Four advanced nuclear reactors reached a self-sustaining chain reaction for the first time in the weeks before July 4, 2026, each one built by a startup and authorized under a federal program designed to accelerate small reactor technology. Aalo Atomics’ Aalo-X was the last to cross the line, completing its zero-power fueled criticality demonstration at Idaho National Laboratory and becoming the fourth reactor to hit the milestone ahead of the holiday deadline. The rapid-fire sequence of criticalities, all achieved within weeks of one another, signals the most aggressive push in decades to prove that compact reactors can be built fast enough to meet surging electricity demand from artificial intelligence data centers and other high-draw industrial users.
Four criticalities in weeks and what they mean for power-hungry data centers
The Department of Energy’s Reactor Pilot Program set a clear target: get multiple advanced reactors to criticality by July 4. The agency hit that mark with room to spare. The sequence began when Antares’ Mark-0 went critical, a milestone that Energy Secretary Chris Wright discussed on the record. Valar Atomics’ Ward 250 followed, then Deployable Energy’s Unity, and finally Aalo Atomics’ Aalo-X. Each demonstration confirmed that its reactor’s fuel and core geometry could sustain a controlled chain reaction at zero power, the essential prerequisite before any reactor can generate electricity.
The clustering of these milestones was not accidental. The DOE framed the July 4 date as a presidential goal, and each company operated under federal authorization to conduct its test. The underlying bet is straightforward: if startups can prove reactor physics quickly under government oversight, they can build a track record that supports faster commercial licensing later. That matters because hyperscale data center operators need firm, carbon-free power sources that can be sited close to their facilities, and conventional nuclear plants take a decade or more to build.
A useful way to test whether these demonstrations amount to a real strategy or a symbolic exercise is to watch for commercial contracts. If the companies that just proved criticality begin announcing power-purchase agreements with data center operators or cloud providers in the coming months, it will confirm that the DOE’s pilot timeline was designed to feed directly into a commercial pipeline. If no contracts follow, the demonstrations will have been engineering achievements without a clear market path.
Who built what, where, and under which federal program
The four reactors span different designs, companies, and locations, but all operated under DOE authorization. According to a department summary, Aalo Atomics’ Aalo-X achieved criticality at Idaho National Laboratory and was designated the fourth DOE-authorized advanced reactor to reach the milestone. The agency emphasized that the timing exceeded its July 4 goal, underscoring how tightly the demonstrations were tied to federal scheduling and oversight.
Valar Atomics’ Ward 250 stands out for a geographic reason: it completed its demonstration at the Utah San Rafael Energy Lab in Emery County, Utah, making it the first DOE-authorized reactor built outside a national laboratory. That distinction matters because commercial deployment will eventually require reactors to operate on private or utility-owned sites, not inside federal research campuses. Proving that a startup can assemble and fuel a reactor at an off-site facility removes one of the practical objections to rapid deployment and hints at how future commercial units might be sited near industrial loads instead of remote research reservations.
Deployable Energy’s Unity completed its own zero-power fueled criticality demonstration at Idaho National Laboratory and was recognized as the third DOE-authorized advanced reactor to go critical by the deadline. The DOE’s program structure splits these efforts between the Reactor Pilot Program and the Nuclear Energy Launch Pad, both of which provide access to federal sites, fuel, and technical support. In practice, that means startups can lean on federal infrastructure for early tests instead of building full-scale nuclear campuses from scratch.
Antares’ Mark-0 was the first of the four to reach criticality. The DOE announced that milestone separately, and Wright’s public comments framed it as proof that the pilot model-pairing private designs with federal sites and fuel-could deliver results on a political timetable. Together, the four reactors represent the largest batch of new reactor criticalities in the United States in years, all completed within a compressed window that was explicitly linked to a presidential directive.
Gaps between a chain reaction and a plugged-in power plant
Zero-power criticality is a physics proof, not an engineering finish line. None of the four reactors has generated electricity. None has connected to a grid or delivered power to a customer. The DOE’s own descriptions consistently use the phrase “zero-power fueled criticality demonstration,” which means the reactors sustained a chain reaction without producing usable thermal or electrical output. The gap between that step and a functioning power plant is significant: it includes thermal testing, power ramp-up, safety validation, and regulatory clearance for commercial operation.
Several questions remain open. No primary DOE records or company filings available in the public record confirm that any of these four reactors was explicitly designed to power AI data centers, even though the broader policy conversation ties small reactors to data center demand. Direct statements about customer commitments, grid interconnection studies, or on-site power infrastructure are absent from the DOE summaries. That leaves a disconnect between the political narrative-advanced reactors as a solution for AI’s energy appetite-and the documented design and siting choices for these specific demonstration units.
Regulatory pathways are another uncertainty. The criticality tests were conducted under DOE authorization, which allows for experimental operation on federal property or at designated partner sites. Commercial deployment, however, will require approval from the Nuclear Regulatory Commission or relevant state authorities. The experience gained during these zero-power runs may help the companies refine their safety cases and operating procedures, but it does not automatically translate into a streamlined licensing process for full-power plants located near data centers or industrial parks.
Engineering scale-up also looms. The demonstration cores were sized and instrumented for controlled, low-power experiments, not for long-term commercial duty cycles. To turn them into revenue-generating assets, the companies will need to validate fuel behavior at higher temperatures, integrate heat removal systems, and prove that their designs can handle transient events and maintenance outages. Each of those steps introduces new technical risks and potential delays that the current DOE announcements do not address.
What to watch next
The four criticality milestones show that advanced reactor startups can move quickly under a focused federal program. Whether that momentum translates into new, reliable power for data centers will depend on developments that have not yet appeared in official records. Power-purchase agreements, siting applications for commercial-scale units, and detailed safety filings will be the next indicators of whether these designs are headed toward real-world deployment or will remain confined to the pilot phase.
For now, the July 4 deadline has done its job: it forced a cohort of companies to prove that their cores work as advertised. The harder task-turning those physics demonstrations into grid-connected plants that can shoulder the relentless load of AI infrastructure-still lies ahead, and it will require more than a celebratory countdown to complete.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
