In 2024, Refactor co-led the Seed round into Hexium — a $12M round alongside MaC Venture Capital, with participation from Humba Ventures, Julian Capital, Overture VC, and R7 Partners. I was one of the first checks into the company.

The company stayed in stealth through most of that year. Then, in 2025, two things happened in quick succession. Hexium emerged publicly with TechCrunch coverage of its laser-based isotope separation platform. And shortly after, the company announced a strategic collaboration with Oklo and TerraPower — alongside Lawrence Livermore National Laboratory — to develop domestic production of High-Assay Low-Enriched Uranium (HALEU), the rate-limiting fuel for the next generation of advanced nuclear reactors.

Watching that arc — pre-seed conviction in 2024, public emergence in 2025, foundational nuclear partnerships within months — has been one of the more remarkable execution sequences I've seen in a decade of investing across hard tech.

The headline most early coverage led with — fusion power has a fuel problem, Hexium has a solution — is true, but it undersells what Charlie Jarrott and Jake Peterson are actually building. Fusion fuel is a wedge. Uranium enrichment is a wedge. The company is the atomic supply chain.

What Hexium Does

Hexium is reviving a piece of American physics that has been sitting in a classified vault for forty years. In the 1970s, Lawrence Livermore National Laboratory developed a technology called AVLIS — Atomic Vapor Laser Isotope Separation — a method for using precisely tuned lasers to pluck individual isotopes out of a stream of vaporized atoms with surgical precision. It was originally built for uranium enrichment. The U.S. spent roughly $2 billion getting it ready for commercial deployment.

Then the Cold War ended. The Megatons to Megawatts program flooded the market with cheap, downblended Soviet uranium — enough to power roughly 10% of U.S. electricity for two decades. The economic case for AVLIS evaporated overnight. The technology was boxed up. The physics stayed classified.

Hexium has brought it back. The mechanism is elegant: vaporize a target metal into a stream of atoms, shine three precisely tuned lasers — at "tattoo removal" energy levels but with picometer precision — at the exact frequency that ionizes only the desired isotope, and pull the ionized atoms onto a collection plate with an electric field. Everything else passes through untouched.

This is, importantly, a platform technology. The same architecture works on uranium. It works on lithium. It works, in principle, on any element with a wave-function-distinct isotope. What changes is the laser tuning. That's the whole company: a laser-based, software-controlled, modular system for isolating individual isotopes at industrial scale.

Uranium: The HALEU Bottleneck

Advanced nuclear reactors — small modular reactors, microreactors, the next generation of fission designs being deployed by Oklo, TerraPower, X-energy, and others — almost universally require HALEU, uranium enriched to between 5% and 20% U-235. The Department of Energy estimates the U.S. will need up to 40 metric tons of HALEU annually by the early 2030s to meet projected reactor demand.

The U.S. currently produces almost none of it commercially. The dominant enrichment infrastructure in the world today is centrifuge-based, and the dominant supplier is Russia. That dependency is no longer politically tenable.

This is where the Oklo/TerraPower collaboration matters. The three companies — joined by Lawrence Livermore — are jointly funding R&D to validate AVLIS-based HALEU production at commercial scale. AVLIS has structural advantages over centrifuge enrichment that make it especially well suited to this moment:

• No chemical conversion to uranium hexafluoride. This eliminates an entire toxic, infrastructure-heavy step from the front of the process.
• A fraction of the energy footprint. AVLIS originally promised to use roughly one-twentieth the power of gaseous diffusion. Modern lasers have only improved that math.
• Compact, modular form factor. Hexium's systems fit in buildings the size of a Starbucks. A traditional enrichment facility is the size of a stadium.
• Domestic, deployable, scalable. This is the part the DOE cares about most.

If AVLIS works at commercial scale — and the early benchmarking work with Oklo, TerraPower, and Lawrence Livermore is designed to validate exactly that — Hexium becomes a foundational supplier for the entire American advanced nuclear industry.

Lithium: The Fusion and MSR Problem

The same platform also unlocks a parallel set of markets in lithium isotope enrichment.

The fusion industry has a problem almost no one outside of it talks about. There are roughly 25 private fusion startups in the United States. Almost all of them rely on a deuterium-tritium reaction. Tritium has to be bred inside the reactor from lithium-6. And the United States has no domestic infrastructure for enriching lithium-6 at scale. The current stockpile dates from the Cold War. The only operating capacity in the world today belongs to China and Russia, both using a 1950s-era process called COLEX that leaches mercury into soil and waterways.

This is, by any reasonable definition, a national security problem. Every fusion company in America is presently making a quiet bet that the lithium-6 supply chain will materialize before they need it. Most of them are not building it themselves.

Hexium is.

In late 2025, the team announced a milestone worth pausing on: in a single pass, using their patent-pending approach to AVLIS, they enriched lithium-6 to 99.99% purity. That number is not a projection. It's not a model. It's measured output from their lab — the time-of-flight mass spectrometer data Charlie shared publicly shows the lithium-7 line essentially flat against the lithium-6 peak. Single-pass enrichment at that purity is an extraordinary result for a technology that was sitting in classified storage three years ago.

