Apollo Atomics bets nuclear's cost problem is the steam generator, not the reactor

Apollo Atomics has closed a $26 million seed round, $5 million of it debt, to build a demonstration reactor ahead of a commercial deployment planned for 2028 ^[21][1]. The reason to look closely is that the company is not selling a new reactor: CEO Assil Halimi told TechCrunch the target is the steam generator, the component that converts heat from the reactor coolant loop into the steam that drives a turbine ^[2][3].

That is a different wager from the one the rest of the field has placed. TechCrunch reports that startups have been raising money to develop advanced reactors, and that recent projections make it unlikely those reactors will be cost-competitive soon ^[4]. Halimi and co-founder Drew Walker argue those companies have been chasing the wrong problem ^[5].

The complaint is about manufacturing, not physics. Most steam generators are hand-built and several stories tall ^[6]. Apollo says its version can be mass-manufactured and is about the size of a person ^[7]. Conventional designs run coolant through a series of large tubes surrounded by the water that becomes steam; Apollo threads both loops through a compact metal block laced with needle-thin channels, which the company says moves heat more efficiently and so allows the shrink ^[8][9]. Halimi adds that existing generator designs borrow heavily from coal and gas plants, whose fuels have far lower energy density than uranium, so a nuclear plant converts only a small piece of what its fuel offers ^[10].

The payoff Apollo claims is a reactor 40 times smaller than one built around a conventional steam generator ^[11]. TechCrunch does not state which plant that ratio is measured against, and no third party has verified it, so treat it as a company figure rather than a benchmark. The same is true of the downstream numbers: factory assembly of reactor and generator instead of on-site construction ^[12], a 300-megawatt plant in under 24 months ^[13], a reactor costing four to five times less to produce than existing designs ^[23], and electricity at 3 cents per kilowatt hour, which Halimi frames as beating natural gas rather than improving on nuclear ^[24][25].

What exists today is smaller. Apollo has built a 40-kilowatt reactor at MIT to demonstrate the technology, and its design is based on MIT research ^[14][15]. The planned product line runs from 10 megawatts-electric to 300 megawatts ^[16]. The distance between the demonstration and the top of that range is a factor of 7,500 in thermal output terms ^[20], which is where a manufacturability bet either survives or does not. Compact channel heat exchangers are attractive precisely because they are small; small channels are also where materials behaviour, fabrication tolerance and inspection get harder, and none of that is settled by a 40-kilowatt unit.

FCVC led the round, with Telesoft Partners, Y Combinator, Alumni Ventures, Robinhood Ventures, Nucleation Capital, Pelion VC, Duke Capital Partners, New Era Ventures, Orange Collective, Stanford University, E14 Fund and Neutron Power Ventures participating ^[17]. Apollo was in Y Combinator's Spring 2026 batch ^[18]. Stripping out the debt leaves roughly $21 million of equity to reach a demonstration reactor ^[22], which is thin for a nuclear programme and implies another raise before 2028.

Three things to watch. Whether Apollo publishes the baseline behind the 40x claim, since a compactness ratio without a reference plant is marketing. Whether the demonstration unit is built at a scale that tests the channel block under commercial heat flux rather than at 40 kilowatts. And whether the 2028 date survives contact with licensing, which the company's account to TechCrunch does not address at all ^[19].