Product1 distinct publisher3 min readUpdated
The Kaleidos microreactor left El Segundo for Idaho National Laboratory on August 12, bound for a five-phase test campaign that ends with 150 hours of unattended full-power operation.
The Product Desk · Product desk

Compiled by The Product DeskSomething wrong?How this is made
Radiant's 1-MWe Kaleidos microreactor began a journey of more than 1,000 miles from El Segundo, California, to Idaho National Laboratory on August 12, riding on a trailer bed [1][2]. The trip is not logistics in service of the product; the company is presenting it as evidence about the product, having promised a reactor transportable by land, sea and air and then pointing at the trailer as proof [7].
That framing only works if the thing on the trailer is the thing customers would buy. Radiant says it is: same design, same scale, same commercial fuel load as the units intended to come off the line at its proposed R-50 factory in Oak Ridge, Tennessee, and "not a simplified stand-in" [8][9]. Test articles in nuclear development are usually derated, partially fuelled, or otherwise trimmed, so building the demonstration unit to production configuration moves the regulatory and manufacturing risk earlier rather than deferring it. It also means a failure in Idaho is a failure of the shipping configuration, not of a prototype that can be quietly redesigned.
At Idaho the reactor goes inside DOME, the Demonstration of Microreactor Experiments test bed, which reuses the containment structure of a former experimental breeder reactor and exists so developers can run experimental designs and gather performance data [3][10]. Once installed, Kaleidos will be loaded with TRISO fuel under Department of Energy and Idaho National Laboratory oversight [11]. Standard Nuclear fabricated that fuel to Radiant's specifications earlier in 2026 [12].
Then comes the sequence that matters: zero-power criticality, heat and power generation, full-power operation, and a closing 150-hour continuous run with no operator intervention [4]. Radiant says every phase feeds data directly into its US Nuclear Regulatory Commission licence application and de-risks manufacturing and deployment at commercial scale [17][5]. The capstone is the interesting claim. An unattended 150-hour run is an assertion about staffing economics as much as about physics, and it is the kind of data a regulator can be shown rather than argued with. One wrinkle in the company's own language: the quoted description refers to "1 MWe thermal operation," mixing electrical and thermal units for a unit elsewhere described as 1 MWe [18].
The paperwork is running in parallel with the hardware. Radiant's Part 70 application for the R-50 production facility is already under NRC review, with the company targeting approval in the fourth quarter of 2026 [13]; it is targeting completion of the Idaho campaign in the third quarter of 2026 [6]. That leaves roughly one quarter between the last test data and the approval Radiant wants [19] - a tight coupling that assumes the campaign runs clean and on schedule. Beyond that sits an agreement with the US Air Force to deliver Radiant's first commercial microreactor to Buckley Space Force Base in 2028, which the company says the testing campaign will ensure [14][15]. That is about two years from the target end of testing to a delivered unit at a military site [20].
What to watch: whether zero-power criticality happens on the schedule implied by a third-quarter 2026 finish, since every later phase inherits any slip [6][4]; whether the 150-hour run is completed in one uninterrupted block or restarted; and whether the NRC's Part 70 review of the Oak Ridge factory moves independently of the Idaho results or waits on them [13]. Also worth noting for anyone dating this timeline precisely: the source gives August 12 as the departure without a year, while placing fuel fabrication earlier in 2026 [16].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Radiant's 1-MWe Kaleidos microreactor began traveling from El Segundo, California, to Idaho National Laboratory on August 12.
The reactor is being transported more than 1,000 miles on a trailer bed.
At Idaho National Laboratory, Kaleidos will be installed inside the Demonstration of Microreactor Experiments (DOME) test bed.
Radiant's five-phase testing program begins with zero-power criticality and advances to heat and power generation, full-power operation, and a final 150-hour continuous run without operator intervention.
The company will use data from the testing program for its US Nuclear Regulatory Commission license application.
Radiant is targeting completion of the testing campaign in the third quarter of 2026.
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.
Single outlet, company-sourced
Every substantive assertion - shipment, design equivalence, test plan, fuel provenance, regulatory target, customer delivery date - traces to one publisher relaying Radiant statements and direct quotes. No independent confirmation from Idaho National Laboratory, DOE, the NRC, Standard Nuclear or the Air Force appears in the supplied material, and no test results exist yet because testing has not begun. Two unresolved internal inconsistencies (unstated shipment year, '1 MWe thermal') further limit the evidentiary floor.
Hardware moved, nothing operating
There is one concrete physical event: a reactor on a trailer en route to a national-lab test bed. Fuel exists and is fabricated, and a licensing application is in review, but the reactor has not achieved criticality, no test phase has been completed, no power has been delivered to any user, and the sole committed field deployment is a 2028 Air Force delivery. Adoption is therefore pre-operational logistics plus paper commitments.
Milestone framed as validation
The company's framing runs ahead of the demonstrated facts: a road move is presented as having proven land/sea/air transportability, the untested unit is declared 'not a simplified stand-in', and the campaign is said to 'ensure' a 2028 base delivery. Against that, the measured record is a shipment in progress with zero completed test phases, an ungranted Part 70 license, and only a single quarter of slack between the Q3 2026 test target and the Q4 2026 approval target. The overstatement is in certainty of outcome rather than in fabricated events, so the gap is moderate, not extreme.
Vendor narrative before licensing
The claims originate from a company with two concrete pending outcomes it can influence through public confidence: an NRC Part 70 application under review for its Oak Ridge production facility and a signed first-delivery agreement with the US Air Force. Emphasising that the test unit is production-identical and that the campaign guarantees the 2028 delivery serves both the regulatory file and the customer relationship. The supplied source contains no independent voice to offset that positioning.
Plausible event, thin verification
The core event is specific, dated and physically checkable, and the surrounding institutional details (DOME, TRISO, Part 70, Buckley) are internally coherent, which supports moderate confidence in the story's shape. Confidence is held down by single-publisher sourcing, total reliance on company statements, an unstated year on the departure date, and a mismatched power-unit description in the key quote.
product
Radiant locks in TRISO to the early 2030s, treating fuel as the binding constraint1 distinct publisher
build
Nuclear AI program adds an AI security vendor, and the $60M is not the company's1 distinct publisher
product
GE Vernova and SHINE argue nuclear accounting belongs in the design, not the retrofit1 distinct publisher
product
The Air Force just published its shopping list of threat signatures it wants to shoot at cheaply1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.