The Truth About the AI Boom: Why This Is Not a Bubble but a Buildout

The Truth About the AI Boom: Why This Is Not a Bubble but a Buildout
Featuring Brandon N. Owens — Hosted by Michael C. Vincent

Artificial intelligence is transforming the American landscape—not metaphorically, but physically. In this episode, host Michael C. Vincent sits down with Brandon N. Owens, one of the country’s leading voices at the intersection of AI and energy, to explore why the explosive growth of AI is not the next speculative bubble but the beginning of a vast, long-term industrial buildout.

Drawing on new research and reporting, Owens argues that the real story of the AI boom is not found in market valuations or venture capital enthusiasm, but in power-flow models, substation blueprints, transformer backlogs, and the load forecasts of utilities now revising decades of assumptions. Across the nation, electric grids are bending under unprecedented demand from hyperscale data centers—far faster, and far more dramatically, than planned. In Texas, data centers now consume an estimated fifteen percent of statewide electricity. In the Tennessee Valley, utilities are preparing for GPU clusters that could require a third of regional generation. In states like Ohio, Indiana, and Virginia, thirty years of anticipated load growth is collapsing into a single decade.

This is not the behavior of a hype cycle. This is what it looks like when a new industrial sector arrives.

Throughout the conversation, Owens traces the historical markers that define this moment. He draws parallels to the railroad boom of the 19th century, the electrification wave of the 1920s and ’30s, the interstate highway buildout of the 1950s, and the fiber-optic surge of the 1990s. In every case, skeptics misread early overbuilding as waste—only to discover that excess capacity became the backbone of future economic growth. According to Owens, AI follows that same arc: early uncertainty, rapid investment, infrastructure that outlasts its financers, and ultimately the emergence of a new economic system built atop the physical foundation laid during the buildout phase.

Michael presses Owens on the controversies now bubbling to the surface: growing tensions between hyperscalers and utilities; lawsuits over power delivery; interconnection queues stretched to breaking; water-use disputes in drought-stressed regions; and the looming mismatch between AI construction timelines and utility permitting processes. Owens explains why these challenges are not anomalies but signals of a structural transformation—one that demands the modernization of permitting, transmission, planning tools, and approach to large-scale load additions.

The discussion then widens to the global arena. Owens outlines how China is treating compute, power, and semiconductor capacity as strategic national assets, building new transmission corridors and dedicated AI zones. Europe, meanwhile, faces permitting bottlenecks and energy constraints that threaten to leave the continent behind. This global race for compute capacity echoes earlier eras when countries competed over steel output, electrification rates, and broadband penetration. The stakes, Owens argues, are no less consequential today: nations that control dense, reliable AI infrastructure will shape the economic and geopolitical landscape of the 21st century.

At its core, this episode makes a simple but profound case: the United States is not living through an AI bubble. It is living through the early stages of an industrial surge that will reshape energy systems, land-use patterns, regulatory structures, and national competitiveness. The question is not whether AI will scale—but whether the country will build the infrastructure fast enough, clean enough, and intelligently enough to unlock its potential.

Transcript

AIxEnergy Podcast — Episode Transcript

“Why This Is Not a Bubble but a Buildout”
Host: Michael C. Vincent
Guest: Brandon N. Owens

Michael Vincent:
Welcome to AIxEnergy. I’m your host, Michael Vincent.

Today, we’re again talking with Brandon Owens—founder of AIxEnergy. In the last podcast episode, we began a conversation about Cognitive Infrastructure Theory and The Cognitive Grid article series on AIxEnergy. We’ll return to that in future episodes, but today we want to tackle a topic ripped from the headlines: Are we in the middle of an AI bubble?

Brandon, welcome.

Brandon Owens:
Thanks, Michael. It’s great to be here. And it’s nice to talk about AI in terms of transformers and substations instead of pure hype cycles.

Michael:
Let’s start right there. You argue that what we’re seeing with AI is not a bubble, but a buildout. That’s a pretty sharp distinction. What does that mean in practical terms?

Brandon:
When people say “AI bubble,” they’re usually looking at stock charts and venture dollars. They’re asking, “Are valuations overheated?” That’s a fair question, but it misses the real story.

If this were just a bubble, it would be weightless. You wouldn’t see it in the physical world. But right now, you can literally fly over Texas, Virginia, Ohio, Tennessee and watch this boom take shape in steel, concrete, and copper.

We’re seeing the fastest surge in electricity demand since the dawn of electrification. Utilities that used to expect maybe one percent annual load growth are now ripping up their forecasts. In some regions, thirty years of expected growth has been compressed into ten.

That’s not a financial bubble. That’s an industrial buildout.

Michael:
You use a phrase in the article that really stuck with me—you say system planners are seeing “industrial shockwaves” on the grid. What does that look like from the inside?

Brandon:
It looks like planners suddenly living in dog years.

For decades, grid planning assumed slow, incremental change. You updated an integrated resource plan every three to five years. You added a new plant here, a line there. Nothing dramatic unless a major industrial customer arrived.

Now you’ve got developers asking for 300, 500, 1,000 megawatts of new load—and they want it energized in two or three years. That’s a different universe.

