Aalo and Crusoe Target Nuclear AI Demonstration

Aalo Atomics and Crusoe formed a partnership to deploy a nuclear-powered Crusoe Spark modular data center at Idaho National Laboratory in 2027. Data Center Dynamics reports that the proof of concept would pair the facility with Aalo's 50 MWe Pod reactor, composed of five 10 MWe Aalo-1 microreactors and one turbine. The companies also stated an intention to deploy Aalo Pods across Crusoe's portfolio by the end of 2029.
Aalo Atomics and AI infrastructure company Crusoe announced a partnership to demonstrate a nuclear-powered AI data center at Idaho National Laboratory (INL) in 2027. According to Data Center Dynamics, Crusoe will deploy a modular Crusoe Spark data center powered directly by an Aalo Pod reactor as a proof-of-concept installation.
The proposed Aalo Pod is rated at 50 MWe and combines five 10 MWe Aalo-1 microreactors with a single power-generating turbine, Data Center Dynamics reports. Interesting Engineering reports that the initial Spark unit is intended to run Crusoe Cloud at the INL site.
A demonstration, not an operating data center
The 2027 deployment remains a proposed demonstration. Public reporting describes it as an effort to validate how a small modular nuclear reactor can supply an AI data center load, rather than evidence that the facility is already delivering nuclear-powered compute.
Crusoe co-founder, president, and chief strategy officer Cully Cavness said the next generation of AI infrastructure requires power that is "reliable, abundant, clean and at scale." He added that the companies look forward to jointly demonstrating what Crusoe called the first AI data center powered by a small modular nuclear reactor.
Aalo has recently reached criticality with its Aalo-X Critical Test Reactor at INL under the US Department of Energy's Reactor Pilot Program, according to Data Center Dynamics. Criticality denotes the point at which a reactor sustains a nuclear chain reaction. Data Center Dynamics reported that Aalo was one of four advanced nuclear companies to reach that milestone under the program.
Interesting Engineering also reported that Aalo has begun work on a second reactor beside the Aalo-X test reactor at INL. The publication described that work as testing a commercial-scale system intended to generate electricity and supply an on-site Crusoe Spark facility.
Potential fleet deployment
Beyond the INL demonstration, Data Center Dynamics reports that Aalo and Crusoe stated they intend to deploy Aalo Pods across Crusoe's data center portfolio by the end of 2029. That target is contingent on the engineering, regulatory, construction, fuel, and interconnection steps required for commercial reactor deployment, none of which are detailed in the available reporting.
The project is notable because it pairs a 50 MWe generation unit with modular AI data center infrastructure. That capacity does not on its own establish available AI compute capacity: usable capacity also depends on the data center's electrical topology, cooling systems, storage, networking, uptime design, and the portion of generation reserved for non-IT loads.
Comparable on-site generation projects are typically evaluated against grid interconnection schedules, delivered energy cost, capacity factor, operational staffing, and the redundancy required for production workloads. The INL deployment could provide an early real-world reference point for those questions if it reaches operation, particularly around coupling modular reactor output with modular AI data center infrastructure.
It does not establish that small modular reactors are commercially ready for broad AI data center deployment, but it creates a defined 2027 test case involving a reactor developer, an AI cloud provider, and a national laboratory site.
Key Points
- 1Aalo and Crusoe target a 2027 INL proof of concept pairing a 50 MWe reactor pod with modular AI data center infrastructure.
- 2The deployment could generate operational evidence on direct nuclear supply for AI workloads, beyond proposed reactor and data center capacity claims.
- 3Comparable AI power projects require evaluation of interconnection, cooling, redundancy, energy economics, and regulatory execution alongside generation capacity.
Scoring Rationale
The proposed INL demonstration addresses a major infrastructure constraint for AI practitioners: reliable power for high-density compute. It is a notable but pre-operational project, so its broader significance depends on execution, regulatory progress, and published operating results.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems

