“Make no little plans. They have no magic to stir men’s blood and probably will not themselves be realized. Make big plans, aim high in hope and work, remembering that a noble, logical diagram once recorded will never die, but long after we are gone will be a living thing, asserting itself with ever growing insistency.”
- Burnham, Daniel H., and Edward H. Bennett. Plan of Chicago. 1909, 40–42.
Burnham’s rhetoric was intended to inspire civic infrastructure improvements in the first decade of the 20th century. A century later and more two nuclear developers have released plans for massive investments in nuclear power plants to be build in fleet mode. They face a gauntlet of financial, engineering, logistical and operational challenges at a scale that goes far beyond what has been considered so far as feasible.
The Nuclear Company and Oklo in aggregate have released plans to build a total of 18 GWe of new nuclear generating capacity. While the Nuclear Company has indicated its focus is new nuclear for mainstream power utilities, Oklo has identified a single customer which builds power hungry data centers.
Nuclear Company RFI for PWRs
First, the Nuclear Company, which is planning on deploying 6 GWe of fleet-scale nuclear power across America, announced this week it has released a solicitation of interest to contractors in engineering, procurement and construction to support its first fleet. The firm notes that contractors of all sizes and capabilities are encouraged to request invitations to the formal process.
The Nuclear Company says in its press statement that it is ‘committed to redefining nuclear construction by adopting fleet-scale deployment, which employs a design-once, build-many approach.”
Accord to the release, the firm says the approach focuses on streamlining processes, utilizing advanced construction and manufacturing methods, and employing digital technologies. These technologies will integrate advanced analytics, AI and IoT to streamline site selection, permitting, construction management, regulatory management, engineering processes and supply chain logistics.
By utilizing this approach, The Nuclear Company aims to achieve significant cost reductions and shortened project timelines, meeting its commitment to on-time, on-budget deployment of its first fleet of 6 GWe of nuclear power. Beyond that, the design-once, build-many approach combined with digital technologies ensures an empowered workforce, data-driven decision-making, improved safety and operational excellence.
Oklo’s Advanced SMRs for Data Centers
Second, Oklo, a developer of a micro reactor which now has power ratings in the small modular reactor class, e.g., 50 MW, announced Oklo and Switch, a data center developer, have signed one of the largest corporate power agreements in history, a 12 GWe non-binding Master Power Agreement to deploy 12 GWe of Oklo Aurora powerhouse projects through 2044.
Under the Master Agreement, Oklo will develop, construct, and operate powerhouses to provide power to Switch across the United States through a series of power purchase agreements. This relationship over several decades will help accelerate Oklo’s early powerhouse deployments and also position the Company to scale in response to a growing demand pipeline. This Master Agreement enables Oklo to leverage Switch’s execution while developing the financial and infrastructure model for scaling advanced nuclear.
The timescale of this Master Agreement underscores Switch and Oklo’s commitment to meeting artificial intelligence’s growing electricity demands with clean, sustainable power. The company believes that Switch’s industry-leading data centers, which support the computing needs of some of the world’s largest companies, are positioned to handle AI workloads well into the future.
Roadmap of Challenges Ahead
The largest single focus nuclear project developments in within the past 20 years have been four 1,400 MW PWRs in the UAE, which took 14 years for South Korean firms to build from the time the first shovel of earth was moved until the fourth reactors was commissioned and in revenue service on the grid. In Turkey Russia is building four 1,200 MW VVERs with each one taking about eight years to from start to finish.
In China, the country has 58 nuclear reactors in revenue service providing 59 GWe of electric generation capacity. The country has another 29 reactors under construction. By the time China finishes the 29 reactors it has under construction, the county will have almost 90 GWe of nuclear power in place.
China has been able to fund these massive commitments to nuclear new builds by using state owned enterprises who’s budget authority is so huge that, and other major infrastructure commitments, like high speed rail, have warped the nation’s economy at the expense of consumer spending and a social welfare net for its aging population.
Short List of Major Challenges
In the US the challenges ahead for both the Nuclear Company and Oklo are as follows;
Raising capital through public private partnerships is on the order of tens of billions of dollars. For instance, using a benchmark of $5,000/Kw, the 6 GWe the Nuclear Company wants to build, priced in today’s dollars, comes to a staggering $30 billion.
With regard to Oklo’s plans for 12 GW of power to be supplied to a single class of customers, e.g., data centers, there are no reliable benchmarks in terms of dollars for kilowatts for advanced reactors because no one has built one, especially Oklo’s unique design, for commercial uses.
Assuming Oklo can raise capital to build its reactors at this scale, it is plausible to assume it will seek to see additional revenue from that scale of investment by selling its reactors to other customers for other uses. It is also likely that with Oklo’s plan to build 12 GW worth of reactors, that it has plans for bigger designs than the current 50 MW version it is currently seeking the license at the NRC.
Getting regulatory approvals for six 1,000 MW nuclear reactors, even spaced out over a two to three decade period, will severely tax the ability of the NRC to review reactor plans and issue licenses.
According to the Nuclear Company, it will look at AP1000 sites that have COLs but which are not expected to be developed by their respective utilities in the near term. These include six COLs held by Duke (two each at Williams States Lee, Harris, Levy County), and two COL held by Florida Power & Light (Turkey Point). The startup also has its eye on Early Site Permits (ESP), which have 20 year shelf lives but some of these ESPs are technology neutral or for other reactor designs.
