Starcloud, the fastest unicorn in Y Combinator history, just raised a $170M Series A at a $1.1B valuation to build orbital data centers aimed at addressing the AI energy bottleneck on Earth.
It’s an ambitious vision, and clearly one that has resonated strongly with investors, helped in part by the broader momentum around space, energy, and compute. That said, I think the key question is not just whether space can provide cheaper energy, but whether it can deliver reliable, high-utilization compute at scale. Frontier AI depends on much more than power alone: networking, maintenance, upgrade cycles, resilience, and utilization all matter enormously.
My sense is that the most viable path may be a more hybrid one at first: keep the most fragile training workloads on Earth, and move only the workloads that structurally benefit from orbit. The opportunity is real, but the long-term winner will likely be the one that matches the vision with the most practical execution path.
