The SMR Startup Field Guide
Making Sense of the Companies Racing to Criticality by America's 250th
The SMR Startup Field Guide
☢️Hi,
A couple weeks ago I shared a primer on small modular reactors (SMRs) and this week I’m excited to come back and dig a little deeper into the companies trying to make SMRs happen.
Last August, the DOE selected 11 projects from 10 companies for its Nuclear Reactor Pilot Program, with the ambitious goal of bringing at least 3 test reactors online by July 4, 2026, less than 6 months away.
So what are these companies actually building? And why are there so many different approaches? When trying to understand the array of different companies and approaches, I didn’t find any great centralized source to learn about them, so I thought it could be helpful to create one myself!
This post is the Uncredentialed field guide to America’s nuclear startup race.
Before we get started, though, welcome to 16 new Uncredentialed members! If you’re new here and interested in seeing more posts like this one, subscribe today! If those first 5 sentences weren’t enough to sell you, that’s understandable too, but give it a read and let me know what you think!
Pick Your Coolant: Three Bets on Nuclear’s Future
Before diving into the companies, it helps to understand that they’re not all building the same thing. The 11 projects in the DOE pilot program fall into 3 distinct categories, each representing a different bet on what the future of nuclear looks like.
Bet #1: Shrink What Works (Light Water Reactors)
Pressurized water reactors (PWRs) power over 300 reactors worldwide. They’re the Honda Civic of nuclear, they’re proven, reliable, and well-understood by regulators. The bet here is that you can take this mature technology, shrink it down, and mass-produce it.
The downside is that it still inherits many of the downsides of traditional nuclear like lower operating temperatures and too much established regulation (notably it replaces the expensive, bespoke nature of traditional nuclear with mass production, though).
Bet #2: Salt of the Earth (Sodium and Molten Salt Reactors)
What if you used something other than water to cool your reactor? Sodium and molten salt can operate at much higher temperatures, making them more efficient, with some designs even running hot enough to burn recycled nuclear fuel.
The main tradeoff here is regulatory uncertainty. It’s a new and unproven technology and regulators are still figuring out how to evaluate it, so there’s an inherent risk of not even knowing what regulatory hurdles you might need to clear.
Bet #3: Too Cool to Melt (High Temperature Gas Reactors)
These reactors use TRISO fuel, tiny uranium particles coated in layers of ceramic that can withstand extreme heat. Combined with a helium coolant, they can operate at temperatures exceeding 900 degrees Celsius and are physically unable to melt down.
While the technology has been tested, the tradeoff with these is that TRISO fuel is more expensive and the tech hasn’t been implemented before at commercial scale, leaving some execution risk/uncertainty.
Shrink What Works: Light Water Reactors
Three projects in the pilot program are using pressurized water reactors, the same basic technology that’s powered commercial nuclear for decades.
Deep Fission: A Reactor at the Bottom of an Oil Well
Deep Fission’s Gravity Reactor is a 15 MW PWR designed to sit at the bottom of a mile-deep borehole. At that depth, the weight of the water column above provides the same pressure that traditional PWRs need expensive containment structures to create.The company claims this could cut construction costs by 80%.
CEO Liz Muller came to the idea through her previous company, Deep Isolation, which developed borehole disposal for nuclear waste. The insight was that the conditions you want for storing spent fuel aren’t all that different from the conditions you want for running a reactor.
Deep Fission broke ground in Parsons, Kansas in December and is targeting criticality by July 4th. They’ve also announced sites in Texas and Utah, with plans to cluster multiple reactors together to scale up to 1.5 GW per site.
Lastly, they plan to use standard fuel enriched below 5%, avoiding supply chain issues with harder-to-get HALEU fuel.
Last Energy: The Full-Service Model
Last Energy is looking to offer NaaS (nuclear-as-a-service), building, financing, and operating 20 MW PWR plants on customer sites, handling everything from construction to decommissioning.
