The Mount Rushmore of New Nuclear

America’s superpower, rebooted

In today’s Diary:

• Heirs of the Manhattan Project

• “New nuclear” starts in coal country

• The uranium gummy bear

• Banana vs. nuclear plant

• 84 no’s. 1 yes.

Hey Rational Optimist,

Last month I rode a bus three hours into the Utah desert with a “who’s who” of the most American builders and innovators today.

We were headed to a launch party for a nuclear reactor.

I was in Orangeville, deep in coal country. This site was bare dirt nine months ago. Then Valar Atomics showed up and built what amounts to a small city.

I was there to see Valar’s microreactor, which went critical on June 18. During the launch party Valar ramped it up, made electricity, and used it to run an Nvidia AI chip live on stage. I even got to see the “Ward 250” reactor up close:

I first met Valar founder Isaiah Taylor in 2024 in his factory near El Segundo. We sat in a back room and talked about reviving nuclear energy, which was a far-fetched dream back then.

Longtime ROS members know we’ve been following the race to turn on America’s first microreactor. In May 2025 the White House demanded three brand-new reactors, built by private companies, go critical by America's 250th birthday. A big ask, considering America hadn’t switched on a new type of reactor in 50 years.

Did we hit the deadline? You bet. Four startups went critical by July 4!

I've spent the past two years visiting the factories where these reactors were built. I’m lucky to have met three of the four founders who hit the deadline.

What’s strange is we’re celebrating reactors going critical… yet America first achieved this 75 years ago. We invented this machine, built hundreds of them and then just… stopped.

This is the story of the rise, fall, and rebirth of America’s nuclear industry.

December 20, 1951: in a cinderblock building in the Idaho desert…

Sixteen men gathered around a machine the size of a van and switched on four lightbulbs.

These were the first watts in human history drawn from a splitting atom. The builders signed their work in chalk: "ELECTRICITY WAS FIRST GENERATED HERE FROM ATOMIC ENERGY."

Source: Idaho National Laboratory

This was the birth of America’s nuclear industry. It should have ushered in unbridled prosperity.

The chairman of the Atomic Energy Commission, Lewis Strauss, said our children would enjoy electricity "too cheap to meter." They would "travel effortlessly over the seas and under them."

Strauss knew more energy = more prosperity.

Robert Caro's biography of Lyndon Johnson shows what getting electricity meant to people in Texas Hill Country in the 1930s.

No electricity meant no pump. Every drop of water a farm family used was hauled from the well and carried on shoulder yokes by women. One told Caro: "Do you see how round-shouldered I am? My back got bent from hauling the water, and it got bent while I was still young."

Every hockey stick in human history is an energy unlock:

Fire let us cook. A lump of coal packed more heat than wood ever could and gave birth to machines that could do the work of a thousand men. Oil packed more energy still into a barrel. Out came the car, airplane, the whole roaring 20^th century.

Uranium is ridiculously dense with energy. One gummy bear-sized fuel pellet holds as much energy as a ton of coal or 149 gallons of oil. A golf-ball-sized lump of uranium would cover your entire life's energy use.

And it was cheap too. The plants America started in the late 1960s made electricity for about 3 cents a kilowatt-hour in today's money. At those prices your monthly energy bill would be roughly $27.

Nuclear is also the cleanest, safest energy source. You absorb more radiation eating a banana than living within 50 miles of a nuclear plant for a year.

America did what it did best back then: built lots of stuff, fast.

We went from one reactor in 1957 to 112 by 1990. At the peak of the buildout the world was adding about 20 gigawatts of clean nuclear energy every single year.

In 1966 the New York Times previewed the 21st century: a world without deserts, without smog, billions living "in unprecedented luxury thanks to nuclear power." The New York Times wrote that!

Doctors were implanting nuclear-powered pacemakers in human hearts. The atom was going to desalinate oceans, green deserts and power flying cars.

Then the villains arrived.

Bureaucrats.

Regulations pushed up the cost of atomic energy ninefold. Utilities cancelled more than 100 planned reactors. And for the next 30 years America didn’t greenlight a single new plant.

Valar founder Isaiah Taylor reminded me the first nuclear “regulator” was the Atomic Energy Commission (AEC). Its job was to build atomic energy. In 1974 we shut it down and replaced it with the Nuclear Regulatory Commission (NRC). Its job was to make nuclear power “safe.”

