Energy

Europe Cannot Buy the Reactor America Just Funded

Sequoia led a 1 billion dollar round valuing Valar Atomics at 6 billion. Its reactor first made 10 kilowatts. The European Commission still targets the early 2030s for the first SMRs here.

EnergyBy Servola Tech Desk2026-08-034 min read

AI-assisted, edited by humans. | Editor-in-Chief: Leon Soliman | Editorial standards

Key takeaways

Sequoia has led a 1 billion dollar Series B valuing Valar Atomics at 6 billion dollars, three times its valuation of a few months ago, and the company has added a 200 million dollar credit facility. The reactor behind that price first produced 10 kilowatts of thermal power in June.
The demonstrated output and the announced 30 megawatt data centre are about three orders of magnitude apart, and that comparison flatters the smaller number because it sets thermal kilowatts against electrical megawatts. The round is priced on closing that gap in a factory.
What the money really bought is a licensing position. Ward 250 is the first reactor authorised by the US Department of Energy to be built and operated outside the national laboratory system, under a pilot programme that has no European equivalent.
The European Commission published its own small modular reactor strategy on 10 March 2026 and targets first deployments in the early 2030s. For any European operator planning capacity before then, behind-the-meter fission is not a procurement option at any valuation.
Plan the next two capacity cycles on grid connection and contracted supply, not on private nuclear. Treat an SMR line in a 2027 or 2028 site plan as a research item rather than a commitment, and keep your interconnection application moving regardless.

Ten kilowatts in Utah, six billion dollars in the round

The US Department of Energy's Office of Nuclear Energy confirmed that Valar Atomics' Ward 250 reached zero-power fuelled criticality at about half past four in the afternoon, mountain time, on 18 June 2026, at the San Rafael Energy Lab in Emery County, Utah. Days later the Generation IV high-temperature gas reactor produced roughly 10 kilowatts of thermal power and ran for more than twenty-four hours continuously. In July the company used fission to generate enough electricity to run an Nvidia AI chip and serve a website, which is a genuine first and a very small amount of power.

On 3 August Sequoia Capital led a 1 billion dollar Series B that values Valar at 6 billion dollars including the new money, roughly three times the valuation it carried a few months earlier. Atreides Management, Point72 and Snowpoint Ventures joined, the company also lined up a 200 million dollar credit facility, and Sequoia partner Shaun Maguire takes a board seat. Isaiah Taylor, Valar's founder, framed the purpose plainly: one reactor can be built as a project, but a fleet has to be manufactured. The round is the money to build the factory, not evidence that the factory works.

The asset being priced is an authorisation, not a turbine

Ward 250 is the second advanced reactor to go critical under the Department of Energy's Reactor Pilot Program, and the first reactor authorised under it to be built and operated outside the national laboratory system. That second fact is the one that carries commercial weight. The pilot programme exists to compress the path from design to a working first-of-a-kind unit by using DOE authorisation rather than the conventional civilian licensing route, and Valar is the first private company to have taken a complete integrated reactor through it on its own site.

Energy Secretary Chris Wright placed the milestone in a sequence that runs from airlifting a small reactor aboard a military transport to zero-power criticality testing, which tells you what the programme is optimised for. Speed of authorisation is the product. A European investor reading the round as a bet on high-temperature gas reactors has read it too narrowly. The bet is that a national government will keep granting an accelerated path, and that the factory can be built before that path closes or a competitor is granted the same one.

Brussels published its own timetable, and it says early 2030s

The European Commission published its strategy for the development and deployment of small modular reactors on 10 March 2026, building on the European Industrial Alliance on SMRs it launched in February 2024. The strategy is explicit that these units could be sited at industrial locations, ports and high-demand data centres, and that they could relieve grids under strain from data centres, low-carbon hydrogen and desalination. It is equally explicit about timing: first European deployments are targeted for the early 2030s.

Commercial interest is already visible against that timetable rather than ahead of it. Westinghouse and Data4 have signed a memorandum of understanding to examine AP300 units for future European data centres, with Data4 naming the design as its technology of choice for evaluation. A memorandum of understanding to evaluate a reactor is a long way from a connection date. For an operator sizing a site for 2028, the European SMR question is not which vendor to choose but whether to include the option at all, and on the Commission's own published timetable the answer is no.

What belongs in a capacity plan this year

The useful discipline here is to separate the two things the announcement bundles together. A reactor achieving criticality on a private site under a government pilot is a real engineering and regulatory milestone. A 30 megawatt facility with a named partner is a stated intention. Between 10 thermal kilowatts and 30 electrical megawatts sit roughly three orders of magnitude, a manufacturing line that does not exist yet, and a licensing regime that applies on one continent.

So keep the interconnection application moving, because it remains the only route with a date attached, and price the queue honestly including the cost of waiting. Where an SMR appears in a site plan for 2027 or 2028, mark it as a watch item with a named trigger, such as a European vendor securing a construction licence rather than a memorandum. Ask any vendor citing Valar's milestone which regulator authorised it and whether that regulator has jurisdiction where you intend to build. The answer will usually end the conversation, and it is cheaper to have that exchange now than after a site has been chosen on the assumption of private power.

Frequently asked questions

Could a European data centre buy a Valar reactor today?

No. The company is moving from demonstration to factory production, and its authorisation to build and operate came from the US Department of Energy under a domestic pilot programme. Deployment in Europe would require licensing by the relevant national regulator under European rules, which is a separate process with its own timetable.

Does the Nvidia partnership change the timeline?

It sets a target, not a date you can plan around. The two companies are working towards a 30 megawatt, nearly waterless facility in Utah pairing the Ward 250 design with Nvidia's water-efficient cooling. That is the first commercial application of the partnership and remains at the planning stage, several orders of magnitude above what the reactor has demonstrated.

Is it worth talking to SMR vendors at all right now?

Yes, but as market intelligence rather than procurement. Early conversations tell you which designs are progressing through European licensing and what site conditions a vendor will eventually require, which is genuinely useful for long-range land and grid decisions. Just keep them out of the budget line until a vendor holds a construction licence in a jurisdiction where you operate.

The reporting settled on the valuation, because tripling to 6 billion dollars in a few months is the better story. The document that matters is the Department of Energy's confirmation that a privately built reactor went critical outside the national laboratories, and the Commission strategy that puts Europe's first units in the early 2030s. One of those is a licensing event on another continent and the other is a timetable that governs your site. Plan against the timetable.

Valar AtomicsSequoiaSmall modular reactorsNvidiaData centre powerEuropean CommissionGrid connection

Antora Raised 550 Million To Sell Heat, Not Power

Antora closed a 550 million dollar Series C on 30 July. Its flagship customer is an ethanol plant, and that says more than the data centre headlines do.

6 min read

Energy

Three 20 Megawatt Halls Count as One Large Load

PJM's board directed a FERC filing on 27 July. Loads of 50 MW or more within a one-mile radius are aggregated, and new ones without their own capacity get cut first.

6 min read

Law & Regulation

The EA Buyout Now Hinges on Brussels

The largest take-private in history has cleared EA shareholders and US antitrust. The last real gate is the EU - and not classic merger review, but the Foreign Subsidies Regulation, facing its first blockbuster test.

3 min read

← All articles