Their PWR-20 (similarly to Deep Fission) uses standard fuel to avoid supply chain issues but also utilizes air cooling to expel waste heat, instead of the traditional river/ocean used by other plants, allowing it to be sited almost anywhere1. The company has already built two full-scale prototypes in Texas.
For the pilot program, Last Energy is testing a scaled-down 5 MW version at Texas A&M’s RELLIS campus, with testing expected to begin this summer. Beyond the pilot, they’ve announced plans for 30 reactors in Haskell County, Texas to serve the data center market.
Lastly, they also have projects advancing in the UK, where they’re targeting site license approval by late 2027.
Atomic Alchemy (Oklo): The Isotope Play
Not every reactor in the program is about generating electricity. Atomic Alchemy, acquired by Oklo in early 2025, is building a 15 MW light water reactor specifically designed to produce radioisotopes.
Their VIPR (Versatile Isotope Production Reactor) can produce over 40 different isotopes used in medicine, research, defense, and semiconductor manufacturing. The medical isotope market alone is expected to grow significantly as new cancer treatments using targeted radiotherapy gain traction, and that’s before even touching on semiconductors.
Atomic Alchemy recently withdrew from the NRC’s application process in favor of pursuing the faster authorization enabled by the DOE’s pathway, a sign of how the pilot program is changing companies’ regulatory strategies.
Salt of the Earth: Sodium & Molten Salt Reactors
5 projects are betting that moving beyond water cooling unlocks major advantages in efficiency and fuel flexibility.
Oklo: The Veteran
Oklo is the elder statesman of the nuclear startup world, being founded back in 2013, a full decade before many of its pilot program peers.
Their Aurora Powerhouse is a 75 MW sodium-cooled fast reactor based on the Experimental Breeder Reactor-II, which ran at Idaho National Laboratory (INL) from 1964 to 1994. Fast reactors can burn recycled nuclear fuel and produce less long-lived waste than traditional reactors.
Oklo has had a rocky regulatory journey. Their first NRC license was denied in 2022 for insufficient information but they’ve since pivoted to the DOE pathway, broke ground at INL in September, and received approval for their fuel fabrication facility in November.
The company has multiple projects in the pilot program (also including Atomic Alchemy and a 3rd project called “Pluto” that doesn’t have much public info yet), more than any other participant. Meta recently signed deals that could see Oklo deploying reactors to power their data centers.
Aalo Atomics: Scaling the MARVEL Design
Aalo is pursuing an “extra-modular” approach. Each Aalo-1 reactor produces 10 MW and 5 of them combine into an Aalo Pod producing 50 MW, designed specifically for behind-the-meter data center power. The reactors are sodium-cooled, factory-built in Austin, and use standard low-enriched uranium fuel.
Aalo was the first company to break ground in the pilot program, starting construction at INL just 16 days after selections were announced. Their CTO, Yasir Arafat, previously led the DOE’s MARVEL microreactor project at INL, giving them deep institutional knowledge of the lab.
In July, the company became the first US advanced reactor company to sign a commercial contract for enriched uranium delivery. If they hit their July 4th target, Aalo-X will be the first new sodium-cooled reactor to operate in the US in over 4 decades.
Antares Nuclear: The Nuclear Appliance
At just 500 kW, Antares is one of the most compact designs in the pilot program, but to be clear, being small is the point.
Their R1 microreactor uses sodium heat pipes for cooling, a completely passive system with no pumps. It’s designed to fit in a shipping container and operate for 3 years without refueling, targeting military bases, remote industrial sites, and, eventually, space applications.
Antares was the first company in the pilot program to receive an approved Nuclear Safety Design Agreement from the DOE, and the first to begin TRISO fuel fabrication from their HALEU allocation.
Natura Resources: The First Liquid Fuel Reactor Licensed in US History
In September 2024, Natura became the first company in US history to ever receive NRC approval to build a liquid-fueled reactor.
Their MSR-1 is a 1 MW molten salt research reactor being built at Abilene Christian University. In a molten salt reactor, the fuel is dissolved directly into the salt coolant, allowing for continuous refueling and the ability to use recycled fuel from other reactors.
