South Korea has selected four university-affiliated research labs to receive up to 10 billion won per year for the next decade, with one lab specifically focused on developing AI-autonomous small modular reactors (SMRs) to meet surging energy demands from artificial intelligence infrastructure. The announcement marks a strategic pivot toward combining nuclear technology with AI control systems, addressing a critical bottleneck in powering the next generation of data centers and industrial facilities.

Why Are Governments Pairing AI With Nuclear Power?

The convergence of AI and nuclear energy reflects a fundamental challenge facing the global energy transition. As artificial intelligence systems consume ever-increasing amounts of electricity, traditional power grids struggle to keep pace. Small modular reactors offer a compact, factory-built alternative to massive conventional nuclear plants, but operating them efficiently at scale requires sophisticated automation. This is where AI enters the picture.

The "SMR2 Platform National Research Lab," led by Professor Jae-Sun Lee of Changwon National University, will focus on building an integrated platform for nuclear power plant-specialized core materials, structural integrity, energy conversion, and system integration. The lab's research agenda includes AI autonomous operation of nuclear power plants, verification of structural integrity in extreme environments, and defect removal technologies.

This approach addresses a real operational challenge. SMRs are smaller and more modular than traditional reactors, but they require constant monitoring and adjustment to maintain safe, efficient operation. AI systems can automate these tasks, reducing the need for large on-site staffing and enabling faster response to fluctuations in power demand, particularly from data centers that experience unpredictable computational loads.

What Technologies Are Being Developed Alongside SMRs?

South Korea's investment in SMR research is part of a broader global effort to develop next-generation nuclear fuels and control systems. The selected labs represent a diverse portfolio of energy and AI innovations:

• Physical AI Robotics: Seoul National University's lab, led by Professor Kyu-Jin Cho, will develop human-centered physical AI robotics technologies that mimic human sensory and motor nervous systems, with applications in manufacturing and elder care.
• Intelligent Energy Solutions: Sungkyunkwan University's lab, headed by Professor Nam-Gyu Park, will combine high-efficiency solar cells and energy storage technologies with AI and digital twins to support industrial electrification and power supply for AI data centers.
• Precision Medicine: Chungnam National University's lab, led by Professor Hak-Soo Choi, will develop theranostics technologies that combine diagnosis and therapy for intractable diseases.

The SMR2 Platform lab represents the most direct intersection of nuclear and AI technologies. According to the announcement, the lab is expected to accelerate the development of core SMR materials, system integration and expansion technologies, and hardware virtual verification technologies.

How Are SMRs Different From Traditional Nuclear Plants?

Small modular reactors occupy just 10 percent of the space needed for a traditional nuclear plant and are designed to be built in sections in a factory, then shipped directly to a site for assembly. This modular approach enables faster and more affordable construction compared to conventional reactors, which can take a decade or more to build and cost billions of dollars.

The BWRX-300, a water-cooled SMR developed by GE Vernova Hitachi, exemplifies this design philosophy. These compact reactors can provide reliable, 24/7 near-zero-carbon power without the massive infrastructure footprint of traditional nuclear facilities. For data centers and industrial operations seeking to decarbonize while meeting intense power demands, SMRs offer a practical middle ground between renewable energy's intermittency and conventional nuclear's high capital costs.

However, SMRs require specialized fuels to operate efficiently. Most advanced SMRs are designed to use high-assay low-enriched uranium (HALEU), which is enriched to more than 5 percent and less than 20 percent uranium-235. Currently, only Russia and China produce HALEU at commercial scale, creating a supply chain vulnerability for countries seeking energy independence.

What Fuel Supply Challenges Are Slowing SMR Deployment?

The global race to develop SMRs has hit a critical bottleneck: fuel availability. Following the United States ban on Russian uranium imports in 2024, the U.S. government has focused efforts on developing domestic HALEU production capacity. Centrus Energy produced over 920 kilograms of HALEU from a demonstration facility at Piketon, Ohio, between October 2023 and mid-2025, but this remains a fraction of what would be needed to power hundreds of SMRs.

In January 2026, the U.S. Department of Energy earmarked $2.7 billion to expand domestic uranium enrichment capacity over the next decade. The United Kingdom allocated £300 million to support HALEU production in January 2024. Despite these investments, some companies are pursuing alternative fuels to accelerate deployment.

"We know we want to get to market fast, and we know we need to scale up to build hundreds of reactors, and we can't do that with HALEU for many years, because the U.S. is still pumping money into that HALEU machine, trying to figure out how to crack the code," explained Yasir Arafat, Chief Technology Officer at Aalo Atomics.

Yasir Arafat, Chief Technology Officer at Aalo Atomics

Companies such as GE Hitachi, Westinghouse, and Aalo Atomics have opted for Low Enriched Uranium Plus (LEU+), which has a uranium-235 concentration between 5 and 10 percent, rather than HALEU. LEU+ can be purchased from existing U.S. facilities, avoiding dependence on Russia or China. Urenco USA was authorized by the U.S. Nuclear Regulatory Commission to produce LEU+ at its facility in Eunice, New Mexico, last September and expects to achieve commercial production by mid-2026.

How to Understand the AI-Nuclear Connection in Practice

The integration of AI with nuclear power operates on multiple levels, from grid management to reactor operation:

• Grid Orchestration: AI systems like Zonal Autonomous Grid Control (ZAC) act as automated air traffic controllers for electricity, automatically balancing the grid, managing surges, preventing outages, and integrating renewable wind and solar power while keeping electrons moving smoothly throughout the network.
• Reactor Autonomy: AI can monitor reactor performance in real time, predict maintenance needs before failures occur, and adjust operations to respond to fluctuating power demands from data centers without human intervention.
• Supply Chain Optimization: AI and digital twins can simulate reactor performance under extreme conditions, accelerating the development of new materials and designs while reducing the need for physical prototypes.

GE Vernova's Advanced Research Center is developing these intelligent systems to manage the increasingly complex puzzle of lower-carbon energy flow as the world rapidly electrifies. At the Aspen Ideas Festival in June 2026, the company showcased an interactive game where attendees stepped into the shoes of a grid operator, facing real-time power challenges to see firsthand how digital intelligence can safeguard energy networks.

South Korea's investment in the SMR2 Platform lab signals that the country recognizes AI-autonomous nuclear operation as essential infrastructure for the AI economy. As data centers consume more electricity and governments push for decarbonization, the combination of compact nuclear reactors and intelligent control systems offers a pathway to reliable, clean power that can scale rapidly without the decades-long construction timelines of traditional nuclear plants.

The first year of funding will provide 5 billion won, equivalent to six months of support, with the government noting that the scale of support may change depending on budget circumstances. Following the announcement on June 29, the government will finalize the selected institutions after a period for objections and will fully launch the project through agreements with implementing organizations.