Starcloud lands $250M boost to scale orbital AI as rocket capacity tightens

Starcloud, which operates AI inference satellites in space, has secured an additional $250 million funding extension to its March Series A round, bringing its total raise to $420 million and lifting the company's valuation to $2.3 billion. The capital will fuel expansion of manufacturing capacity and accelerate development of Starcloud-3, a next-generation orbital data center designed for SpaceX's Starship rocket.

Booking Launch Capacity Before It Disappears#

The startup faces an unusual funding pressure: the tightening supply of available rockets. CEO Philip Johnston emphasized that securing launch capacity has become as critical as engineering capability, with SpaceX scheduled to retire its Falcon 9 workhorse by 2028 while Starship remains unproven and competing launchers like Blue Origin's New Glenn and ULA's Vulcan operate inconsistently.

"As soon as we can, we want to get under contract with things like Starship," Johnston told TechCrunch. The company has already requested FCC approval to deploy 88,000 spacecraft, a venture that will require massive launch infrastructure. Starcloud is weighing options including dedicated Falcon 9 missions and contracts with alternative providers to ensure reliable access to orbit.

Near-Term and Long-Term Deployment Plans#

Starcloud plans to launch two of its 8 kilowatt Starcloud-2 compute satellites on commercial rideshare missions during 2027, fulfilling contracts with US government customers who require in-orbit AI processing. Yet the company's long-term economics hinge on Starship succeeding as a rapidly reusable heavy-lift vehicle—a bet Johnston maintains is sound, despite SpaceX's recent announcement of delays in catching returning boosters.

"Obviously if we can't book any SpaceX launch capacity in 2029, that will be challenging for us," Johnston acknowledged.

Nvidia's Vote of Confidence#

The funding round, led by Manhattan West Ventures and backed by VCs including Benchmark and EQT, carries particular weight due to Nvidia's $25 million participation. Starcloud is the only company currently operating an H100 GPU in orbit and the first to train machine learning models using it—data that Nvidia is leveraging as it develops Vera Rubin Space-1, a custom GPU designed specifically for space environments.

Johnston credited Nvidia's deep technical vetting: "They, more than any other VC, did way more technical duty on this than anybody else." The space-ready chip targets a 2028 orbital debut, with engineers working through critical design trade-offs including thermal management, radiation hardening, and structural durability for rocket launch.