The second-order play is the byproduct. Lithium-7 — what comes out the other side of enriching lithium-6 — is critical for molten salt reactors (MSRs), an advanced class of fission reactors with billions of dollars of Department of Energy backing behind them. Companies like Kairos Power and Natura Resources are building MSRs now, with federal money, and the only suppliers of high-purity lithium-7 today are, again, China and Russia. The lithium-7 market alone is projected to reach $1–3 billion annually as MSR deployment scales.

One process. Two products. Both critical. Both undersupplied domestically.

Why I Wrote One of the First Checks in 2024

When I co-led the Seed in 2024, the company was still deep in stealth. The Oklo deal hadn't happened. TechCrunch hadn't covered them. The 99.99% lithium-6 result was still ahead. There was just Charlie, Jake, the physics, and the thesis.

A few of the things that got me to yes early:

The technical lineage was real. Charlie spent years inside Lawrence Livermore, where AVLIS was originally built. He knows the physics down to the laser frequency. The unrealized promise of AVLIS was sitting in classified archives, and the people who once worked on it have been quietly waiting for a credible team to bring it back. Charlie is reactivating his Q-clearance — the same security classification once held (and famously revoked) by Robert Oppenheimer — which may eventually unlock decades of accumulated lab knowledge for the company.

The market need was deterministic. This wasn't a thesis investment about whether some future technology might emerge. The DOE was already pouring money into HALEU and MSRs. Fusion companies were already raising at billion-dollar valuations on the assumption that the fuel supply would exist. The only question was who would build it.

The unit economics were almost too good. Natural lithium costs about $30 per kilogram. Enriched lithium-7 sells for roughly $10,000 per kilogram. Lithium-6 sells for at least 10x that. HALEU pricing is similarly favorable for a domestic producer with structural cost advantages. When the price arbitrage between input and output runs three to four orders of magnitude, you do not need to be early to the customer demand to build a real business.

The team understood the deeper opportunity. Lithium and uranium are the wedges. The long-term ambition Charlie and Jake describe is becoming the Dow Chemical of isotopes — the supplier of atomic building blocks for an entire generation of advanced industries. Silicon-28 for quantum computing. Copper-64 and copper-67 for nuclear medicine. Iron isotopes for radiation-hardened structural materials. Once you have a universal laser-based isotope separator running at industrial scale, almost every frontier technology in the deep tech stack — fusion, advanced fission, quantum, propulsion, advanced materials — touches it eventually.

That's a generational hard tech business. Not a fuel supplier.

A Note on Picks and Amps

There's a line Jake Peterson uses to describe Hexium that I have not stopped thinking about since I first heard it: in a world where everyone wants to be a rockstar, we are content making the picks and amps.

This is, I think, exactly the right framing for one of the most important kinds of company being built in critical industries right now. Fusion reactors and quantum computers and orbital launch vehicles and advanced manufacturing all get the magazine covers. But somebody has to build the atomic supply chain that makes any of it work. Somebody has to enrich the lithium and the uranium, separate the isotopes, supply the materials, manufacture the components.

These are the unsexy companies that compound for decades. The Dow Chemicals. The Linde Industrial Gases. The TSMCs. They are the foundation underneath the rockstars, and over a long enough timeline, they often become more durable and more valuable than the rockstars themselves.

Hard tech and frontier tech investing rewards founders who choose to build picks and amps when the cultural pressure is to build the rockstar product. Charlie and Jake chose correctly.

Why This Was a Refactor Investment

People sometimes ask me what makes something a Refactor deal. Hexium is the cleanest case study I have.

We invest at the intersection of computation and the physical sciences. We back hard tech, deep tech, and frontier tech across critical industries — aerospace, critical materials, energy, nuclear, defense, advanced manufacturing — where technical execution is the moat and the market need is structural rather than speculative. We lead or co-lead seed rounds with $1–2M checks for 5–10% ownership. We do this roughly 20 times per fund.

Hexium fit on every dimension. The science was proven. The market was urgent. The geopolitical case was airtight. The founders had the rare combination of physics depth, fusion-industry experience, and the temperament to build something deliberately unsexy.

Co-leading the round in 2024 let us be a foundational partner at the moment of company formation. That's exactly the seat I want every time.

What I'm Watching Next

A few signals worth tracking:

• The HALEU benchmarking work. The Oklo/TerraPower/LLNL evaluation will produce a validated technoeconomic assessment of AVLIS-based uranium enrichment. That's the document that unlocks the next phase of nuclear industry adoption.
• The pilot plant. Hexium is building toward a first commercial pilot for lithium isotope production, targeting tens to hundreds of kilograms of capacity.
• Fusion industry partnerships. Watch which fusion companies sign supply agreements first. Whoever locks in lithium-6 access ahead of the rest gets a structural advantage.
• MSR contracts. The lithium-7 revenue line is the near-term commercial story on the lithium side. Kairos and Natura are the early signals.
• Isotope expansion. The most interesting long-term question isn't lithium or uranium. It's whether Hexium successfully ports the platform to the next isotope, and the next.

A Note for Founders

If you're building at this kind of edge — hard tech, deep tech, frontier tech, in a critical industry where the science is real and the market is structural — I want to hear from you.

‍