Kairos Power has emerged as one of the most closely watched US advanced nuclear developers, backed by Big Tech and pushing firsts with its unconventional design.
Having pulled in $643m from investors including Google Ventures, the US Energy Department and Samsung C&T, the Hermes demonstration reactor is set to become the world’s first operational pebble-bed, fluoride salt-cooled high-temperature reactor (FHR). It was also the first US Gen IV reactor to receive a construction permit from the Nuclear Regulatory Commission (NRC).
Go deeper with GlobalData
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
In April, the company broke ground on the Hermes 2 Demonstration Plant, marking the first deployment under Kairos’ landmark deal with Google to develop an advanced reactor fleet to help decarbonise data centres in Tennessee and Alabama.
Edward Blanford, co-founder and CTO of Kairos Power, tells Power Technology about the benefits and trade-offs of its unique design choice, supply chain strategies and the path to commercial deployment.
Jackie Park (JP): Kairos Power has opted for a FHR design, which uses FLiBe (molten-salt coolant made of lithium fluoride and beryllium fluoride) instead of pressurised water, and TRISO fuel in pebble form. What inspired this design choice?
Edward Blanford (EB): Historically, people have looked at advanced reactors through the lens of the past, with much of the work done at national labs. They have thought about molten salt reactors as having fuel dissolved within them – but as one of our co-founders, Per Peterson, and a few others further developed the technology, we found that combining the solid fuel with the fluoride salt coolant offers a unique safety case in which all the [dangerous] fission products are contained within the fuel itself. This creates an opportunity to build a safe advanced technology that can operate at extreme temperatures. Being able to keep the integrity of the vessel with the fuel and the coolant is what is really needed to ensure the integrity of a plant.
Limiting the safety risks also opens up opportunities to improve the economics. Because the primary reactor system is safe, we don’t need to over-engineer everything else surrounding it. Rather than building large containment structures that are needed for high-pressure systems, our plants are more like a chemical plant that operates at lower pressure. By taking advantage of that safety case downstream, we drive down the costs of the building and design structures and leverage more conventional construction techniques.
JP: Apart from these cost and safety benefits, are there any trade-offs in using FLiBe as the primary coolant that influence plant efficiency and scaling?
EB: One important trade-off to recognise is that testing for components, materials, behaviour, qualification and so on requires a clean coolant, not a coolant with fuel in it. That means that, if we want to measure the equivalent of the steam table for FLiBe and get thermophysical properties, we have to do it without fuel dissolved in it. It is a unique requirement upfront that needs specific facilities capable of performing that type of testing and qualification work, which can be challenging to find and quite expensive.
As for some other challenges, nuclear FLiBe is enriched in lithium-7 [Li-7] for use in the Hermes reactor and beyond, which adds cost. Kairos has invested significantly in internally developing Li-7 enrichment capabilities and strengthening the FLiBe supply chain.
In terms of the thermophysical characteristics of the coolant itself, FLiBe offers a range of benefits, including from a chemical volatility standpoint; it is not reactive like steam. However, melting at a high temperature, around 459°C, creates some challenges. Designing systems that can operate at those temperatures can be difficult.
Having said that, there is a large window between when FLiBe melts and when it boils, and with it comes a huge opportunity to drive these systems to higher temperatures. Long-term, that is one of the main drivers behind FLiBe.
Industry is excited about these types of molten salts’ volumetric heat capacity and the ability to move large amounts of heat with very compact systems. That is why it was one of the coolants originally selected for aircraft reactor experiments. We see more opportunities than challenges in further developing these capabilities.
JP: Speaking of the FLiBe supply chain, the US has been ramping up efforts to develop a domestic enriched lithium supply chain for advanced nuclear technologies. What key bottlenecks have you identified, and how is Kairos navigating the geopolitical and scaling risks of the supply chain?
EB: FLiBe for our purposes requires even higher purity levels of Li-7 enrichment. We require 99.995%, whereas the current fleet of pressurised water reactors (PWRs) that operate in the US utilise enriched lithium hydroxide, also enriched in Li-7, but at 99.95%. This enriched lithium is critical for the security and long-term operations of these PWRs, as well as the new fleet of AP-1000s that the government is considering, and the supply has typically come from China or Russia.
