Starcloud raises $250M for orbital AI data centers

Starcloud, a startup building satellites that perform AI inference in orbit, has added a $250 million extension to its $170 million Series A round from March, valuing the company at $2.3 billion. The new capital will fund a larger manufacturing facility and further development of Starcloud-3, the company’s largest planned orbital data-center spacecraft, which is meant to fly on SpaceX’s forthcoming Starship rocket. CEO Philip Johnston is also stockpiling capital to secure launch capacity as the market for rocket transportation tightens: Starcloud has already asked the FCC for permission to operate 88,000 spacecraft. “We can see what’s coming — we’re going to need to book an enormous amount of launch,” Johnston told TechCrunch.

Launch costs are already one of the biggest constraints on orbital data-center startups — enough that one rival has decided to build its own rockets. SpaceX is moving to phase out its workhorse Falcon 9, which is scheduled to end in 2028, in favor of the larger, still-unproven Starship, complicating planning for satellite operators; competing options like Blue Origin’s New Glenn and ULA’s Vulcan aren’t flying regularly, and Rocket Lab’s Neutron isn’t on the pad yet. For now, Starcloud is focused on launching two new 8 kW compute satellites, Starcloud-2, on rideshare flights in 2027 to run orbital inference tasks for customers including U.S. government agencies, while considering a dedicated Falcon 9 launch and contracts with other providers.

The extension was led by Manhattan West Ventures, with participation from Nvidia and Cisco; a person familiar with the deal said Nvidia contributed $25 million. Other backers include Benchmark, EQT, Soma, NFX, 776, Cedar Capital, Goanna Capital, and Standard Capital. Johnston points to the Nvidia investment as validation: Starcloud says it’s the only company currently operating an Nvidia H100 GPU in orbit and the first to train a model there, and it’s sharing that data with Nvidia as the chipmaker develops the Vera Rubin Space-1, its first purpose-built chip for space — which hasn’t been built yet, though Starcloud hopes to fly it by late 2028. The company, now 25 employees, is building production lines at a 100,000-square-foot facility in Woodinville, Washington.

Why it matters

Starcloud’s bet is that Starship eventually makes launch cheap enough to build an orbital inference layer that competes with terrestrial data centers — but that bet depends entirely on a rocket that hasn’t yet proven it can be reused quickly and often, at a moment when the industry’s current workhorse is being phased out.