By Craig Bushon
Welcome to The Craig Bushon Show, where we don’t just follow the headlines… we read between the lines to get to the bottom line of what’s really going on.
Most people think the artificial intelligence revolution is happening inside their phone, their laptop, or inside a chatbot window.
That is not where the real transformation is happening.
The real transformation is happening in massive buildings being quietly constructed across the United States.
Recently I spoke with someone involved in building data centers for Meta Platforms. The scale of these facilities is almost difficult to comprehend. Some of them are approaching two million square feet. To put that into perspective, that is larger than many regional shopping malls.
Yet when fully operational, some of these facilities may employ only a few dozen people onsite.
Imagine a building larger than a mall, filled with some of the most advanced computing systems ever built, operating around the clock with perhaps forty employees managing the entire facility.
That reality reveals something important about the next phase of artificial intelligence. This revolution is not primarily about hiring more people. It is about building infrastructure that allows machines to perform more intellectual work than ever before.
Why This Matters
Artificial intelligence is often discussed as a software story—chatbots, search engines, and automation tools.
But the real transformation underway is physical.
Across the United States, technology companies are building enormous data centers that will power the next generation of AI systems. These facilities require unprecedented amounts of computing power, electricity, and physical infrastructure.
That buildout has implications far beyond the technology industry.
It affects energy policy, labor markets, construction demand, nuclear power development, and the long-term structure of the American economy.
Understanding what is happening behind the scenes helps explain why artificial intelligence may reshape not just software, but the physical and economic infrastructure of the country.
The Physical Engine Behind Artificial Intelligence
Artificial intelligence depends on three key ingredients.
Data.
Software architecture.
Computing power.
The public conversation tends to focus on algorithms and software, but computing power has quietly become the limiting factor.
Training modern AI systems requires extraordinary amounts of processing power. Instead of a single computer solving a problem, modern AI models run calculations across tens of thousands of processors simultaneously.
These processors perform trillions of calculations per second for days or even weeks at a time.
That is why the technology industry is racing to build hyperscale computing facilities across the country. These data centers are essentially industrial plants whose sole purpose is producing computation.
Inside these facilities are enormous clusters of advanced processors supplied by companies like NVIDIA. These clusters allow engineers to train AI models on massive datasets, refining them until they can perform complex tasks such as writing software, analyzing documents, diagnosing technical problems, or generating realistic images and video.
As more computing capacity comes online, the scale of AI models grows with it.
Many engineers believe that once the newest generation of data centers becomes operational, artificial intelligence capabilities could advance much faster than the public currently expects.
And these projects are no longer theoretical.
For example, Meta Platforms has announced plans for a massive AI data center campus known as Hyperion in Louisiana that could eventually scale toward roughly five gigawatts of power capacity — a level of electricity consumption comparable to several large nuclear power plants combined.
Meanwhile, Microsoft has received approvals for major data-center expansion in Wisconsin as technology companies race to secure the infrastructure needed to train future AI systems.
In other words, the AI arms race is no longer just about software.
It is about physical infrastructure.
Some technologists, including Elon Musk, have begun describing the potential acceleration created by this infrastructure as the approach of the technological singularity.
Understanding What the Singularity Means
The technological singularity is a theoretical moment when artificial intelligence becomes capable of improving itself faster than human engineers can manage.
If an AI system can design a better version of itself, and that improved system can repeat the process again, technological progress could begin accelerating exponentially.
Some researchers believe this could lead to artificial general intelligence — a system capable of performing most intellectual tasks at or beyond human levels.
Other experts believe the timeline is much longer, or that major breakthroughs are still required.
Regardless of the exact timing, the scale of infrastructure being built today suggests that technology companies are preparing for a world where artificial intelligence becomes far more powerful than the systems we use today.
Which raises an uncomfortable question.
What happens to jobs?
Why Some Parents Are Steering Their Children Toward the Trades
The person I spoke with has nine children. His advice to them was simple: consider learning a skilled trade.
His reasoning reflects a growing concern among people working closest to artificial intelligence development.
Many white-collar professions involve tasks that follow predictable patterns. Writing reports, summarizing documents, analyzing spreadsheets, drafting contracts, preparing marketing materials, and even writing routine software code all involve structured cognitive work.
Artificial intelligence systems are increasingly capable of performing these tasks.
