The most interesting infrastructure bet of the week did not involve a large language model. It involved a pressure vessel. Apollo Atomics, a Y Combinator alumnus, is redesigning a core component of nuclear reactors that has been ignored for decades, with the goal of producing electricity below the cost of natural gas. This is the miniaturization thesis applied to atoms, and it is a much harder problem than applying it to chips.
The Overlooked Part as Business Strategy
Apollo's move is structurally identical to the best hardware startups of the last decade: find the component that everyone takes for granted, make it smaller or cheaper or both, and watch the entire system economics shift. The strategy has a long pedigree in Silicon Valley. What's different here is that the regulatory environment for nuclear is far more complex than for consumer electronics, and the failure modes are existential rather than merely expensive. TurboFund's seed-stage investor database tracks the growing cohort of deep-tech funds entering climate and energy hardware, a category that YC's Apollo bet suggests is heating up fast.
YC's Hardware Moment
The a16z story this week, about how AI gives foreign founders a structural advantage through the Borderless Founder network, is not obviously connected to nuclear miniaturization. But it points to the same underlying shift: the locus of infrastructure innovation is moving away from pure software and toward founders who bring domain expertise from outside the Silicon Valley monoculture. The nuclear engineer who grew up near a plant in Eastern Europe and the designer who reverse-engineered DJI's camera app are both working on the same problem: making closed, expensive systems open and cheap. Brewster Kahle has spent decades arguing that access to tools is a political act. Apollo Atomics is making that argument in megawatts.