Microsoft is using AI to get its nuclear projects approved in the US
Microsoft is using an AI model to help speed up the approval process on nuclear reactor projects
Microsoft is using large language models (LLMs) to expedite the approval process for nuclear power projects that will fuel its future artificial intelligence (AI) and supercomputing energy requirements.
A team at Microsoft has spent the previous six months training an AI model with US nuclear regulations and licensing documents to assist with generating paperwork for nuclear power projects, according to the Wall Street Journal.
The tech giant hopes the model will help speed up the lengthy and costly approvals process for nuclear power projects using the latest reactor technologies.
The energy demands of AI systems are expected to be a significant constraint for organizations moving forward, particularly for hyperscalers hosting data centers underpinning the biggest foundation models.
As hyperscalers grow and these models become more sophisticated and provide new functionalities, their energy demands will soar.
Microsoft’s senior director Michelle Patron said if the company wants to scale its models in a sustainable way it will need to use every weapon in its arsenal, including AI.
“If we're going to do that [grow] carbon-free, we're going to need all the tools in the tool kit,” she told the WSJ.
Sign up today and you will receive a free copy of our Future Focus 2026 report - the leading resource for IT decision-maker insight on priorities and investment areas in AI, security and more.
Microsoft nuclear power ambitions are gathering pace
In September 2023, Microsoft made its nuclear ambitions clear by posting a job description for a ‘principal program manager for nuclear technology’.
The program manager was sought to lead a technical investigation using small modular reactors (SMRs) to provide energy for Microsoft’s AI and cloud computing data centers.
SMRs can generate around one-third of the production capacity of conventional large nuclear reactors and their smaller footprint gives added flexibility for site selection.
Its modular design also means deployment is far simpler, where manufacturing facilities can fabricate SMR units to be shipped and installed across a wide range of locations, driving down installation and design costs.
With these advantages, SMRs present a tempting value proposition for reliable green energy production, but some SMR developers in the UK have called for planning rules to be relaxed as they don’t reflect the flexibility SMRs can bring.
RELATED RESOURCE
The enterprise’s guide for Generative AI
Make the most of the opportunities GenAI offers
DOWNLOAD NOW
Microsoft faces a similar challenge in the US, with the Vogtle Unit 3 nuclear power plant that came online in July 2023 being the only nuclear power project to open since 2016.
At the moment, NuScale Power is the only SMR developer to have had its design approved by the US’s Nuclear Regulatory Commission (NRC), and the project is reported to have cost around $500 million with a 12,000 page application and two million pages of supporting documentation.
Microsoft hopes its model trained on the NRC regulatory framework will assist in producing the application quickly and for a fraction of the cost.

Solomon Klappholz is a former staff writer for ITPro and ChannelPro. He has experience writing about the technologies that facilitate industrial manufacturing, which led to him developing a particular interest in cybersecurity, IT regulation, industrial infrastructure applications, and machine learning.
-
Amazon targets agent safety gains with investment in team behind Lean programming languageNews The tech giant hopes support for the open source programming language could drive AI agent safety improvements
-
At AMD Advancing AI, Helios was the star around which everything else revolvedOpinion The company's new rack-scale infrastructure is finally rolling off the production line, but there's a more petite offering to consider too
-
‘We are now seeing MAI models outperform general-purpose frontier models’: Microsoft CEO Satya Nadella touts in-house models to cut spiralling AI costs – and reduce growing reliance on frontier labsNews The Microsoft chief says pricey frontier models don't have to be used for every task, and its own in-house MAI models could be the key to reducing costs.
-
‘A company should be able to use a model without giving up the knowledge that makes it unique’: Microsoft CEO Satya Nadella says enterprises shouldn’t be sharing so much data with AI providersNews The Microsoft chief warned that corporate data could be at risk thanks to AI models
-
Copilot Cowork is now generally available: Everything you need to know, including pricing, usage limits, and new featuresNews A host of partner plugins are already available for Copilot Cowork, and more are coming
-
How Welsh councils are improving services with Microsoft CopilotCase study AI use has reduced staff toil, improved customer service, and increased team collaboration at three Welsh councils
-
‘LLMs are unreliable delegates’: Microsoft researchers say you probably shouldn’t trust AI with work documentsNews A research paper from Microsoft shows AI degrades documents over longer workflows
-
Microsoft joins competitors in handing over AI models for advanced testingNews US and UK government agencies will evaluate the firm's frontier models, along with those from Google and xAI
-
The AI operations gap is reshaping the Microsoft channelHow are AI advancements shaping the moves channel partners are making and need to make going forward?
-
Accenture has been trialling Microsoft Copilot since 2023 – now it’s rolling out the AI tool to all 743,000 staffNews Accenture will roll out Microsoft Copilot to nearly three quarters of a million employees after years of testing