The Electric Power Research Institute and other organisations are looking into means of replacing this supply, but at the end of the day, we need access to this material one way or another for US fleets. To ensure that Kairos does not have to depend on international supply, or wait for other domestic developments, we are actively contributing to building out US lithium enrichment capabilities. That process is a trade secret, but we are developing technology in Albuquerque, New Mexico, that builds on years of research, development and testing.
Over time, as Kairos goes into mass commercial deployment, we will need more beryllium to supplement demand – but eventually, I see our facilities servicing not only light water reactor (LWR) customers but also lithium-6 and lithium-7 for both fusion and fission technologies at different levels of enrichment.
JP: What are support mechanisms that could help secure beryllium supply?
EB: Both beryllium and lithium are critical minerals, and they have been identified by this and prior administrations as critical for not just advanced nuclear applications, but the country’s broader needs, including manufacturing and national security. Further policy support to ensure sufficient beryllium supplies is crucial across industries. This could involve supporting domestic beryllium fluoride producers, like one of our partners, Materion. Today, they supply our non-nuclear programme, but going forward, we want to collaborate with them on nuclear applications.
JP: What are some key technical learnings so far working with the Oak Ridge National Laboratory?
EB: Much of our work with Oak Ridge has focused on developing means and methods to construct and deploy the technology more quickly, rather than the technology itself. The unique safety case of our technology enables different civil structural strategies for buildings that look more like chemical facilities than expensive LWR structures. We are working with Oak Ridge to leverage methods that support that type of prefabricated modular concrete construction.
One method we are exploring is using 3D printing to investigate novel concrete shapes and forms, which enable a more compact design and quicker implementation. We are looking into a precast construction method, where you essentially build Lego structures that then come together in final form. Making those structures come together faster, and the erection process more efficient, will reduce the overall construction timeline and labour. Oak Ridge has been very helpful in demonstrating some of those advanced methods prior to executing them in the field with the NRC.
JP: With demonstration milestones approaching, what are the biggest remaining technical challenges in moving from prototype to commercial deployment? Could they affect the timeline, and how are you managing risks to stay on schedule?
EB: Right now, Kairos is focused on Hermes 2. Standardisation has been key to the architectural and construction decisions we have made for Hermes 2, our first power-producing plant. Building on lessons learned from Hermes 1 (low-power demonstration reactor), we have designed Hermes 2 to be a more standardised design that can be reproducible downstream for widespread commercial deployment.
Getting to market quickly is important, but hiding risk doesn’t get rid of it. We have decided to do more testing and development, and some might think that is because our technology is immature – that is not true. Rather, we are dedicated to minimising risk down the line. There are plenty of technologies not fully tested that have entered the market, with risks buried within the design and deployment strategy. Then these technologies, including mature and proven LWRs, have created challenges for utilities with even simple component replacements.
In general, our industry needs to have more humility. The commitment to testing isn’t just about retiring technology risk but also standing up capabilities.
JP: When do you expect commercialisation, and more broadly, when might your technology become a popular design choice across the industry?
EB: Right now, we are looking at a 30–34-month (as of October 2026) commercial construction timeline. We have broken ground on Hermes 2, and we are targeting 2030 for a commercial operation date, pursuant to our Google deal.
Following that, we see opportunities to build multiple units. In terms of siting and deployment, those decisions will happen in the next year or two. With Hermes 2, we will see a commercial uptick under our Master Plant Development Agreement with Google.
It is important to be thoughtful, however. You can have a very large order book on the back end, but if you can’t deliver on the first plant, then that is all for naught. In the past, there was a huge order book following expectations for a huge LWR renaissance in the early 2000s, but the industry didn’t get past the first few projects. We are trying to avoid that and focus on having a credible path to deploying the first commercial unit. Subsequent units can be like a hockey stick in terms of the inflexion point.