The NRC adopted a rule called ALARA (as low as reasonably achievable), which declared there is no safe dose of radiation. Sounds sensible. But if no dose is safe no amount of spending to reduce a dose is ever overkill.

By 1978 the NRC was issuing more than one new rule every working day. It included requirements like "nuclear-grade" steel valves, identical to those in any oil refinery, costing up to 50X more.

At the Sequoyah plant in Tennessee crews were ordered to tear out their own finished work and redo it to new specifications 23 separate times. The plant went from costing $300 million to $1.7 billion.

Multiply that across the whole industry, and this essentially made building new power plants in America illegal.

In its entire half-century of existence the NRC licensed two reactors that turned on. The old AEC licensed 133:

NRC requirements stretched the average nuclear build from roughly five years to 12. This, more than any single rule, was the death knell.

Think of a reactor as the world's largest mortgage. You borrow billions on day one, and you don't earn a cent until it’s finished… often decades later. Year after year interest compounds.

Congress's own budget office found a single month of delay added roughly $44 million to a plant build cost. A big nuclear project burns millions of dollars a day just standing still. Stretch build times to 12 years and the interest becomes prohibitive.

Time is money, especially in nuclear. It’s why nuclear is one of few technologies to get more expensive over time.

It’s why more US nuclear plants have closed this century than opened. The fleet peaked at 104 reactors in 2012 and slid to 94 today:

Effectively outlawing new nuclear builds was the single biggest mistake America made in the 20th century. I’ve caught flak for that view, but I stand by it.

The one attempt to build again proved the skeptics right. Vogtle, in Georgia, was meant to cost $14 billion. The final bill was $35 billion, and it arrived seven years late. Westinghouse went bankrupt in 2017 building it!

And the first wave of nuclear startups fared no better. NuScale spent half a billion dollars and two million engineering hours just to win a license… then costs killed its first project anyway. TerraPower was founded by Bill Gates in 2008, and 18 years passed before it was allowed to pour nuclear concrete.

When I started writing about and investing in nuclear back in 2018, the most optimistic thing you could say was: "We're probably not going to shut down as many reactors as we thought." That was the bull case! Nobody seriously believed America would ever build new reactors again.

I believe many of America’s problems today can be traced back to our failure to “go nuclear.”

My friend, the economist Tyler Cowen, coined the term “Great Stagnation” for what happened to America after the early 1970s.

That’s when energy use per person topped out and never recovered.

Author J. Storrs Hall went back through the sci-fi predictions of the 1950s and 60s… flying cars, moon bases, undersea cities… and noticed every prediction needing more than about 10 kilowatts of electricity failed. Nearly every “energy-lite” idea came true.

A line of code costs the same to run whether oil is $3 a barrel or $300. So for 70 years, America's brightest talent fled the expensive world of “atoms” for the cheap world of “bits.” We got Facebook instead of flying cars.

COVID was the wake-up call.

Think back to April 2020. The richest city (NYC) in the richest country (USA) in history lacked basic protective medical equipment. Doctors wore ponchos instead.

A country that spent decades assuming stuff just shows up looked around and said: “Holy cow, we don't make anything here anymore.”

That same month the price of oil went negative. For one surreal day sellers were paying people to take barrels off their hands!

Millions of us, cooped up at home, woke up and started asking what went wrong. Many of the smartest founders I know found the same answer: energy. The factories, flying cars and future we didn't get were downstream of the moment we stopped adding power.

We track 55 nuclear companies in the ROS startup database, which ROS patrons can access here. Thirty of those 55 were founded within the past five years. For the first time in my lifetime, there’s new blood in nuclear.

(If you’re not a patron and you’d like to become one, go here. I truly appreciate your support. ROS is 100% bootstrapped and not-for-profit, so your patronage goes a long way toward helping us take these primary research trips.)

I've been lucky enough to get to know three of the most ambitious “new nuclear” founders. All three have built startups now worth billions. I believe when the history of the second atomic age is written their faces will be on the Mount Rushmore of nuclear.

Isaiah Taylor's great-grandfather was a nuclear physicist on the Manhattan Project.

