I was watching a factory tour video last night.

The workers were Texans. A lot of them young. A few of them, by the look of it, were probably working a fast-food drive-thru window not long ago.

And they're building the most advanced communications satellites ever designed.

Placing components on phased arrays the size of a tennis court. Running calibration sequences. Hitting tolerances that the legacy aerospace primes — Boeing, Lockheed, Northrop — would tell you are impossible without a PhD-heavy workforce and a five-year program timeline.

The facility is in Midland, Texas. Permian Basin. Oil country. The satellites weigh several tons, with 2,400 square feet of unfurled phased array — physically comparable to traditional geostationary communications satellites that have historically cost hundreds of millions each and taken three to five years to build.

These are being built on a target cadence of multiple units per month at roughly an order of magnitude lower unit cost. In weeks instead of years. These are AST SpaceMobile’s (ASTS) BlueBird-class satellites — what I’m calling gigasats here because of the massive scale.

And the workforce doing this? The workforce that 40 years of cable news told you couldn't do advanced manufacturing anymore?

It's Texans, mostly young ones, who weren’t building satellites six months ago. Now they are.

America Is Building the Impossible

Midland is one site in something far bigger. It's not an isolated story.

Drive 12 hours east from Midland, and you'll hit Memphis. A few months ago, Elon Musk's xAI stood up the largest AI training cluster on the planet — 100,000 Nvidia H100 GPUs, networked, liquid-cooled, operational — in roughly four months. Jensen Huang publicly called it impossible-but-done. Industry consensus said two years minimum.

Memphis tradesmen did it in just four months…and they've since scaled past 500,000 GPUs.

Drive west to Phoenix, and you'll find TSMC's Arizona fab. The fab people spent years saying could never match Taiwan's yields. The fab built by ex-retail, ex-service, ex-military Arizonans who were given good tools and trusted to use them.

It is now beating Taiwan's yields.

Cutting-edge semi is the most complex industrial process humans perform — and the Phoenix line is winning.

Drive to Boca Chica and watch SpaceX. South Texas welders and former oil-field workers, building Starship — the most complex flight hardware in human history — at a cadence Aerojet Rocketdyne couldn't dream of with their credentialed legacy workforce.

Drive to Mississippi, and Anduril is building Arsenal-1. Drive to Ohio, and Intel is building in Silicon Heartland. Drive to upstate New York, and Micron is building America’s largest memory fab. And in Abilene, Texas, the $500 billion Stargate AI campus is rising out of the prairie.

The last U.S. aspirin plant closed in 2002. The last domestic penicillin plant in 2004.

But drive to Lebanon, Indiana or Clayton, North Carolina and you'll see something else: the full reconstruction of American pharmaceutical manufacturing.

Eli Lilly alone is putting $50 billion into four new U.S. plants. AstraZeneca has committed another $50 billion by 2030.

COVID exposed the strategic catastrophe of that offshoring, and the response is now visible from orbit — being built by Hoosiers, North Carolinians, and Midwestern building trades who, six months ago, were not building GMP-grade bioprocessing facilities.

Now they are.

The American hyperscale data center buildout alone is running at roughly $320 billion per year in capex across Microsoft, Meta, Google, and Amazon. Total committed capex through 2030 is conservatively a trillion dollars. Some estimates push two trillion.

For context: the Manhattan Project cost about $35 billion in today's dollars and employed roughly 130,000 workers across three primary sites.

The current American AI and advanced manufacturing buildout is, by every measurable input, somewhere between 30 and 60 times larger. Multiples more workers. Dozens of sites.

And almost nobody is calling it what it is — measured in dollars committed, the largest industrial mobilization in American history.

Here's why.

A New Era of American Industrialization

Everyone assumes American manufacturing labor is a declining input. Every analyst note worries about "skills gaps." Every consultant deck explains why complex assembly must be offshored.

They're all working from a map of a country that no longer exists.

The country that does exist is reshoring its industrial base at unprecedented speed, in the towns the consensus wrote off. Midland. Memphis. Abilene. Phoenix. Boca Chica. Lordstown. Columbus. Lebanon. Clayton.

The limiting factor was never the worker. It was the system around the worker. Legacy primes deskilled, offshored, treated labor as a cost line to minimize.

The new builders — SpaceX, ASTS, Anduril, Hadrian, TSMC Arizona, the hyperscalers — inverted every one of those assumptions. Good tooling. Real training. Standard work. Trust. Fair pay. AI-assisted process control.

And here's the part that should stop you cold: this workforce isn't just doing what the Manhattan Project workforce did. It's doing it with tools the Manhattan Project workforce couldn't have imagined.

The kid in Midland placing a component on a phased array now has real-time AI-assisted process control telling him whether the placement is within tolerance before he even lifts his hands.

