Build1 distinct publisher3 min readPublished
Actinide's accredited assay is real and very small. The claim worth testing is that a direct-to-solid separator removes the deconversion step America never built.
The Engineer · Build desk

Compiled by The EngineerSomething wrong?How this is made
The accreditation covers the ruler, not the run. Actinide says an ISO/IEC 17025-accredited laboratory measured the sample independently [3], which is worth having, because it means the 15.38% figure was not self-graded. It certifies nothing about how the material reached that assay, and the company's separate claim to be the first startup to enrich natural uranium into HALEU has not been independently established [4]. The report also carries no mass for the batch [19]. Without kilograms, the result cannot be placed on the same axis as anything else in the fuel chain, and that axis is where the whole argument lives.
Against what the NRC actually licenses, the assay is high. Louisiana Energy Services is listed as licensed to enrich to 5.5% [12], and 15.38% is roughly 2.8 times that level [1]. The licence is the more instructive half of the comparison. LES holds one for production; Actinide's batch exists under a laboratory-scale exclusion in the federal regulations, and the company is not a commercial fuel supplier [c8a]. That exclusion is the legal category built to let exactly this kind of work happen without a production licence.
Demand is not the soft spot. The Department of Energy says HALEU remains unavailable from domestic suppliers at the scale commercial deployment requires [10], and the NRC's HALEU band runs from 5% to 20% U-235 [2], so a research batch at 15.38% is squarely the material reactor developers say they are short of.
The timeline is worth reading as two clocks. Robert Mendelsohn spent seven years researching and prototyping the separation technology before he and Eric Olszewski incorporated Actinide in September 2025 [14]. Olszewski says he financed the work before the company existed, moved his family to Dallas in early 2024, and put more than $1 million of his own money into it [15]. The seed round, led by Onto Ventures with Neo Ventures, Mana Ventures, Discipulus Ventures, Shor Capital and angel investors, closed in March 2026 [16], about six months after incorporation [2]. Long prototyping, short industrial record, and institutional money that is younger than most of the engineering.
That gap explains the product framing. Actinide positions Endurance as a general-purpose separator for isotopes used in nuclear energy, medicine, quantum systems and electronics [18]. Selling materials that need no fuel licence is a credible way to fund the years of work that a fuel licence takes. It also means the fuel claim gets validated last, because it is the line with the least revenue attached and the most regulators watching. Actinide's own framing concedes the point: its distinction rests on startup status, the laboratory-scale regulatory basis and the electromagnetic process, not on being the first American organisation to make HALEU [13].
The figure that would settle this is kilograms per month from one Endurance, at 15% or better, at a purity a fabricator will accept, under something other than an exclusion. Until that exists, what happened in Texas is a well-measured small batch of the right material, produced by a route that would skip a step the country has not built.
Ranked by verification strength, evidence, and original report placement.
Actinide said Tuesday that a Texas-built electromagnetic isotope separator enriched a laboratory batch of uranium to 15.38% uranium-235.
An ISO/IEC 17025-accredited laboratory independently measured the sample, according to the company.
The Nuclear Regulatory Commission defines HALEU as uranium containing between 5% and 20% uranium-235.
The machine is a calutron called Endurance, an electromagnetic separator descended from the technology used at Oak Ridge during the Manhattan Project; Actinide reconfigured it for uranium and says it produced HALEU directly as a solid material.
Weeks before the uranium run, Endurance had been separating rare ytterbium isotopes for Oklo Isotopes.
Centrifuges output uranium hexafluoride gas, which must be converted back into a fuel-ready solid before advanced-reactor developers can use it, and the United States does not yet have a commercial HALEU deconversion industry.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One company announcement with an unpublished third-party assay
The central technical fact rests on a single publisher relaying a single company announcement plus a founder's post. An ISO/IEC 17025-accredited measurement is asserted but the laboratory, its report and the assay method are not identified, and no batch mass or throughput figure appears. Surrounding structural facts are better grounded, being drawn from NRC, DOE and EIA statements. The primacy claim is explicitly not independently established, and the cost, reconfiguration and Fortitude capacity figures are untested company estimates.
Research-scale only, one disclosed customer run
Adoption evidence is minimal: one laboratory uranium batch under a regulatory exclusion, and one prior ytterbium separation run for Oklo Isotopes. There is no licensed production, no fuel fabrication, no reactor-supply relationship and no disclosed revenue or offtake. Comparable adoption in the same market belongs to others: Centrus delivered 20 kilograms of HALEU to the government and incumbents received the January 2026 federal task orders.
Company framing runs ahead of a single lab batch
The claims being marketed - first startup to make HALEU, a machine that deletes the deconversion step, a second-generation unit at half the federal electromagnetic fleet's separation capacity - are considerably larger than the demonstrated evidence of one unquantified research-scale batch with no licensing, safeguards or fabrication path. The gap is positive but not extreme because the reporting itself discloses most of the limits, attributes the estimates to the company and states plainly that the HALEU shortage is unsolved.
Self-announced milestone from a recently funded startup
Every load-bearing technical figure originates with the company or its founders, and the announcement lands months after a seed round led by Onto Ventures and amid billions of dollars of federal commitments to domestic enrichment and deconversion, giving Actinide a direct interest in establishing priority and technical credibility with investors and government buyers. The founders also disclose more than $1 million of personal capital at stake. Countervailing incentive on the reporting side is limited, since only one publisher covers the story and it names the company announcement as its primary source.
Low-moderate: single publisher, no independent verification
Confidence is limited by a one-source, one-publisher cluster with no regulator, competitor or laboratory corroboration, and by the absence of the quantities that would make the result assessable. Confidence is not lower because the report is internally specific and hedged, attributes contested items, and anchors the surrounding supply-chain picture in NRC, DOE and EIA statements that are independently checkable.
product
TerraPower hires a builder: Hyundai E&C signed for up to eight Natrium units1 distinct publisher
product
Nuclea buys Moltex's entire patent shelf out of administration, in cash1 distinct publisher
product
Apollo Atomics bets nuclear's cost problem is the steam generator, not the reactor2 distinct publishers
product
A barge, a berth lease and a standards deal: how nuclear gets back into California1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 26, 2026