It wasn’t long ago that the very mention of “nuclear” was enough to unsettle people across the ideological spectrum.
Between the meltdown of a nuclear reactor at Three Mile Island, explosions in Chernobyl, and an earthquake-powered disaster at Fukushima, the one thing people seemed to agree on was that nuclear was a dirty, if not dangerous, word.
But times have changed, new insights and new data have emerged, and what once seemed unethical has been rebranded as clean, in every sense.
The nuclear energy renaissance — fueled by Big Tech investment, a more favorable legislative climate, and shifting public opinion — hinges on a change in perspective: Nuclear energy has the potential to be one of the cleanest and most reliable forms of energy on Earth.
But the construction of nuclear reactors in the United States — where nuclear accounted for 19% of electricity generated in 2023, according to the US Energy Information Administration — is lagging compared to efforts in China and elsewhere, according to the Boston Consulting Group.
That’s part of why consulting firms are searching for ways for nuclear energy companies to cut costs, improve efficiencies, and remove the obstacles holding up progress.
Public sentiment is still a headwind
There are 94 nuclear reactors in the United States, according to the Nuclear Energy Institute. And tech giants are investing considerably to bring new ones online.
Last October, Amazon invested $500 million in X-Energy, a developer of small nuclear reactors and fuel. Around the same time, Google said it would purchase nuclear energy from Kairos Power, a California-based company developing small modular reactors. The month before, Constellation Energy struck a deal with Microsoft to provide the tech giant with nuclear power for the next two decades by resurrecting part of the Three Mile Island nuclear plant in Pennsylvania.
The Trump administration is also working to jumpstart nuclear reactor construction, building on Biden’s ADVANCE Act last year, which sought to remove regulatory hurdles. Trump issued four executive orders on Friday to accelerate the development of the domestic nuclear energy industry, touching on areas from fuel cycle development to reactor construction, testing, and licensing, to workforce training.
However, hurdles persist.
“Growth in nuclear power is projected to be almost flat to 2050 due to more stringent regulatory requirements than for other low-carbon energy sources, negative public perception, perceived safety issues, supply chain constraints, and uncertainty around waste disposal,” McKinsey & Company wrote in its 2024 Global Energy Perspective.
Some of the persisting negative sentiment can be attributed to concerns about how the radioactive waste that nuclear reactors generate will be stored. The federal government once planned to direct all waste to Yucca Mountain in Nevada, but lawmakers and the public objected.
There are also lingering concerns about the government’s capacity to mitigate nuclear disasters and nuclear power’s association with nuclear weapons proliferation.
There is hope, however. In 2024, three-quarters of respondents, out of a sample of 3.5 million people used by Bisconti Research, said they favored the use of nuclear energy for electricity, up from about half in the 1980s and 1990s.
The efficiency of construction
Large-scale nuclear reactors require a significant investment in time and money. They necessitate significant upfront capital and require long construction windows that are often beset by delays, cost issues, and regulatory hurdles.
BCG says that the “simpler the design, the better.” The firm said the optimal design should follow “a ‘design-for-manufacturing’ approach, which has delivered compelling results in other industries such as aerospace and defense.”
The goal, according to BCG, should be to standardize the components and material needed for construction, along with reducing the number of construction steps to promote “modularity,” the firm said. That applies to both large-scale and small-scale reactors, Benjamin Vannier, managing director and partner at BCG, told BI by email. The AP1000, a large-scale pressurized water reactor, is an example of a reactor designed for modular construction, he added.
Modularity is perhaps the biggest trend in nuclear right now. There are a number of small modular reactor companies, like Oklo, which was until recently chaired by OpenAI CEO Sam Altman; TerraPower, which is backed by Bill Gates; and X-Energy, which Amazon invested $500 million in last year. These more compact nuclear reactors are designed to be built in factories and then shipped to sites for installation and are, therefore, easier to standardize for production.
The economics of scaling
Nuclear fission, which produces energy by splitting uranium atoms, releases almost no greenhouse gas emissions. Many AI executives consider nuclear fission the only reliable way to power their data centers.
Tech leaders are also excited by recent advances in fusion technology, which involves combining two atomic nuclei to release energy and is considered safer than fission. Type One Energy, funded by Gates, published research in March that shows there are no scientific barriers left to making commercial fusion a reality.
However, nuclear power remains only a “medium-term solution,” McKinsey said in a 2024 report on data centers and power. “The timeline to scale nuclear so it can achieve rapid, repeated deployment is nearly a decade, while constraints on data center power are appearing today.”
It added, “The early economics of nuclear are challenging compared with other energy options, and implementing various technologies to try to reduce its costs may or may not work.”
While power plants are relatively cheap to run in the long term and can last as long as 60 years, they require considerable upfront investment.
In a report published in 2017, the World Nuclear Association said that nuclear plants are strongly influenced by capital cost, which accounts for at least 60% of their total levelized cost of electricity, which is essentially a measure of a plant’s economic efficiency. It is calculated by dividing the total cost to build and operate a power plant over its lifetime by the total electricity output.
Generative AI may help improve efficiencies on the operating side. Rafee Tarafdar, the chief technology officer of Infosys, a global consulting firm, told BI that the firm is helping companies integrate AI into their plans.
“One proof of value that we built is: How do we use all the logs, all the sensors that come from all these machines in order to triage, predict failures, and help resolve issues much earlier,” he said, in reference to the work the firm did with a large US energy company.
The consensus among consulting firms is that solving early construction bottlenecks is key to scaling nuclear energy as a viable power source for AI and beyond.
Read the full article here
