Valar Atomics Q&A
Valar Atomics hopes to make energy 10 times cheaper by using nuclear fission to pull oil and gas out of thin air
Deep inside the sunny bastion of American optimism known as El Segundo there is a start-up company called Valar Atomics, which plans to use nuclear fission to extract hydrogen and carbon from the atmosphere.
The company’s unique market strategy directly addresses the considerable challenges of nuclear reactor approval and construction.
Traditional approaches to nuclear development projects often involve gigawatt-scale projects — i.e., building multiple large-scale reactors to amortize regulatory and environmental costs, which can lead to billions in cost overruns and decades-long timelines.
In contrast, Valar proposes to develop numerous small, manufacturable reactors, bunched together as “gigasites,” which are co-located with other industrial entities e.g., data centers and Bitcoin mining facilities. This strategy aims to achieve economies of scale and potentially reduce costs and deployment times.
Since the existing grid has numerous limitations in its ability to accommodate nuclear capacity, Valar plans to diversify its suite of products. In addition to supplying electricity, the company will produce synthetic fuels (such as hydrocarbons and hydrogen) which are more easily transportable and deliverable to a broad market. This approach involves constructing, operating, and maintaining Valar’s reactors and selling associated bi-products to end-users directly.
Currently, Valar is engaged with governments in West Virginia, Wyoming, Utah, and Nevada to explore potential pilot-scale (or large-scale) generation sites. West Virginia is of particular interest due to its substantial CO₂ emissions from coal plants, which could be utilized in Valar's carbon recycling processes. Valar is also in preliminary discussions with some large utilities to explore partnerships that would transform carbon emissions from coal plants into marketable products.
The company’s founder, Isaiah Taylor, has the pedigree to pull off the impossible — his great grandfather worked as a physicist on the Manhattan Project.
Taylor and his team have landed on the striking notion that, despite the fact that nuclear seems perfect for grid power due to its baseload capabilities and capacity factor, its full scaling potential is unlikely to be realized with the grid as its anchor customer. Rather, the grid must become one customer among many, with gigasites producing multiple products that are not tied to local grid needs, e.g., hydrogen, data-center power, heavy industrial power, and synthetic hydrocarbon fuels.
I recently talked with Valar’s head of operations, Kip Mock, to learn more. Below is a transcript. [Minor readability edits and hyperlinks were added for clarity.]
UPDATE: Last evening, the EOS Organization announced a grant of $192,000,000 Philippine pesos (approximately $3.3 million U.S. dollars) to the Philippine Nuclear Science Foundation for the advancement of nuclear research. This work entails the development of a Generation IV Research Reactor by 2027. EOS is engaged with Valar Atomics to achieve this goal, as facilitated by Administratum, who we wrote about on November 11. Remarks from U.S. Ambassador to the Philippines MaryKay Carlson are available here.
Q: Do you see any weak links in the supply chain or production chain that could make it difficult to manufacture some of your products?
A: There are a couple. Our main philosophy is: where there's a weak link, it’s there because it hasn't been scaled out yet. An obvious one is nuclear fuel. Uranium, the type of fuel we're using, is way under-produced right now, and even the U.S. government is having a hard time getting ahold of the amount they need for allocations for different research projects. The DOE is a few kilograms short — just a few kilograms — and it's causing a big headache for them. That's a big problem.
Graphite — nuclear-grade graphite — is about 300 times more expensive than it ought to be because it's under-supplied.
Helium is a concern, but we just tapped the largest helium reserve here in the U.S. a few months ago. So we think that problem is well on its way to being solved, but the fuel and the graphite problems are pretty significant in the short-term. That just means our costs for our pilots and our prototypes are higher than they should be. In the long-term, we hope to vertically integrate those processes in and amortize those costs out.
Q: Has Valar interacted yet with the Nuclear Regulatory Commission or other regulatory bodies? What kind of challenges do Small Modular Reactors face with these regulatory bodies?
A: We haven't engaged much with the NRC yet. We're hoping to zero-in on a site [and] have a real plan before we just talk with them about nothing. We've engaged with the DOE a little bit and then with with some foreign regulators that are allied to the U.S.
There are a number of problems with the regulatory regime as it stands right now, and it seems like it's become a bipartisan issue. The Advance Act1 was just passed (very bipartisan) and there was another bill passed in 2019 to refine the NRCS process for advanced reactor deployment. So we think that the problems are being solved.
To put it simply, I could talk for hours about challenges with the regulatory environment. The main problem is that the NRC was built around large Light-Water Reactors. That's what it was built for. And so now that there are dozens of companies that are developing non-Light-Water Reactor tech that's inherently safer, inherently cheaper and more reliable…the tide is shifting, and the NRC is is catching up to that. It's taking a long time, as government tends to do, so there is no clear path for Gen IV Reactors like ours to get to market in a timely fashion. It takes a decade to license a gigawatt reactor. It shouldn't take a decade to license a megawatt reactor. This seems to make sense just from the risk calculus, but there's no sliding scale yet. Hopefully there will be soon.
Q: Valar anticipates using its DOLE Loop2 (“Deploy-Operate-Learn-Educate”) as a blueprint for continuously reducing the cost of nuclear — this process suggests a self-contained, experimental laboratory. What kind of innovations would you pursue in that scenario?
A: This is another question of regulatory challenge. The technology only gets cheaper through iteration. Obviously, with nuclear you can't do destructive testing like SpaceX does with rockets, so you still need extremely rigorous safety margins.
