with the need to build generation and transmission from scratch, city by city, wire by wire.
AI is no different.
It just took a little longer for people to notice because the first phase of this revolution felt purely digital.
You typed a prompt, you got an answer, and it felt like it happened out of thin air, but it never happens out of thin air.
It happens because somewhere a construction crew poured a foundation.
An electrician ran conduit, a turbine got shipped, installed, commissioned, a regulator approved an interconnection agreement after months of hearings.
A state legislature argued for a year about how to balance a tax exemption against a climate goal.
The intelligence economy still runs on the same stuff.
Every economy has always run on energy, land, labor, and the will to actually build something.
And that to me is the real headline hiding inside this story.
We keep being told that software is eating the world, that the future belongs entirely to the people who write code.
But the AI boom is proving in real time that the future still belongs just as much to the people who can build, the developers who can site a power plant, the crews who can run transmission lines faster than regulation usually allows the states smart enough to write policy that keeps growth and responsibility pointed in the same direction.
Software gave us the vision.
Builders are the ones who are going to have to make it real.
So if there's one thing I want you to walk away with today, it's this.
Don't let anyone tell you the physical world doesn't matter anymore.
The most advanced technology on the planet just ran straight into a power grid built decades ago.
And it's the electricians, the energy developers, the permitting offices, and the construction crews who are going to determine how fast this next era of intelligence actually arrives.
Not the other way around.
Intelligence needs electricity.
Electricity needs infrastructure, and infrastructure needs builders.
That's it for today's episode of Builders Creed.
If this made you think differently about what's actually powering the AI boom, share it with someone who needs to hear it.
And if you haven't yet, follow the show so you don't miss what's next.
Until next time, keep building.
That's a wrap on this one.
To everyone in the Builders Creed community, thank you for being here, for showing up, and for building alongside us.
None of this works without you.
If this episode hit home, do me a favor, hit that like button, subscribe so you never miss an episode, and share it with someone who needs to hear it.
Until next time, keep building.
EP-15Exercise may be doing far more than changing what you see in the mirror. Research into trained older muscle suggests long-term physical activity may help preserve parts of the molecular machinery tied to energy, adaptation and resilience.
EP-14Biology does not operate one gene or pathway at a time. AI is helping researchers model the interactions between those pieces, creating a new way to explore how cells behave, respond and change before every question reaches the lab.
EP-13Aging touches every system we build on; healthcare, families, businesses, economies. This episode breaks down how AI is turning longevity from something we react to into something we can predict, plan for, and build around.
Resources
