Samsung Backs Kairos Power's Reactor Build for Google with $100 Million Stake

A Strategic Bet on Advanced Nuclear

Samsung C&T will invest up to $100 million in Kairos Power and supply engineering resources to help the nuclear startup deliver its 50-megawatt demonstration reactor, according to Kairos Power. The arrangement splits into $70 million in equity and the remainder in engineering services, positioning Samsung as both a financial backer and construction partner for a project that must reach operation by 2030.

Samsung C&T brings experience from roughly a dozen nuclear builds globally, a track record that matters when timelines are compressed. Kairos needs the first commercial-scale unit operational within five years, and subsequent reactors will need to arrive even faster to satisfy the full scope of its agreement with Google, which calls for approximately 500 megawatts by 2035.

At Opentechwire, we've tracked the capital flows into advanced nuclear since late 2024, when data centre operators began signing power purchase agreements with startups promising reactors that could sidestep the decade-long construction cycles of conventional gigawatt plants. The Samsung investment is the clearest signal yet that established industrial players see the bottleneck shifting from financing to execution, and they want a foothold before the sector scales.

Two Reactors in Tennessee, Two Different Roles

Kairos is constructing two units at Oak Ridge, Tennessee. Hermes 1 is a low-power test bed designed to validate the fluoride salt-cooled high-temperature reactor design and refine operational procedures. Hermes 2 is the commercial-scale demonstration plant, rated at 50 megawatts, and its output will count as the first tranche of power delivered under the Google deal.

The Nuclear Regulatory Commission granted construction approval in late 2024, clearing the path for physical work to begin. The 2030 target for Hermes 2 aligns with the original timeline Google set when it signed the deal in autumn 2024, making this one of the tighter schedules in the current wave of advanced reactor projects.

Fluoride salts have boiling points well above the operating temperatures of the reactor core, which keeps internal pressure low. That characteristic reduces the likelihood of high-pressure failures and simplifies containment design. The reactor will also use TRISO fuel, a particle form in which uranium kernels are coated in layers of carbon and ceramic, then packed into spheres roughly the size of billiard balls. The layered structure is intended to contain fission products even under accident conditions, eliminating the potential for core melt scenarios.

Why Google Needs Half a Gigawatt by 2035

Google's power demand is rising faster than grid capacity in key data centre markets, driven by the computational requirements of large language model training and inference workloads. The company committed to Kairos for around 500 megawatts by the mid-2030s, a volume that implies at least ten additional reactors beyond Hermes 2 if the design remains at 50 megawatts per unit.

That pace, roughly two reactors per year after 2030, is aggressive by nuclear industry standards but closer to the cadence Samsung C&T achieved in its earlier reactor programmes in South Korea and the United Arab Emirates. Modular construction and factory fabrication of reactor components can compress site work, but regulatory approvals, supply chain coordination and grid interconnection still impose lead times that few startups have managed at scale.

The equity stake gives Samsung exposure to Kairos' intellectual property and a potential template for future projects in Asia, where several governments have restarted nuclear programmes to meet decarbonisation targets. South Korea's own energy policy has swung back toward nuclear expansion under recent administrations, and Samsung C&T is positioned to bid on domestic builds as well as export projects.

Engineering Services as Strategic Entry

The in-kind engineering contribution, valued at $30 million, covers design support, procurement coordination and construction management. For Kairos, this arrangement offloads some of the project risk and taps Samsung's experience in managing complex supply chains for nuclear-grade components. For Samsung, it provides early visibility into Kairos' design choices and operational constraints, information that becomes more valuable if the fluoride salt-cooled architecture proves commercially viable.

Advanced reactor startups have struggled to move from regulatory approval to physical construction because the vendor base for specialised components is narrow and lead times are long. Pressure vessels, heat exchangers and control systems for non-light-water reactors often require custom fabrication, and few suppliers have capacity to serve multiple projects simultaneously. Samsung's involvement could help Kairos secure priority in procurement queues and negotiate better terms with manufacturers.

The partnership also signals a shift in how capital is entering the sector. Early-stage nuclear ventures relied on venture funding and government grants, but moving to commercial scale requires construction finance and performance guarantees that traditional investors are reluctant to underwrite. Strategic investors with engineering capabilities can de-risk projects by taking on execution responsibilities directly, a model that may become more common as other startups approach first-of-a-kind builds.

A Compressed Timeline in a Slow Industry

Five years from construction start to commercial operation is fast for nuclear, even for a 50-megawatt unit. Conventional pressurised water reactors in the United States have averaged more than a decade from groundbreaking to grid connection, with cost overruns driven by design changes, regulatory delays and workforce shortages. Kairos is betting that a simpler design, factory-fabricated components and a smaller physical footprint will cut that timeline in half.

The startup received its construction permit under the Nuclear Regulatory Commission's Part 50 licensing process, which requires detailed design documentation upfront. That front-loaded regulatory work can accelerate construction once approval is granted, but it also means any design changes during the build phase trigger additional review cycles. Kairos will need to hold its design stable and avoid the scope creep that has plagued larger projects.

Google's commitment provides revenue certainty, but it also locks Kairos into a delivery schedule with limited flexibility. Missing the 2030 target for Hermes 2 would delay the entire programme and potentially trigger contract penalties. Samsung's equity stake aligns incentives, since delays would erode the value of its investment, but construction risk remains high for a technology that has never operated at this scale.

What Comes After Hermes

If Hermes 2 performs as designed, Kairos will need to replicate the build process at least ten more times to meet the Google contract. That requires not just engineering repeatability but also site identification, interconnection agreements and workforce scaling in markets where nuclear construction skills are scarce.

The fluoride salt-cooled design has advantages in siting flexibility because the low-pressure operation and passive safety features reduce the required exclusion zone around the plant. That could allow Kairos to place reactors closer to load centres, cutting transmission costs and improving grid resilience. Data centre operators prefer on-site or near-site generation to avoid congestion charges and reduce exposure to wholesale power price volatility, making smaller, distributed reactors more attractive than remote gigawatt plants.

Samsung's participation suggests the model may extend beyond the United States. South Korea, Japan and several Southeast Asian countries are evaluating advanced reactors as part of decarbonisation strategies, and Samsung C&T is well positioned to bid on those projects if it gains experience with Kairos' technology in Tennessee. The equity stake gives Samsung a claim on intellectual property and potentially a licensing path for international deployments.

The next eighteen months will clarify whether Kairos can hold its schedule. Hermes 1, the low-power demonstrator, is expected to begin testing in late 2027 or early 2028, providing the first real-world data on how fluoride salts and TRISO fuel perform together in a working reactor. If that test phase uncovers issues requiring design changes, the 2030 timeline for Hermes 2 becomes difficult to defend. Samsung's engineering team will be watching those results closely, and so will the rest of the advanced nuclear sector.