Natura’s pursuing both the NRC pathway (for commercial applications) and the DOE pilot program, and plan to scale up to 100 MW commercial systems for grid power and produced water desalination in the Permian Basin.
Terrestrial Energy: The Biggest Reactor in the Program
While most pilot program participants are building microreactors, Terrestrial Energy is going big. Their Integral Molten Salt Reactor (IMSR) produces 195 MW of electricity.
Terrestrial has pursued a route that gives them a similar competitive advantage to Last Energy and Deep Fission in that they use standard low-enriched uranium fuel (under 5% enrichment), avoiding HALEU supply constraints that most other advanced reactors face. The entire reactor core is designed as a replaceable unit with a 7-year lifespan, sidestepping the corrosion issues that plagued earlier molten salt designs.
Terrestrial went public by SPAC in MArch 2025 and is planning commercial deployment at Texas A&M’s RELLIS campus in the early 2030s. They’re also the furthest along in traditional regulatory approval, having completed Canada’s Vendor Design Review and received NRC approval for their Principal Design Criteria.
Too Cool to Melt: High Temperature Gas Reactors
Lastly, there’s 2 companies betting on combining TRISO fuel and helium cooling to boost safety by creating a reactor that can’t meltdown.
Radiant Industries: SpaceX for Nuclear
Radiant was founded by former SpaceX engineers who wanted to bring aerospace-style iteration to nuclear. Their Kaleidos microreactor produces 1 MW of electricity and is designed to replace diesel generators anywhere they’re used, whether it be military bases, data centers, disaster relief, or more.
The reactor fits in a shipping container, and can be set up within 48 hours of delivery. Radiant is one of the key disciples of “productizing” nuclear, aiming to scale their factory up to producing 50+ reactors per year.
They’ve signed contracts with the US Air Force and Equinix (a data center operator), among others and will be the first reactor tested at INL’s new DOME facility in 2026.
Valar Atomics: First to Criticality
In November, Valar became the first company in the pilot program to achieve criticality, reaching zero-power criticality at Los Alamos’ Nevada facility.
Their Ward 250 is a 100 kW helium-cooled reactor using TRISO fuel, with a vision of building towards 1000-reactor gigasites where they’ll produce not just energy, but also hydrogen and synthetic fuels (like those used by planes) by using the reactor’s heat.
The company broke ground in Utah in September but is also pursuing deployment in the Philippines to accelerate their development with a lower regulatory hurdle.
What to Watch For
The July 4th deadline is less than 6 months away. Here’s what I’m watching:
Who hits criticality first? Valar already hit cold criticality, but that’s more of a check box from my perspective. Hot criticality, where the reactor actually produces usable heat, is the most important milestone to watch out for. Aalo and Antares are both pushing hard here but it’s early, and there always could be multiple winners.
Which approach scales? PWRs (Bet #1) have a regulatory edge. Sodium and molten salt reactors (Bet #2) have an efficiency edge. HTGRs (Bet #3) have a safety edge. The market will ultimately decide which tradeoffs matter most.
The fuel question. Several companies (Last Energy, Terrestrial, Deep Fission) deliberately chose designs that use standard low-enriched uranium, avoiding dependence on scarce HALEU fuel. Others are counting on DOE allocations and new enrichment capacity coming online. Fuel availability could play a factor in which companies can actually deploy at scale.
Uncredentialed Growth
As promised, I want to continue to share some of the numbers I’m looking at as Uncredentialed grows. It feels kind of nuts, but at this pace we’re on track to halve our days to each 100 down to about a month!
Really grateful that people seem to be resonating with what I’m writing and all the support. Not much beats checking your phone and seeing a whole bunch of people you respect are sharing and reposting the words you strung together on the internet!
If you enjoyed this post, consider sending it to a friend who’d like it too! And if this is your first time here, today’s the day! Subscribe below to join the Uncredentialed family!
Note: it still uses water for its primary cooling loop on the nuclear side, that heat is just then transferred to air that can be expelled into the atmosphere