Starcloud raises $250 million for orbital data centers as launch options dry up
TechCrunch Tim Fernholz ● Covered by 2 sources
Starcloud just added $250 million and says it needs it for orbital AI gear. The catch: it’s racing to lock down rocket launches before SpaceX’s plans shift.
Based on reporting by TechCrunch, Tim Fernholz — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Starcloud has tacked a $250 million extension onto its March Series A, lifting the round to $420 million and the company’s valuation to $2.3 billion. The startup builds satellites that do AI inference in orbit, and this cash is meant to help it scale up fast: a bigger factory, more production lines, and its largest spacecraft yet, Starcloud-3.
That bigger ship is supposed to ride SpaceX’s Starship, which is exactly why CEO Philip Johnston sounds so eager to get money in the door. He says launch capacity is getting tighter, and that securing rides is now one of the biggest costs in the whole business. Starcloud has already asked the FCC for permission to operate 88,000 spacecraft, and Johnston wants contracts with SpaceX as soon as possible.
The timing is awkward for anyone building around rockets that don’t quite exist yet. SpaceX plans to phase out Falcon 9 in 2028, while Starship is still working toward regular reuse. Blue Origin’s New Glenn and ULA’s Vulcan are not flying regularly, and Rocket Lab’s Neutron is still on the pad. That leaves satellite operators trying to plan a fleet around a launch market that feels short on options and long on promises.
For now, Starcloud’s near-term plan is more modest. It wants to launch two of its 8 kW Starcloud-2 satellites on rideshare missions in 2027, with orbital inference work for customers that include U.S. government agencies. The company is also weighing a dedicated Falcon 9 flight and contracts with other providers to keep future missions from getting stranded.
Nvidia and Cisco joined the extension, and a person familiar with the deal said Nvidia put in $25 million. Johnston clearly thinks that matters. Starcloud says it is the only company currently running a Nvidia H100 terrestrial data center GPU in orbit, and the first to train a model on one. It has been feeding those results back to Nvidia as the chipmaker works on a space-specific Vera Rubin Space-1 chip, which Starcloud hopes to fly in late 2028.
My take — AI-written commentary, not fact-checked reporting
This is the most honest part of the space-compute pitch: the bottleneck is not the AI, it’s the rocket schedule. A billion-dollar orbital cloud dream still lives or dies on whether SpaceX can actually behave like an airline instead of a very expensive science project. A startup buying chips for space is cute; a startup stockpiling launch access is the real business plan.
Read more about this at: TechCrunch