When Isaiah decided to follow in his footsteps, he pitched 85 venture capitalists on his nuclear startup. 84 said no. The one yes was all he needed.

On June 18, 16 days ahead of the president's deadline, Valar’s reactor went critical in Emery County, Utah.

One thing separates Isaiah from the many brilliant new nuclear engineers. He is laser-focused on making reactors rather than designing them. "To build nuclear you have to build nuclear," he told me… his way of saying the only path is to build many, watch what fails, and fix it in the next one.

When I quizzed him about the cost of all that physical trial and error, he said: "Steel is cheaper than software." Five software engineers cost you a million dollars a year, he explained. A million dollars buys a lot of steel.

Standing in front of Valar’s Ward 250 microreactor in Utah (named after his great-grandfather) Isaiah told me the first reactors were beautifully simple. They were uranium and graphite blocks and not much else. Isaiah is taking nuclear back to that era of simple, cheap, and fast.

The Ward reactor's shield, “the Citadel” is a 78-inch-thick fortress of pre-cast concrete blocks stacked by crane. No grout, bolts or fasteners. The reactor protection system was quoted at $5 million and two and a half years. Valar engineers built one from scratch in six weeks for about $400,000.

Valar went from an idea to a critical reactor in under three years. Isaiah Taylor will make the most driven person question their ambitions.

Idaho desert. June 4, 2026.

Antares’s Mark-0 microreactor went critical. It was the first privately developed reactor to reach criticality in America in more than 40 years!

I visited founder Jordan Bramble at its HQ, Antares Prime, in Los Angeles. The factory is HUGE, stretching seven football fields. "It's sized for the first 50 reactors," he said.

I like to geek out during tours, and Jordan didn’t disappoint. He explained Antares’s reactor uses rotating drums instead of control rods, freeing room to pack in more fuel. Rather than pumps it uses sealed metal tubes that shift heat with no moving parts. The engineering is so solid NASA has Antares working on heat pipes for its Moon program.

I think Antares has the best business plan in nuclear. Jordan ranked customers by willingness to pay. At the top sits the Pentagon, which will pay almost anything to power remote military bases.

Sell to premium customers until you're cheap enough for everybody. This is how technology evolves. The US military bought the first microchips and the first satellite navigation. No ICBM program, no Google Maps.

Antares has 16 reactors on its order backlog, including a contract with the Air Force to power a cyber-warfare center and the Space Force to put nuclear power in orbit! It also inked a deal with the Army to build a microreactor at Fort Bragg within two years.

Antares’s roadmap in Jordan’s words is: “Criticality in 2026, electricity in 2027, power the warfighter in 2028.”

In an industry where missing deadlines by a decade is the norm, that’s refreshing.

A reactor on the Moon over or under 2035? Jordan took the under. Maybe 2030, he said. And the limiting factor is whether SpaceX’s Starship is ready in time.

At the end I asked him my favorite question: How does America look different if you win? "Energy dominance," he said. "You make everything cheaper."

Antares has one of the best teams in the game. It doesn’t surprise me it won the criticality race.

Matt Loszak told his parents he wanted to start a nuclear company.

So they bought him a book about Chernobyl.

He was undeterred. After building a software startup and selling it for nine figures, he moved from Canada to Texas and founded Aalo Atomics.

I've visited Aalo's Austin plant four times. The first time, in late 2024, Matt had just moved in, and the enormous building held nothing but a giant American flag:

He told me: "We're going to mass-produce 50 reactors a year from this facility."

When I returned and saw Aalo's reactor I thought: that's a nuclear reactor? It looks almost too simple… which is the whole point.

Aalo’s reactor is designed around one question: what’s the most powerful reactor you can fit on the back of a normal truck?

The key is the coolant. Sodium doesn’t need to be pressurized. No pressure means you can delete the massive, forged vessels which take months to make.

The Aalo Pod is five of these reactors stitched together, purpose-built for AI data centers.

At Idaho National Labs, Aalo assembled its reactor in under four weeks. The 540 fuel rods were bundled into assemblies in two and a half days. It was the largest fuel load ever taken to criticality by a startup.

At 12:20 am on July 4, Matt stood in the control room in Idaho and watched his reactor go critical.

Aalo has since inked a deal with Crusoe, which built OpenAI's giant Stargate data-center campus. It’s targeting real electricity by next July.

