Starcloud, a startup developing satellites that perform AI inference in orbit, told TechCrunch it has closed a $250 million extension to its March $170 million Series A round, valuing the company at $2.3 billion. The new capital will fund a larger manufacturing facility and advance its largest orbital data center spacecraft, Starcloud-3, which is intended to launch on SpaceX‘s Starship rocket. CEO Philip Johnston said the company needs to book a large amount of launch capacity early, citing market constraints: Falcon 9 is scheduled to end service in 2028, Starship remains unproven, and competing vehicles like Blue Origin’s New Glenn, ULA’s Vulcan, and Rocket Lab’s Neutron are not yet flying regularly. “We can see what’s coming—we’re going to need to book an enormous amount of launch,” Johnston said. Starcloud has already requested FCC permission to operate 88,000 spacecraft.

Near-term plans focus on launching two Starcloud-2 spacecraft—a new generation of 8 kW compute satellites—on rideshare flights in 2027. Those will perform orbital inference tasks for customers including U.S. government agencies. Starcloud is also considering buying a dedicated Falcon 9 launch and signing contracts with other providers to secure future capacity. The company’s long-term strategy depends on Starship lowering launch costs enough to build an orbital inference layer competitive with terrestrial data centers. Johnston said he remains confident in SpaceX‘s ability to quickly and often reuse the rocket, even after SpaceX CEO Elon Musk said this week that an attempt to catch a returning Starship will be delayed a few months, with the first re-flight now targeted for late 2026 or early 2027. “Obviously if we can’t book any SpaceX launch capacity in 2029, that will be challenging for us,” Johnston said.

The extension was led by Manhattan West Ventures, with participation from Nvidia and Cisco. A person familiar with the deal said Nvidia invested $25 million. Other participants included Benchmark, EQT, Soma, NFX, 776, Cedar Capital, Goanna Capital, and Standard Capital. Johnston pointed to Nvidia‘s investment as a signal of Starcloud‘s advantages: it is the only company currently operating an Nvidia H100 terrestrial data center GPU in orbit and the first to train a model using it. Most other space GPUs are designed for edge processing, not full data center-class inference. Starcloud is sharing its on-orbit data with Nvidia as the chipmaker develops its first purpose-built GPU for space, the Vera Rubin Space-1 chip. “The reason they’ve chosen to do this investment now is because of all of this data that we got from Starcloud One,” Johnston said. “They, more than any other VC, did way more technical due diligence on this than anybody else.”

The space-ready chip has not been built yet. Starcloud hopes to fly it in late 2028. Engineers are tracking key design choices: the relationship between chip running temperature and radiator size to dissipate heat, radiation shielding placement, and the ruggedization needed to survive launch. The company, now 25 employees, is building production lines at a 100,000-square-foot facility in Woodinville, Washington, near where SpaceX and Amazon build their communications satellites. The article notes that launch costs were already a major hurdle for orbital data center startups, to the point that one startup has decided to build its own rockets. Starcloud‘s approach is to pre-book capacity and deepen its Nvidia relationship while waiting for Starship to mature.