Two companies are about to test whether small nuclear reactors can power the next generation of artificial intelligence data centers. Aalo Atomics and Crusoe Energy announced a strategic partnership on July 30 to develop a nuclear-powered "AI Factory" data center that combines Aalo's advanced reactor technology with Crusoe's modular data center platform. The demonstration will begin in 2027 at Idaho National Laboratory (INL), marking a concrete step toward solving one of AI's biggest challenges: finding enough clean, reliable power.

Why Is Nuclear Power Suddenly Critical for AI?

For 15 years, electricity demand in the United States barely budged. From 2005 to 2020, businesses and consumers used roughly the same amount of total energy year after year, meaning almost no new power plants were built. Then everything changed. U.S. power consumption hit its second straight annual record high in 2025, driven almost entirely by energy-hungry artificial intelligence data centers. Experts now expect electricity demand to grow consistently for decades to come, a dramatic reversal that has caught the power industry's attention.

The commercial sector, which includes data centers, is expected to outpace residential electricity demand in 2026 for the first time on record. This surge has created an urgent problem: there simply isn't enough power infrastructure to support the buildout of AI globally. Goldman Sachs concluded in a recent report that "a lack of capital is not the most pressing bottleneck for AI progress, it's the power needed to fuel it". That realization has made nuclear energy suddenly attractive again, particularly in the form of small modular reactors, or SMRs.

Goldman Sachs

What Are Small Modular Reactors, and Why Do They Matter for AI?

Small modular reactors are essentially miniature nuclear power plants. Unlike conventional nuclear facilities, which take decades to build and cost billions of dollars, SMRs have smaller footprints, can be constructed faster, and are significantly cheaper to deploy. Bank of America identified five major advantages SMRs offer over traditional large-scale nuclear plants:

• Lower Upfront Costs: Smaller capital requirements make nuclear power accessible to more companies and regions.
• Enhanced Safety: Modular design includes passive safety features that reduce accident risk.
• Modularization: Multiple reactor units can be added together to increase output capacity as demand grows.
• Smaller Physical Footprints: Reactors require less land, making them suitable for data center campuses.
• Reduced Carbon Dioxide Production: Nuclear power generates electricity without greenhouse gas emissions.

For AI data centers, which consume enormous amounts of electricity continuously, SMRs offer a way to secure dedicated, carbon-free power without relying entirely on the grid. This is why companies like Aalo and Crusoe are moving fast to commercialize the technology.

What Is Aalo's Plan, and How Far Along Are They?

Aalo Atomics has made remarkable progress in a short time. The company's Critical Test Reactor (CTR) reached criticality at INL on July 4, making it the fourth Department of Energy-authorized advanced reactor to achieve this milestone in about a month. Criticality means the reactor achieved a sustained nuclear chain reaction, a crucial proof that the core physics and control systems work as designed.

Aalo is pursuing a staged development approach. The CTR is a zero-power physics reactor designed to validate core behavior and operating procedures. Nearby, the company is building Aalo-0, a full-scale sodium test system that will circulate 60,000 pounds of sodium under operating conditions to validate heat transport and power conversion. Data from both projects will feed into Aalo-X, a planned 30-megawatt thermal (10-megawatt electric) demonstration plant that Aalo expects to operate in the coming months.

The ultimate goal is the Aalo Pod, a 50-megawatt extra-modular reactor (XMR) purpose-built for data centers. The Pod uses five 10-megawatt reactors connected to a shared turbine, arranged as 50-megawatt modular blocks that can scale to larger campuses. Aalo says the Pod is designed for on-site, grid-parallel, hybrid, or independent operation, with a compact footprint of 100 megawatts on less than five acres and limited water requirements.

"As one of only four companies to reach criticality by President Trump's deadline for companies in the DOE Reactor Pilot Program, we are now well positioned for not only the full power operation of our 10 megawatt Aalo-X reactor, but also the deployment of the Aalo Pod to power commercial data centers," said Matt Loszak, CEO of Aalo Atomics.

Matt Loszak, CEO, Aalo Atomics

How Will the Aalo-Crusoe Partnership Work?

Under the partnership announced July 30, Aalo will initially power a Crusoe Spark modular data center running Crusoe Cloud at INL in 2027 as a proof of concept. This is a critical test: it will demonstrate whether a small nuclear reactor can reliably power a real AI data center workload in a commercial setting. Over the longer term, Aalo and Crusoe intend to deploy Aalo Pods and Aalo's 50-megawatt XMR power plants at Crusoe data centers by the end of 2029.

Crusoe Spark is the company's modular AI factory product, manufactured at a 352,000-square-foot factory in Brighton, Colorado. Spark is a turnkey, prefabricated modular data center platform that integrates power, cooling, remote monitoring, fire suppression, and high-density racks into repeatable units. The company says Spark deployments can scale from hundreds of kilowatts to tens or hundreds of megawatts and can be delivered in as little as three months. The first factory-produced Spark modules are expected in the third quarter of 2026.

"The next generation of AI infrastructure demands power that is reliable, abundant, clean and at scale, and Aalo's nuclear reactors are built to meet that need," said Cully Cavness, co-founder, president, and Chief Strategy Officer of Crusoe.

Cully Cavness, Co-Founder, President and Chief Strategy Officer, Crusoe

How to Evaluate Nuclear Power Options for Data Centers

As companies race to secure power for AI infrastructure, they face a strategic choice about how to source electricity. Here are the key approaches emerging in the market:

• Grid-Connected Utility Scale: Companies like NuScale Power focus on larger, utility-scale SMR deployments through direct partnerships with major utility companies. This approach assumes most AI data centers will connect to the grid in a conventional manner, with utilities purchasing nuclear power and distributing it to customers.
• On-Site Direct Generation: Companies like Oklo take a more direct approach by selling reactor systems directly to AI data centers for on-site power generation. Data center operators produce their own power next to their facilities, reducing reliance on grid connections and cutting out the utility middleman.
• Hybrid Bridge Models: Some companies, like Blue Energy, use natural gas generation to supply AI data center load immediately, then transition to SMR power as nuclear units come online. This approach allows data centers to begin operations while waiting for nuclear reactors to be built and licensed.

Each approach has trade-offs. Grid-connected models benefit from existing infrastructure and utility expertise but depend on regulatory timelines and grid capacity. On-site models give data centers direct control over their power supply but require companies to become nuclear operators themselves. Hybrid models spread risk but add complexity and cost.

What Happens Next?

The Aalo-Crusoe partnership is significant because it tests a specific model: a smaller, nuclear-first pairing between an on-site advanced reactor and a modular data center platform. Unlike the Blue Energy model, which uses gas as a bridge to later SMR deployment, Aalo and Crusoe are going straight to nuclear. If the 2027 INL demonstration succeeds, it could validate the entire approach and accelerate deployment at commercial data center campuses.

However, success is not guaranteed. Aalo's claims about delivering Pods within 12 months of order placement remain contingent on successful Aalo-X operation, factory scale-up, regulatory approvals, fuel supply, site-specific requirements, and customer deployment conditions. The company is moving faster than traditional nuclear vendors, but nuclear technology still faces regulatory hurdles, supply chain challenges, and the need to train a workforce capable of operating these systems safely.

What is clear is that the AI industry's power crisis has made nuclear energy a serious contender again. For decades, nuclear power was considered too expensive and too slow to build. Now, with AI data centers consuming gigawatts of electricity and demanding clean, reliable power, small modular reactors are no longer a niche technology. They are becoming a central part of how the world plans to power artificial intelligence.