RTS Nuclear + AI Briefing · May 2026
The small modular reactor story is moving from broad theme to specific deals. In spring 2026, small reactor developers and operators were matched to specific AI workloads, and the pace of announcements ran ahead of the regulatory, licensing, and operational frameworks meant to support them. That gap is where the most interesting work lives.
AI compute meets atomic power: NANO Nuclear and Supermicro sign an MOU
NANO Nuclear signed a strategic MOU with Supermicro to explore pairing NANO’s microreactor designs with Supermicro’s AI server infrastructure. NANO’s reactors are not yet operational, but the MOU joins a growing set of pre-deployment agreements pairing microreactor developers with data center customers.
The RTS perspective
Microreactors are entering the data center conversation faster than most operators expected. The deals that move from MOU to operating asset will be the ones where reactor developers, hyperscalers, and regulators align early on siting, licensing, and interconnection. The licensing pathway behind the MOU is the harder problem.
Source: NANO Nuclear Energy
Fuel and compute: Oklo, Los Alamos, and NVIDIA partner on plutonium-bearing fuels
Oklo announced a partnership with Los Alamos National Laboratory and NVIDIA to pursue AI-enabled research on plutonium-bearing fuels, with science-based AI models supporting materials science, fuel performance modeling, and fabrication R&D.
The RTS perspective
Fuel is one of the longest-lead constraints on advanced reactor deployment. Applying AI to candidate screening and materials work this early shortens the discovery and modeling timeline without bypassing the experimental, qualification, and licensing steps that still have to follow.
Source: American Nuclear Society
First watts from the TRIGA: University of Utah to generate electricity from a research reactor
The University of Utah’s TRIGA research reactor will generate electricity for the first time in summer 2026, in partnership with Elemental Nuclear Energy, to run a “mini AI data center” on campus.
The RTS perspective
This is a meaningful demonstration, not a commercial deployment. Expect more projects like it, and expect them to surface 10 CFR Part 50 licensing questions around possession limits, fuel use, and authorized end-use of generated electricity that have not been worked through at scale.
Source: American Nuclear Society
Research news
- Ames National Laboratory released DuctGPT, combining machine learning with physics-based modeling to predict fusion-grade alloys, cutting discovery from months to hours.
- Texas A&M released AROMA-GPT, a research tool pairing generative AI with a digital twin for advanced reactor monitoring, with human operators kept in the control loop.
- PJM reopened its interconnection queue with 220 GW filed, including 18 GW of nuclear.
- The OECD NEA RegLab published first-cycle findings on AI in real-time nuclear plant monitoring.
What this signals
- Microreactor developers are signing pre-deployment agreements with data center hardware companies.
- Advanced fuel R&D is being accelerated with AI and high-performance computing.
- A university research reactor will demonstrate electricity for a campus AI workload.
- Regulators are starting to use AI on their own side of the table.
The organizations that lead will treat SMR deployment as a near-term execution challenge and integrate AI within nuclear standards, workforce models, and regulatory expectations from day one.
Talk to RTS
One conversation. No pitch deck. Schedule a roadmap session or browse RTS research.
Comments are closed