Analysis
Samsung C&T Engineering & Construction Group agreed to provide up to $100 million in combined equity and in-kind engineering support to Kairos Power, the advanced nuclear startup building a demonstration reactor for Google, according to TechCrunch and ExchangeMonitor. Roughly $70 million of the total is a direct equity investment, with the remainder structured as in-kind engineering services on the project itself.
What Kairos Is Actually Building
Kairos Power was founded in 2016 by Per Peterson and a team out of UC Berkeley's nuclear engineering department, after years of federally funded research into advanced reactor concepts. Its design runs at near atmospheric pressure using molten fluoride salt as coolant instead of water, with fuel that encases uranium kernels in ceramic and graphite pebbles engineered not to melt even at high reactor temperatures. The specific project Samsung is joining is a 50-megawatt demonstration reactor being built for Google, using fluoride-salt cooling and TRISO fuel, targeting completion by 2030 -- a milestone toward Google's broader goal of securing 500 megawatts of nuclear capacity by 2035.
“- TerraPower -- founded 2008 by Bill Gates, who remains among its largest backers; raised over $1 billion from investors including Gates, South Korea's SK Inc.”
Why An Engineering Firm, Not Just A Fund
Samsung C&T has built or helped build roughly a dozen nuclear reactors globally, which makes this a different kind of investment than a typical growth-equity check: Samsung is putting real construction expertise into a reactor design that has never been built at commercial scale anywhere. Kairos has also been awarded a Risk Reduction project under the Department of Energy's Advanced Reactor Demonstration Program worth $629 million total, with the DOE contributing $303 million and Kairos matching the rest from private sources -- federal derisking on top of Samsung's engineering commitment.
The Advanced Reactor Field Is Crowded And Capital-Intensive
Kairos is one of several advanced-reactor startups racing to be first to actually deliver power to a data center customer, all chasing the same underlying customer base: hyperscalers with enormous, firm-power demand for AI data centers and a willingness to underwrite first-of-a-kind construction risk that traditional utilities have historically avoided.
- TerraPower -- founded 2008 by Bill Gates, who remains among its largest backers; raised over $1 billion from investors including Gates, South Korea's SK Inc. and SK Innovation.
- X-Energy -- Rockville, Maryland-based; anchored a $500 million Series C-1 round in October 2024 led by Amazon.
- Oklo -- backed by OpenAI's Sam Altman; already public after a 2023 reverse merger with Altman's AltC SPAC.
Pulse has previously covered Google's other nuclear bet, a federal loan behind the revived Duane Arnold plant -- this Kairos deal is a second, distinct track toward the same 500-megawatt-by-2035 goal.
The Numbers In Context
$100 million is a modest figure next to TerraPower's billion-dollar cumulative raise, but it's meaningfully different in kind: this is a strategic construction partner putting capital and engineering hours directly into a specific, named reactor project with a real 2030 delivery date, not a diversified venture round. Google underwriting reactor-level construction risk -- rather than signing a power-purchase agreement with an already-operating plant -- is itself the bigger signal: Big Tech is now willing to fund first-of-a-kind nuclear construction directly to secure future capacity.
What The Announcement Doesn't Guarantee
A $100 million commitment and a dozen prior reactor builds don't eliminate the core risk: Kairos's molten-salt design has never operated at commercial scale anywhere, and every advanced reactor company chasing a Big Tech customer this cycle -- TerraPower, X-Energy, Oklo included -- is making some version of the same 2030-ish promise. Samsung's engineering hours reduce execution risk on construction specifically, but they don't derisk the underlying physics and regulatory approval timeline, which has slipped for nearly every advanced reactor project attempted in the US over the past two decades.
What Founders And GPs Should Watch
The test for Kairos isn't the $100 million, it's whether the 2030 delivery date holds -- advanced reactor projects have a long history of schedule slips, and Samsung's construction expertise reduces but doesn't eliminate that risk given the molten-salt design has no commercial operating precedent. Whether Kairos hits 2030, and whether Google's 500-megawatt-by-2035 target ends up leaning on Kairos, TerraPower, X-Energy or some combination, is the story to track as AI power demand keeps outrunning traditional grid buildout timelines.