The Memphis electrician wiring a GPU rack is working from prints generated, optimized, and stress-tested by software that didn't exist three years ago.

The Arizona fab technician is running calibration sequences on equipment that self-corrects in milliseconds.

The Lebanon pipefitter is laying out a bioprocessing facility that was digitally twinned, walked through in VR, and pre-validated before a single beam went up.

Robotic arms. Computer vision. Predictive maintenance. Real-time defect detection. Generative design. AI-optimized scheduling. Autonomous material handling.

None of this is automatic. Power interconnection queues, skilled-trades pipelines, and regulatory timelines remain real constraints. But the velocity and the inversion of prior assumptions are unmistakable.

The Greatest Generation built the atomic bomb with slide rules, hand calculations, and human intuition refined over years of practice. They built something miraculous despite their tools, not because of them. They were great men formed by hard times and harder work, and they earned every bit of the mythology.

But the men and women being formed in Midland and Memphis and Lebanon and Phoenix right now are being formed by something different.

They are learning to operate at productivity levels that would have been incomprehensible to the WWII generation, no matter how skilled or how dedicated.

And none of this runs without energy.

America, Powered By American Energy

The United States is currently producing more oil than it ever has. More oil than any country has ever produced, ever. Over 13.5 million barrels per day, exceeding Saudi Arabia, exceeding Russia, exceeding the U.S. conventional peak that the experts of the 1970s said could never be surpassed.

Natural gas: same story. Over 105 billion cubic feet per day. The United States is now the world's largest LNG exporter. Permian associated gas alone exceeds the total output of most OPEC nations.

And where is most of this energy being produced? The Permian Basin. Around Midland. The same Midland where they're building gigasats. The same West Texas geography where Stargate is rising in Abilene. The same Gulf Coast where Starship is being built in Boca Chica. The same TVA grid powering xAI Colossus in Memphis.

The American manufacturing renaissance is sitting, almost literally, on top of the American energy renaissance — and being built by the same archetype of worker. The Permian roughneck and the Midland gigafab tech might be brothers — or the same man at different points in his career.

The roughneck didn't get the memo either. He was told for a decade that his industry was dying, that his work was obsolete, that capital was abandoning him for ESG-compliant alternatives. He kept drilling. He kept innovating.

Horizontal wells, multistage fracking, real-time downhole telemetry, AI-optimized completions.

Same grit. New leverage. A new kind of man, building a new kind of country, on top of an energy base that the rest of the world cannot match.

And the energy story isn't just hydrocarbons. It's what's coming next.

America is quietly preparing the largest nuclear buildout since the 1970s. Small modular reactors are being designed, licensed, and contracted to power the same hyperscaler campuses we just walked through.

Amazon has signed for nuclear capacity at Susquehanna. Microsoft paid to restart Three Mile Island. Google has contracted with Kairos. Meta is in active negotiations. TVA is racing toward the first U.S. utility-scale SMR at Clinch River. Permitting timelines are being compressed from a decade to under three years.

This is the first American nuclear renaissance in 50 years, and it is being driven by the AI buildout. The data centers need power. Wind and solar can't provide baseload at the scale required. Gas turbines bridge the gap today. Fission is the answer for tomorrow.

And it doesn't stop at the grid.

The Next Frontier

NASA's Fission Surface Power program is contracting reactors for the lunar surface. DARPA's DRACO program is building nuclear thermal propulsion engines for crewed Mars missions, with a demonstrator targeted for orbital test inside three years.

BWXT, Lockheed, and a new generation of nuclear-focused startups are building the reactors that will power humanity's permanent presence beyond Earth.

The same country producing more oil than any nation in history is simultaneously building the next generation of fission reactors for its data centers and the first generation of fission reactors for its spacecraft.

That is not a country in decline. That is a country reaching escape velocity.

The result is the emergence of the most valuable, most mispriced, most underestimated combination of assets in the developed world: untapped American human capital, amplified by a technology stack that didn't exist a decade ago, powered by an energy substrate no other nation can match, building toward a civilizational trajectory that puts boots on Mars and reactors in lunar regolith inside the next decade.

The next time someone tells you America can't build things anymore, send them a map of where it's being built.

And then ask them what they're long.

This country is turning 250 years old. Two and a half centuries since a small group of farmers, lawyers, printers, and tradesmen in Philadelphia decided they were going to build something new in the world.

They built it.

Their descendants — the welders in Boca Chica, the electricians in Memphis, the pipefitters in Lebanon, the technicians in Phoenix, the roughnecks in the Permian, the kid placing components on a gigasat in Midland — are still building it.

And we're just getting started.

So when someone asks me what I'm long?

I'm long the kid in Midland… and the country he's building.

I'm long America.

Happy 250th, you magnificent country.