That does not mean all white-collar work disappears. But it does mean that a single professional equipped with powerful AI tools may eventually perform the work that once required multiple employees.
Trades work operates in a completely different environment.
Electricians, plumbers, welders, HVAC technicians, mechanics, and construction specialists operate in physical spaces that are constantly changing. Every building is different. Every repair job presents unexpected challenges.
These professions require manual dexterity, situational awareness, and real-time problem solving.
Robotics is advancing quickly, but replacing the adaptability of a skilled tradesperson in unpredictable real-world environments remains one of the hardest challenges in engineering.
Many robotics researchers believe widespread automation of complex trade work could still be decades away.
But there is another important piece of the story.
The construction of the AI economy itself is creating a massive surge in demand for skilled labor.
Building hyperscale data centers requires enormous teams of electricians, pipefitters, welders, plumbers, structural steel workers, heavy equipment operators, and specialized technicians.
Each large facility can require thousands of workers during construction and months or years of specialized labor to complete.
Across the United States, the rapid expansion of AI infrastructure is already contributing to a skilled-labor boom in parts of the construction and engineering economy.
In other words, the technology that may automate some office work is simultaneously creating large new demand for physical infrastructure and the people capable of building it.
Electricity: The Hidden Constraint of Artificial Intelligence
Artificial intelligence runs on electricity.
Massive amounts of electricity.
A single advanced AI data center can consume hundreds of megawatts of continuous power. That is comparable to the electricity demand of tens of thousands of homes.
Unlike residential demand, AI training clusters cannot simply power down when electricity becomes scarce. Training runs often last for weeks and require uninterrupted power.
This creates a growing challenge for the technology industry.
Where will all of this electricity come from?
Some technology companies are exploring small modular nuclear reactors as a potential long-term solution.
Companies such as NuScale Power and TerraPower are developing reactor designs intended for industrial users.
However, despite growing excitement around this technology, the nuclear timeline for powering AI infrastructure is more complicated than many headlines suggest.
As of early 2026, no U.S. data center is currently powered by a newly constructed small modular reactor.
Most of the nuclear capacity being discussed for AI in the near term comes from existing nuclear plants that are being extended, upgraded, or potentially restarted to support growing electricity demand.
Technology companies are already moving to secure future power through long-term power purchase agreements, investments in nuclear technology companies, and partnerships with utilities exploring advanced reactor deployments.
Most energy analysts expect meaningful commercial SMR deployment to begin in the early-to-mid 2030s, with the potential to scale significantly in the decades that follow.
In other words, the nuclear-powered AI future being discussed today is real, but it is still several years away from widespread implementation.
For now, the AI boom is placing enormous pressure on existing electricity grids while energy companies and technology firms race to build the next generation of power infrastructure.
Read Between the Lines
Artificial intelligence is often described as a software revolution.
But what is quietly unfolding right now looks much more like a new industrial buildout.
Gigantic data centers.
Massive electricity demand.
Advanced semiconductor manufacturing.
And a rapidly growing need for skilled construction labor capable of building the physical backbone of the digital economy.
The public conversation about AI often focuses on algorithms, chatbots, and automation.
But the deeper story is about infrastructure — the factories of computation, the power plants that feed them, and the workforce required to build it all.
Bottom line
Artificial intelligence may reshape the economy, but its future will not be determined by software alone.
It will be determined by how quickly the United States can build the infrastructure that powers it — computing facilities, electricity generation, semiconductor supply chains, and the skilled workforce capable of constructing and maintaining these systems.
The quiet construction projects happening across America today may determine how fast the next technological era arrives.
And as always, on The Craig Bushon Show, we don’t just follow the headlines… we read between the lines to get to the bottom line of what’s really going on.
Craig Bushon
Host, The Craig Bushon Show
Disclaimer
This commentary reflects analysis and opinion from Craig Bushon based on publicly available information, industry reporting, and discussions with individuals familiar with technology and infrastructure development. Projections regarding artificial intelligence, labor markets, robotics capabilities, and energy infrastructure involve significant uncertainty and should not be interpreted as guarantees or precise forecasts. Technological development, regulatory decisions, economic conditions, and market dynamics may materially change the outcomes discussed. This content is provided for informational and educational purposes only and should not be considered financial, investment, career, or professional advice.
