Nuclear power needs AI to thrive

But if the U.S. is to meet its growing energy demands, the nuclear industry must embrace artificial intelligence and digital tools to transform how reactors are built — on time, on budget, and at the scale our future requires.

Nuclear’s around-the-clock energy generation provides the baseload power that AI-focused data centers require and does so without carbon emissions. It’s why Amazon, Google, Meta, and Microsoft have all announced nuclear initiatives in recent months and why approximately $80 billion was invested in nuclear power in 2024. However, just as AI needs nuclear power, the nuclear industry needs AI to solve its most pressing problems, especially when the U.S. government sets expectations for the industry. 

INSIDE THE EFFORTS TO RESTART A SHUT-DOWN NUCLEAR PLANT FOR THE FIRST TIME IN THE US

For nuclear to scale to meet America’s energy challenges, construction must become significantly faster and, for once, be on time and on budget. I was fortunate to lead the efforts to build units three and four at the Vogtle nuclear power plant, serving as the Project Director and Vice President of Operational Readiness during critical periods of the build. 

These are the only two new nuclear power plants built and operating in the United States in the last three decades, a project initially budgeted at $14 billion. The units ultimately cost more than twice that at $35 billion and took seven years longer than initially scheduled. And that’s not just an American issue. Britain’s Hinkley Point C is more than two years behind schedule and, at $59 billion, estimated to cost more than twice as much as anticipated.

The reasons for delays and added costs are not related to the technology of the reactors themselves — it is the way these projects are contracted, structured, managed and supported. 

Nuclear power plants are among humanity’s largest, most elaborate construction feats. Yet construction sites continue to rely on outdated, disconnected enterprise software tools that were not designed for projects of this scale and complexity. Projects frequently suffer from insufficient planning and coordination, fragmented supply chains, incomplete or immature designs, and the lengthy time required to meet complex regulations. To think that legacy software tools responsible for typical commercial construction projects can be optimally utilized is ludicrous. And that’s when software is even being used.

At recent nuclear projects, reams of excessive paperwork orders — often obsolete as soon as it is printed — were being pulled across the construction site in wagons to the work front. Workers were required to follow revisions and changes from one paper document to another, with the sheer volume of information and inefficiencies making cost increases, reduced productivity, and schedule delays inevitable.

AI and advanced digital tools offer us an entirely new way of approaching nuclear construction. It is one of the key missing ingredients that will make nuclear deployment not just possible, but repeatable, scalable, and efficient. Leveraging AI, digital twins, blockchain, and other technological advances that nuclear plants themselves are being built to support can keep new builds on time and on budget, and allow the industry to reach its full potential. 

For instance, in the past, construction teams required hours of preparation before work began in order to collect paperwork, visit tools rooms to review equipment, and locate the rest of their crews for the prerequisite pre-job briefs. AI-powered scheduling can provide workers with instantaneous, context-aware guidance — from the availability of certain parts and materials to the weather — that adapts to real-time constraints, ideally saving teams hours each morning.

Augmented reality can also streamline processes and increase productivity in the field. Construction workers can be supplied with augmented reality welding helmets that offer training solutions, such as how to ensure welds or job details are completed properly. In the future, we expect to see these technologies transfer to in-field solutions for weld accuracy, quality inspections, and real-time updating.

A blockchain-based supply chain makes supply chain management transparent, efficient, and trustworthy. Digital systems can track and verify all parts, preventing shipment errors, material shortages, and lost documentation. And when delays appear imminent, the system can initiate backup options immediately or prioritize other work in its place.

Sensors and other edge devices placed across a construction site can feed data in real time to a digital twin model of the site. This would allow project managers to track progress with precision, comparing what’s actually happening on-site to the original engineering plans. By using predictive analytics, teams can spot problems early — catching issues before they become expensive mistakes.

REPUBLICANS EMBRACE CLEAN ENERGY SOLUTIONS WITH OR WITHOUT DEMOCRATS

AI can even assist with regulatory compliance, turning a traditionally labor-and time-intensive task to a process that becomes nearly instantaneous. With today’s advanced technology, we can use large language models to rapidly review tens of thousands of documents and AI agents trained on the regulatory requirements to validate the data recorded automatically at the construction site.

American support for nuclear power has never been higher, but unless the industry commits to changing its track record of over-budget and behind-schedule construction, that support will quickly diminish. Technological innovations offer an unprecedented opportunity to reestablish American leadership in nuclear power, where China has built 37 reactors in the last decade, with 30 more under construction. It’s time to engineer a nuclear renaissance — one powered by the very technologies that will depend on our most complex energy source.

The Nuclear Company’s Chief Nuclear Officer Joe Klecha has more than 30 years of experience in both commercial and military nuclear operations and construction.