Nuclear-powered AI… the future is here.

Honey, I shrunk the reactor.

"Nuclear" makes people picture giant cooling towers with steam drifting over a plant the size of a small town. Nuclear = big.

You know the trouble with big. Every plant becomes a bespoke, decade-long megaproject. Nuclear plants are some of the most expensive buildings we’ve ever built.

But remember: atomic energy started small… the first reactor only lit four lightbulbs.

Blame the Navy for nuclear’s “size” troubles.

The Navy needed a reactor to fit inside a submarine. The only way to shrink one back then was to switch to water cooling at enormous pressure. It worked brilliantly… for submarines. Then the civilian industry borrowed the idea and blew it up to the size of a city block.

Isaiah, Jordan, Matt and the rest of the new nuclear crew are making nuclear small again. Microreactors are small enough to be built in a factory, shipped on trucks, installed in months, and constantly improved with every new version.

"How do you make things cheap?" Isaiah said. "You iterate." For 50 years reactors were built like NASA built spaceships. This new wave of nuclear startups is more like SpaceX.

This completely shifts who builds nuclear and who buys it.

Build: The last two big reactors America finished cost roughly $35 billion. Aalo reached criticality having spent less than $100 million.

Remember, much of the cost of big plants was interest on money spent waiting. Ten small reactors coming online one by one start earning revenue years before a giant plant gets going.

Buy: The only customers nuclear ever had were giant regulated utilities. A factory-built reactor is something data centers, military bases and even towns can buy.

Matt told me data-center demand was "off the charts… they need lots of power, and they need it yesterday."

All four companies that beat the July 4 deadline are less than 5 years old.

You've met Antares, Valar, and Aalo. Fourth was Deployable Energy, which I’m ashamed to say I didn’t even know about until it went critical.

Houston-based Deployable was founded in 2025 by oilfield veterans. It built a one-megawatt "nuclear battery" inside a 20-foot shipping container. It went from project kickoff to a critical reactor in about 150 days. Whoa!

Then, on August 5, Oklo's Groves facility went critical. I met Oklo's CEO Jake DeWitte on July 4 in Proto-Town, the capitalism commune outside Austin.

This is the Great Handoff. For most of the 20^th century megaprojects were the exclusive domain of governments.

Now spaceships, supersonic jets and nuclear reactors are being built by founders in their 20s and 30s.

In 1973 President Nixon unveiled Project Independence. His plan was to build 1,000 nuclear reactors by the year 2000. We built 112.

I believe this new breed of nuclear founders will make good on Project Independence and usher in the era of American energy dominance. What happened this summer is only the start.

Walking Valar’s site in Utah, lips cracking in the desert heat, I kept thinking about how we stopped celebrating engineering feats. When the Brooklyn Bridge opened in 1883 New York fired 14 tons of fireworks at once.

Valar’s launch in Utah was, as far as I can tell, America's first infrastructure party in half a century. It won’t be the last.

Here’s how you can take part in the nuclear renaissance.

Help build it. Nuclear startups are hiring welders, electricians, machinists, coders, and, in Isaiah's words, people with extreme agency who want to build hard things. In years to come people will brag they were on those factory floors the way people claim they were at Woodstock.

“Yes in my backyard.” If a reactor ever comes to your county, say yes. One American town saying yes can do more for the atom than decades of studies.

Tell a better story. Nuclear has always had the facts on its side. It lost because the other side told a better story. One about Homer Simpson asleep at the control panel while green goo leaks into the Springfield River.

You now know better stories. Share them by sending this Diary to your friends and family.

Here's to the rational optimists building the future!

—Stephen McBride

PS: We’ve spent 50 years treating a crucial energy supply as trash.

America spent decades and billions of dollars trying to figure out where to bury its spent nuclear fuel. I just met the founder who wants to dig it all back up.

Stafford Sheehan says the 95,000 tons of nuclear “waste” sitting in vats across the country is actually the biggest untapped energy reserve in the US. That’s because spent fuel still contains 95% of its original energy.

Project Omega is Stafford’s way of turning that wasted energy back into fuel, medical isotopes, and batteries that can run for 30 years:

https://www.youtube.com/watch?v=tMc-m1VcETM