In Texas, data centers are already estimated to consume around fifteen percent of statewide electricity. In the Tennessee Valley, the utility is planning for GPU clusters that could eventually draw a third of regional generation. In the Midwest, utilities from Indiana to Ohio are literally telling regulators, “We’re compressing decades of growth into a single planning cycle.”

So the shockwave is this: tools, processes, governance models built for incremental change are being thrown into a world of step-changes. And they’re buckling under the pressure.

Michael:
That’s the demand side. Let’s talk about capital. You basically say: follow the money if you want to understand why this isn’t hype. Walk us through that.

Brandon:
Sure. If this were a classic tech bubble, you’d expect fast money—speculative capital that wants in and out quickly. Instead, we’re seeing long-duration, infrastructure-grade capital.

Meta has issued more than $26 billion in long-term data-center bonds. Microsoft is signing power contracts measured in decades and exploring dedicated nuclear and gas assets. Google is building data centers across multiple time zones simultaneously.

And behind them you have sovereign wealth funds, pension funds, infrastructure funds. These are institutions that invest in pipelines, airports, toll roads. They don’t chase fads. They look for durable, unavoidable demand. Their presence tells you this is not a passing phase.

Michael:
You draw some historical parallels—railroads, electrification, highways, fiber optics. Explain why those analogies matter.

Brandon:
Every U.S. infrastructure revolution follows the same pattern.

First, uncertainty. People say, “Is this worth the money?”

Second, overbuilding. Railroads in the 1880s, electrification in the 1920s, fiber in the 1990s—massive investment that critics called reckless.

Third, infrastructure outlasts the financers. Bankrupt railroads still carried freight. Bankrupt utilities still delivered power. “Excess” fiber became the backbone of the internet.

Finally, a new economic system emerges on top of the physical foundation.

What I’m arguing is that AI is following that same playbook. What people call “overbuild” is actually the steep capital curve required to unlock a new general-purpose technology.

Michael:
One of the strongest sections in your piece is about the grid as the limiting factor. You mention Amazon literally suing a utility over power delivery. That’s not a subtle signal.

Brandon:
No, it’s not. The lawsuit in Oregon captures the clash of two clocks: Amazon is building on a tech timetable—months and quarters. The utility is building on an infrastructure timetable—years and sometimes a decade.

When those clocks collide, something has to give. In Oregon, it was litigation. In Ohio, it’s regulatory debates about who pays for massive upgrades. In Virginia, it’s interconnection moratoria because the system can’t absorb any more large requests.

These are not the symptoms of hype. They’re the symptoms of a physical system pushed beyond what it was designed to handle.

Michael:
Let’s talk about the environmental side. High load, high water, large footprint. How should listeners think about the environmental ledger of AI?

Brandon:
AI is materially intensive. In 2023, U.S. data centers consumed tens of billions of liters of water—and that number is rising. Much of it is used for evaporative cooling. In Arizona, New Mexico, West Texas—these are real tradeoffs: AI campus cooling vs. agriculture vs. municipal supply.

The land footprint is massive. Some campuses cover hundreds or thousands of acres, plus associated substations and transmission corridors.

AI fits into the lineage of hydropower reshaping rivers, coal plants reshaping air-quality policy, highways carving through cities. If we get this buildout wrong, we’ll carry the consequences for decades.

Michael:
Now bring this down a level. What are the big decisions over the next five to ten years?

Brandon:
Three buckets: risks, challenges, opportunities.

The risk is that the grid becomes a brake on the AI economy—constraints, delays, reliability issues.

The challenge is that our institutional machinery—permitting, interconnection, planning—was built for a slower era. Reforming that machinery is hard and unavoidable.

The opportunity is enormous: a chance to modernize the grid, accelerate renewables and storage, build new transmission corridors, upgrade distribution, and harden resilience. AI can become the anchor customer that pays for the cleaner, stronger grid we need.

Michael:
You also say this looks less like an app cycle and more like an industrial arms race. What do you mean?

Brandon:
Globally, nations aren’t asking, “Do we like AI?” They’re asking, “Do we control the stack?”

China is building compute zones, new transmission, domestic semiconductor capacity. Europe is struggling with permitting and prices.

Historically, nations that mastered rail, electricity, telecom, broadband gained strategic advantages. AI is the next chapter in that story.

Michael:
Last question. The piece ends with: “This is not speculation. This is construction.” What do you want listeners to picture when they hear that?

Brandon:
The ground.

Concrete pads, cooling towers, gigawatt substations, high-voltage lines, chip fabs, hydrogen hubs, water lines. Steelworkers, line crews, control-room operators.

Once you see AI as a physical system—being poured and welded into the landscape—you realize we are building the infrastructure the 21st-century economy will run on.

The only question is whether we build it wisely.

Michael:
That’s a perfect place to stop. Brandon, thanks for joining us and for helping us see the AI boom not as a mood swing in markets, but as a buildout in steel and megawatts.

Brandon:
Thank you, Michael. It’s been a pleasure.

Michael:
You’ve been listening to the AIxEnergy podcast. Today’s episode—Why This Is Not a Bubble but a Buildout—featured Brandon Owens.

Join us next time. And visit AIxEnergy.io to stay current on the convergence of artificial intelligence and energy.