But every other industry in the U.S. and around the world has a kind of flight-test program. The FAA has a certification process for aircraft if they're going to be flying people across the Pacific, and then they have a certification process for getting the prototype in the air. They have certain stress-tests for the Triple 7X that apply when it's in prototype phase, and then certain stress-tests that apply when it's in commercialization phase, and that doesn't exist for nuclear.
What we would love to see — this is getting way off into the policy weeds — but there are a lot of very small reactor prototypes that are in design right now (or even moving towards construction) that have to abide by those commercial standards in order to turn on. You're talking about tens — hundreds of millions of CapEx — just in regulatory overhead for something that's not going to produce any revenue, and this just doesn't make economic sense.
For nuclear, what needs to be done is: you take the maximal negative outcome, and you draw a circle on a map, and maybe there is some location — somewhere in the world, somewhere in the U.S. — where you can allow that reactor to go into testing without the commercial level of oversight and approvals.
That's what we are trying to get at with this idea of a regulatory sandbox, I guess you could call it — it's effectively giving the NRC a flight-test program for nuclear reactors to allow innovation to happen.
Q: I like that approach — it’s a clever way to harmonize with regulators. What kind of data would you share with them?
A: Well, the nice thing about this concept is it really simplifies everything, because you have what's called a source term for for nuclear reactors. We have been doing this math for 90 years now, and you can simulate the worst possible event and what radio isotopes would be released; what level of damage could occur. And you can draw a very clearly defined line on the map and say “nothing bad will ever happen outside this line no matter what happens.”
It's a really simple case to make. Now that being said…as a nuclear company, you don't want your reactor to blow up. That's going to be bad PR. So you're still going to design the reactor to be inherently, passively safe. But this allows you to make a simpler case for iterative testing, rather than going through to the Nth degree on every single nut, every single bolt, every single pipe — to tolerances that are not expected — even in commercial aviation.
Q: I’m very curious about the company culture at Valar — I read this Vanity Fair piece that features Valar and I was struck by El Segundo being a hotbed of American optimism, with American flags, guys lifting weights and bringing their bibles to work. It’s alien to someone like me in Portland, Oregon who loves all that stuff but is surrounded by bleakness and despair, particularly the last four years.
A: They represented the the vibes here and the culture pretty well. It's light-hearted. We believe that America was built by not that many people, and there's a lot to America that needs to be rebuilt, and we're confident that it's going to happen. We might be crazy, but it's it's more fun to live life thinking it's going to be OK than thinking everything's going to be terrible. If you think that, then it will be.
Q: Can you discuss the staff at Valar — it looks like you’ve assembled some top-tier talent, with people from Argonne Labs and some nuclear-engineering adjunct professors...
A: Our first really, really important hire was bringing on the former president of USNC-Power, Mark Mitchell. That guy is the best advanced reactor designer in the world. We firmly believe that. He and his team designed the Pebble Bed Modular Reactor back in the late ‘90s. They're a team of South Africans, and they are incredible nuclear and mechanical engineers.
But South Africa had some political turmoil. The reactor never got turned on into nuclear capacity in South Africa. Just a few months ago, China turned on the Pebble Bed Modular Reactor, renamed the HTR-PM, with a few slight modifications, but basically what our team designed.
And they did one of one of the most momentous tests that's been done on this type of reactor. They turned off the coolant flow and just let it sit there, and it proved that this type of reactor design — high-temp gas reactor, correctly designed — is completely, passively safe. So that was a great proof-point for us — proving that our team knows how to design safe reactors.
[But] our design is different from the Pebble Bed. It's a prismatic core, rather than a pebble bed. This allows us to have a far more deterministic safety analysis. A pebble-bed core is constantly changing geometry, and so while the HTR-PM is incredibly safe, we think ours is going to be even more safe and more replicable in terms of its safety performance. That's why we made that one design pivot.
But yeah, our engineering team is absolutely outstanding. The vast majority of them come from Mark Mitchell and the companies he's worked for — PBMR and USNC. Some of them from from other companies as well. X-Energy, BWXT, etc. Our neutron guys — we’ve had some from Argonne National Labs, some from INL. And all of them are absolutely outstanding.
Q: Here’s an innovation question from my co-author — are fission engines conceptually possible — basically anything beyond submarines that go into cars, motorcycles, blimps, rockets, etc?
A: The smaller you get, the more difficult it is for nuclear fission to occur. You have what's called a critical mass. You have to have a certain amount of fissionable material, and that mass is actually pretty large. I would love to be wrong about this, but I don't think it's ever going to be something that goes in [for instance] a passenger vehicle.
You have the weight of the fissionable material, the weight of the reactor and the weight of the shielding — it’s just too heavy, so I don't think it'll ever be practical for that application.
There's also the consideration that nuclear-powered submarines and aircraft carriers, they’re military vehicles with extremely highly enriched uranium, well over the threshold that's considered weapons-grade at about 80%. The idea of that getting into anything other than military vehicles is probably pretty unlikely.
That being said, things like blimps could be an application for sure. Basically, the larger the vehicle, the more practical it could be for a lower enriched uranium reactor to operate and power that vehicle. So cargo ships are an obvious application, particularly for low enriched uranium, say five to 10%.
The Advance Act directs the NRC to develop guidance to license and regulate microreactor designs within 18 months and reduces certain NRC licensing fees to facilitate administrative efficiency.
Where “Deploy” entails working with regulators to continuously deploy known reactors over and over on a gigasite; “Operate” involves selling heat, hydrogen, and electricity at good margins to co-located industry and distributors; “Learn” entails using a gigasite’s paid-down infrastructure as a private lab to continuously innovate on reactor designs, in proximity with regulators; and “Educate” means using experimental test data to educate regulators and formalize the results into accepted design standards; and