A couple of years ago I sat in a packed conference room listening to a startup founder explain why his “walk-away safe” micro-reactor was going to make investors rich. The slides were beautiful, the vision even better. Fast-forward to today and that company is basically dead. The story repeats itself across the nuclear space more often than most people realize.
I’ve been following the nuclear sector closely ever since, and one part fascination with the technology, one part morbid curiosity about the money. What I’ve learned is simple: the physics is the easy part. The hard part is figuring out which teams can actually turn drawings into gigawatts without burning through billions first.
So consider this your cheat-sheet – the guide I wish I’d had before I started writing checks (or almost writing checks) in this space.
Why Nuclear Suddenly Feels Unstoppable Again
Let’s start with the macro picture, because none of the company-specific noise matters if the tide isn’t rising.
Data centers, electrification, and the sheer embarrassment of watching coal plants get emergency extensions have created a perfect storm. The IEA now says global electricity demand will grow 80% by 2050 in its net-zero scenario. Renewables alone can’t keep up – especially when the sun isn’t shining and the wind isn’t blowing.
Even Big Tech has noticed. Microsoft signed a deal to restart Three Mile Island Unit 1. Amazon bought a data-center campus next to a nuclear plant. Google and Oracle are both in talks for SMR power. When the companies that literally run the internet decide nuclear is the answer, you should probably listen.
The Four Families of Reactors You’ll Actually Hear About
Strip away the marketing and almost every company falls into one of four buckets. Understanding the differences is the difference between investing and gambling.
1. Light Water Reactors (LWR) – The Old Reliable
Think of these as the Toyota Camry of nuclear – boring, but they just work. About 90% of operating reactors worldwide are LWRs (PWR or BWR) using ordinary water to both cool and moderate the reaction, with uranium-oxide pellets in zirconium tubes.
Pros: insane operational history (thousands of reactor-years), regulators know them inside out, fuel supply chain is mature.
Cons: big, expensive to build, run at relatively low temperature (~300°C), which limits efficiency and industrial-heat applications.
Who’s playing here: NuScale (already has NRC design certification for its 77 MWe module), GE-Hitachi, Westinghouse (AP1000), Holtec, Rolls-Royce (UK SMR).
My take? If you want the highest probability of something actually getting built in the 2020s, bet on an LWR-based SMR. The technology risk is basically zero. The execution risk is still massive, but at least the physics won’t surprise you.
2. Sodium-Cooled Fast Reactors (SFR) – The Breeder Dream
No moderator, liquid sodium coolant, often metal fuel. These can theoretically “breed” more fuel than they consume and burn nuclear waste. Russia’s BN-800 has been doing it for years.
Pros: high temperature (~550°C), excellent for industrial heat, potential fuel-cycle closure.
Cons: sodium explodes on contact with air or water (ask anyone who worked on EBR-II about the 1986 leak), most designs need HALEU (fuel enriched 19.75%), supply chain basically doesn’t exist at scale yet.
Key players: TerraPower (Natrium, backed by Bill Gates), Oklo (Aurora), Aalo Atomics.
“Sodium is a marvelous coolant… until it isn’t.”
– Veteran fast-reactor engineer I had coffee with last year
3. Molten Salt Reactors (MSR) – The Exotic Promise
Liquid fuel dissolved in molten fluoride or chloride salt. Some designs drain the core into a tank if things go wrong – true walk-away safety in theory.
Pros: very high temperature (600-700°C+), low pressure, online refueling possible, can burn thorium or waste.
Cons: almost zero operating experience at scale, corrosion is a nightmare, tritium management, salt freezes at ~400°C so you can never let it cool completely.
Companies: Kairos Power (fluoride salt, building demo in Tennessee), Terrestrial Energy, Moltex, ThorCon, Natura Resources, Copenhagen Atomics – the list is long because the barrier to a nice PowerPoint is low.
Reality check: Oak Ridge ran a molten-salt reactor for 20,000 hours in the 1960s. Zero commercial plants since. That should tell you something.
4. High-Temperature Gas-Cooled Reactors (HTGR) – The Quiet Contender
Helium coolant, graphite moderator, TRISO fuel particles that can withstand 1600°C+. The fuel is basically indestructible – they literally tried to destroy it in tests and failed.
Pros: highest safety margin of any design, excellent for process heat (hydrogen, desalination, synthetic fuels), demonstrated in real plants (though old).
Cons: huge vessel size because gas has low heat capacity, TRISO fuel is expensive, still needs HALEU in most designs.
Players: X-energy (Xe-100), BWXT, Kairos (different from their salt design), Ultra Safe Nuclear (USNC), Radiant Nuclear (Kaleidos micro).
The Fuel Question Nobody Wants to Talk About
Almost every “advanced” reactor wants HALEU – uranium enriched to 19.75%. Right now the only commercial supplier in the western world is… Russia. Centrus Energy is building capacity in Ohio, but we’re talking a few tonnes a year when the demand pipeline is hundreds of tonnes.
If your favorite startup says “first core in 2028” and they’re not LWR-based, mentally add 3-5 years until domestic HALEU exists at scale. That single bottleneck has already delayed multiple programs.
What Actually Kills Nuclear Projects (Hint: Not Physics)
- First-of-a-kind syndrome – The first AP1000s at Vogtle cost ~$15 billion each instead of $3 billion. Even with everything “proven,” costs ballooned.
- Regulatory learning curve – Oklo’s Aurora was rejected by the NRC because staff literally didn’t understand fast reactors.
- Capacity factor surprises – Large LWRs hit 93% today, but it took decades. New designs often start in the teens.
- Supply-chain infancy – Forgemasters that can make 600-ton reactor vessels are booked years out.
My Personal Investing Framework
After watching a dozen nuclear SPACs implode, here’s what I look for now:
- Do they have a license or at least a realistic path? (NuScale yes, almost everyone else no)
- Do they have a signed customer with skin in the game? (Dow Chemical for X-energy, Google rumors don’t count)
- Is the fuel available today or in the next 36 months?
- Do they have an operating test reactor or hard commitment to one? (Kairos at ORNL, TerraPower decisions this year)
- Can management build hardware or just PowerPoints? (Look for ex-Navy, ex-NuScale, ex-large EPC experience)
If the answer is “no” to two or more of those, I pass – no matter how cool the coolant is.
Where I’m Putting My Money in 2025
Full disclosure – I own shares in uranium miners, Centrus (HALEU), BWXT (components + HTGR), and a tiny position in NuScale because they’re the only ones with actual NRC approval.
I’m watching X-energy and Kairos Power closely – both have serious engineering cultures and real customer traction. Everything else feels like lottery tickets.
The nuclear renaissance is real. But most of today’s darlings will be tomorrow’s cautionary tales. The winners will be the ones who under-promise, over-engineer, and remember that in this industry, time is measured in decades, not TikTok trends.
Invest accordingly.
