Nuclear startup Oklo Inc. is running plutonium experiments at a Nevada facility once reserved for weapons work, partnering with Los Alamos National Laboratory and chipmaker Nvidia to validate fuel for a new class of commercial microreactors. The collaboration, anchored by federal programs with a July 4, 2026 deadline, represents one of the most ambitious attempts yet to move plutonium from the national security domain into the commercial energy sector.
As of late April 2026, the clock is ticking. The Department of Energy’s Reactor Pilot Program aims to “construct, operate, and achieve criticality of at least three test reactors” by Independence Day, roughly ten weeks away. Oklo holds two of the program’s 11 selected project slots, making it the most heavily represented company in the initiative.
Inside the Nevada experiments
The plutonium validation work is taking place at the National Criticality Experiments Research Center, or NCERC, a facility at the Nevada National Security Site operated by Los Alamos National Laboratory. NCERC houses critical assembly machines with names like Godiva, Flattop, Planet, and Comet, all designed to study how fissile material behaves at the threshold of a sustained chain reaction. These are not power-generating reactors. They operate at essentially zero power, functioning as precision instruments for nuclear physics measurements.
Oklo’s access runs through a Strategic Partnership Project, or SPP, the mechanism the National Nuclear Security Administration uses to pair its modeling, simulation, and experimental capabilities with outside partners. For a private company to conduct plutonium criticality experiments at a weapons-complex facility is notable on its own. For decades, plutonium handling in the United States has been almost exclusively a national security activity. Opening that infrastructure to a commercial reactor developer reflects a deliberate policy shift, reinforced by a May 2025 presidential directive on reforming nuclear reactor testing at the DOE.
Where Nvidia fits in
The AI connection comes through the Genesis Mission, a DOE-led national effort established by presidential executive action that unites department laboratories, industry, and academia around an integrated AI platform. The mission calls for identifying federal and industry computing resources to accelerate scientific goals, and Nvidia is among the industry partners contributing to that computing infrastructure.
In practical terms, AI-driven simulation could help Oklo model neutron behavior, predict fuel performance, and reduce the number of physical experiments needed to qualify a new fuel type. But the specifics remain thin. No available DOE or NNSA documentation spells out whether Nvidia is providing GPU clusters for simulation, training physics-informed machine learning models, or playing a more limited advisory role. The Genesis Mission documents outline intent and structure but do not tie deliverables to individual companies.
Why plutonium fuel matters
Oklo’s Aurora microreactor is a sodium-cooled fast reactor, a design that can potentially run on fuel derived from spent nuclear material, including plutonium. If the company can demonstrate that plutonium-based fuel works reliably, it could open a pathway to recycling waste that currently sits in storage at sites across the country, reducing both the volume of long-lived radioactive material and the demand for freshly mined uranium.
That potential, however, comes with significant technical and regulatory hurdles. The zero-power experiments at NCERC measure fundamental nuclear physics parameters, such as critical mass and neutron multiplication, but they do not replicate the thermal stress, mechanical wear, or radiation damage that fuel experiences inside an operating reactor. How Oklo plans to bridge the gap between NCERC benchmarking data and full fuel qualification, whether through DOE authorization or the traditional Nuclear Regulatory Commission licensing process, has not been publicly detailed.
Oklo’s history with the NRC adds context here. The company’s initial license application for the Aurora reactor was denied in 2022 after the commission found the submission lacked sufficient technical detail. Oklo has since resubmitted, but the Reactor Pilot Program’s use of DOE authorization under the Atomic Energy Act, rather than NRC licensing, offers a parallel and potentially faster route to operation.
What the deadline really means
The July 4, 2026 target applies to the Reactor Pilot Program as a whole, not specifically to Oklo’s plutonium fuel validation. The program’s language calls for “at least three” of its 11 selected projects to reach criticality by that date, leaving open which companies will cross the finish line first. Whether Oklo’s two projects are among the frontrunners, or whether the deadline functions more as an aspirational benchmark, is not clear from publicly available DOE materials.
For the broader advanced nuclear industry, the combination at work here is striking: weapons-lab experimentation, AI-accelerated simulation, and a federal authorization pathway designed to bypass years of traditional regulatory review. The institutional infrastructure is real, the deadlines are set, and the policy direction is unmistakable. What remains unproven is whether the technical path from zero-power criticality experiments to a functioning commercial reactor can be traveled as quickly as the program’s architects intend.
Investors and energy planners watching Oklo should pay close attention to what happens at NCERC over the next several months. The data generated there will determine whether plutonium-fueled microreactors move from policy aspiration to engineering reality, or whether the gap between federal ambition and demonstrated performance proves wider than the timelines suggest.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