It is unclear whether the Nuclear Company plans to co-develop these sites, acquire them, or create other kinds of collaboration to build reactors. In all the startup expressed a goal of building 6 GW of power or the equivalent, more or less, of six Westinghouse AP1000s.
A plan to build 6 GW of PWRs by the Nuclear Company and the number of Oklo SMRs that are needed to deploy at 12 GW by 2040, will place enormous strains on supply chains which will have their own capital financial needs to ramp up to meet this level of demand.
Workforce requirements are similarly daunting and it is going to be problematic whether the necessary engineering and skill trades levels of staffing can be trained and deployed within the time scales proposed by both firms. The current EPCs capable of building large or small reactors would undoubtedly leap at the chance to work at this scale, but the project management and controls capabilities of these firms would have to scale accordingly.
In terms of fuel, while the PWRs will have a less challenging time getting LEU fuel, Oklo will have to build its own fuel fabrication plant, and secure the necessary feedstock from it.
So far the history of building new nuclear power plants in the US, UK, and France, has not been cost competitive results. The schedule delays and cost overruns for these plants are well known. At least for now US utilities have shown no interest in committing to “fleet mode” construction campaigns much less new single large units. The collapse of NuScale’s SMR project in Idaho, due to inflationary factors far beyond its control, chilled the interest of some publicly traded utilities smaller reactors.
IT Platforms Want to Buy Power, Won’t Pay to Build Plants
While some of the big IT platforms have recently shown interest in advanced reactors to power their data centers, the amounts of money they have committed to put their toes in the water won’t build even one SMR much less a fleet of them. Amazon, Google, and Microsoft have all made it clear they are not going to pay to build reactors. Their expectations are to offer favorable terms in power purchase agreements to get to the head of the line for access for any new power generation capacity.
Oil Companies are a Competitive Threat to Plans for Nuclear Fleets
More recently, Exxon and Chevron has have announced they plan to build natural gas power plants adjacent to data centers, bypassing delays in authorizing new grid connections. The implications of a report in the New York Times for 12/11/24 is that the outlook for nuclear energy to power data centers is now facing competition from oil companies. Given that the time to deliver a gas fired power plant is less than half that of an SMR, and comes without the regulatory overhead of nuclear licensing, data centers grasping for any reliable source of power will likely flock like pigeons chasing popcorn on a city street corner to sign up for them.
The earliest dates small modular reactors, or micro reactors, will be available to customers to power data centers is the end of this decade, or more likely, in the start of 2030 or even later. Data centers need power now.
Oklo, and any other advanced reactor developer with a time to market of more than five years from today, which has targeted data centers as their customers, may see some of their initial market share eroded by the fast track construction of gas fired generation capacity for data centers. Over the long term, as decarbonization of the industry becomes a more significant priority due to climate change, nuclear solutions for data centers will likely get more traction with data centers as well as with other industries such a replacing coal-fired power plants.
DOE’s brightest stars in the world of advanced reactors, TerraPower and X-Energy, both have start dates for their respective unique reactor designs in the early 2030s and will face competitive pressures from the oil company initiatives assuming they can delivery at scale.
Can the TVA Model Be Extended to the Nation?
In terms of building nuclear reactors, large and small, in the US in fleet mode, the US government could create a national quasi private corporation like TVA to support private sector building of reactors in fleet mode by being a source of financial support through loans, stimulation of production of nuclear fuel for light water and advanced reactors, and coordination of streamlined regulatory review and oversight processes. Whether Congress or the public would accept creation of such an entity with a mandate this broad and complex is unknown.
The gas companies won’t have their early time to market as a competitive advantage forever. Back in December 2010, John Grossenbacher, then Director of the Idaho National Laboratory, told this blog in an interview, that there is a need to look beyond the headlines in three areas.
- Developing public/private partnerships to build reactors
- Finding ways to assure financing and certainty in the marketplace for utilities and investors
- Rebuilding America’s nuclear energy infrastructure to manufacture components
Grossenbacher said in a comment 14 years ago that seems to predict today’s headlines, “if you just let market forces drive energy policy, you will get natural gas plants being built as long as it is plentiful and cheap, but it may not always be that way.”
He said that the nation needs to look beyond market mechanisms to address two key goals (1) energy security, and (2) environmental stewardship.
Grossenbacher, who had a long, distinguished career in the nuclear navy before taking his civilian post in Idaho in 2005, knows a thing or two about building reactors. He said that to think usefully about nuclear energy, you have to work on a timeline of 50-100 years.
A full size reactor built today has a plausible service life of at least 80 years. Add in the time to build it, e.g., five-to-eight years, and the decommissioning phase after the end of its service life, and you hit the century mark without breathing hard.
To look for an comparable level of commitment for major infrastructure in US history, In 1956 then President Eisenhower initially allocated $26 billion for the project, building the Interstate System ended up being $425 billion and took 35 years to complete. However, over the years, many states have added to the system by constructing national highways and beltways around cities to reduce congestion and traffic.
If nothing else, if both firms gain customers with their plans, a time lime of three decades, or more, is as likely a scenario for success as any.
# # #
