Blykalla Partnered With Microsoft to Automate Reactor Licensing
The firm intends to leverage generative AI to accelerate the regulatory approval process for lead-cooled reactors.
Updated on Sept. 22, 2026 in Artificial Intelligence
Live Poll
Do you trust the use of AI to automate regulatory licensing for nuclear energy plants?
Blykalla has partnered with Microsoft to utilize generative AI and agentic workflows to automate the drafting of complex nuclear regulatory filings. The collaboration aims to address the significant documentation requirements involved in securing approval for new reactor designs.
Why it matters
Nuclear reactor developers face extreme regulatory burdens that often delay project timelines by years. By automating the data evaluation and filing process, Blykalla expects to reduce the time and resource requirements for permitting by 90%.
Blykalla is targeting a 90% reduction in regulatory documentation efforts by deploying Microsoft's Energy Permitting Solution Accelerator. The AI system automates the evaluation of technical reactor data against established regulatory criteria to produce compliant permit applications.
The players
Blykalla
A reactor developer specializing in lead-cooled Small Modular Reactor designs and advanced steel materials.
Microsoft
A global cloud and software provider developing industry-specific AI solutions for energy sector permitting.
U.S. Nuclear Regulatory Commission
The federal agency responsible for regulating nuclear power plants and licensing reactor operations in the United States.
The details
The system employs agentic AI—autonomous software that completes tasks through multi-step reasoning—to process complex reactor data for regulatory review. Blykalla is concurrently developing proprietary aluminium-alloyed steel formulations designed to resist the corrosive effects of molten lead, which acts as a coolant in their Small Modular Reactor (SMR) designs. This material research complements the digital workflow by ensuring the physical components meet the performance standards documented in the AI-generated filings.
Timeline
2013: Blykalla spun out of KTH Royal Institute of Technology.
2024: Construction began on a non-nuclear electrical prototype in Oskarshamn.
February 2025: Groundbreaking occurred for the Oskarshamn testing facility.
March 2026: Blykalla opened a corporate office in New York.
2029: Target date for first criticality of the SEALER-One pilot plant.
The Tech Race
Blykalla is currently positioning its lead-cooled reactor technology to compete with traditional water-cooled SMR designs by streamlining the historically slow permitting phase. This software-enabled approach aims to accelerate the company's roadmap toward factory mass production of SEALER-55 units throughout the 2030s.
The transition to AI-assisted licensing directly influences the commercial timeline for regional power grids, with the Norrsundet park planning six SMRs and the Untra park aiming for eight units totaling 440-MW. Residents and utility customers can expect these facilities to potentially begin operations in the 2030s if the accelerated regulatory path succeeds.
The takeaway
The success of this initiative will be confirmed if the automated filings pass the rigorous review standards required by the U.S. Nuclear Regulatory Commission. Watch for the 2029 target date for first criticality of the SEALER-One plant as the primary indicator that these digital efficiencies are yielding physical results.
Further reading
For broader coverage of how developers are using generative tools in complex infrastructure, visit /tech/artificial-intelligence/.
Source note: This article includes information reported by Nuclear Engineering International.
Live Poll
Do you trust the use of AI to automate regulatory licensing for nuclear energy plants?
