Starcloud, a startup developing satellites capable of performing artificial intelligence inference in orbit, announced that it had raised an additional $250 million for its $170 million Series A funding round announced in March. The deal increased the company’s valuation to $2.3 billion, according to TechCrunch, citing the company.
The additional funding was led by Manhattan West Ventures, with participation from Nvidia, Cisco and other investors, including Benchmark, EQT, Soma, NFX, 776, Cedar Capital, Goanna Capital and Standard Capital. A person familiar with the deal said Nvidia invested $25 million of the total.
Funding Manufacturing and Securing Launches
Starcloud plans to use the capital to open a larger manufacturing facility and advance development of its largest orbital data-center vehicle, Starcloud-3, which it plans to launch on SpaceX’s Starship rocket. CEO Philip Johnston is also raising funding to reserve launches in advance, at a time when access to orbit is expected to become scarcer.
The company says it has asked the U.S. Federal Communications Commission, or FCC, for authorization to operate 88,000 spacecraft. This ambition makes securing launches a central part of its operating model, as Johnston noted that launch costs represent one of the largest expenditure items, particularly given the planned schedule to end the Falcon 9 program in 2028.
In the nearer term, Starcloud plans to launch two vehicles from its new Starcloud-2 generation, each with 8 kilowatts of computing capacity, on rideshare missions in 2027. The two vehicles will perform orbital inference missions for customers, including U.S. government entities. The company is also considering purchasing a dedicated Falcon 9 launch and signing contracts with other providers to support its future missions.
What Is Holding Back Orbital Expansion?
Starcloud’s long-term vision depends on Starship’s ability to reduce launch costs enough to create an orbital computing layer that can compete with terrestrial data centers. But this path faces practical uncertainty: SpaceX is working to phase out reliance on Falcon 9 and introduce the larger Starship, while competing rockets such as Blue Origin’s New Glenn and ULA’s Vulcan are not operating regularly, and Rocket Lab’s Neutron has not yet reached the launch pad.
The pressure is increasing as SpaceX delays an attempt to catch the returning Starship by several months and postpones reusing the vehicle for the first time until the end of the year or early 2027, according to what Elon Musk said this week. Johnston therefore warned that Starcloud’s inability to reserve launch capacity with SpaceX in 2029 would pose a major challenge for the company.
Nvidia’s Role in Developing Space Computing
Starcloud says it is currently operating, to its knowledge, an Nvidia H100 graphics processing unit inside an orbital data center, and that it used the GPU to train a model in orbit for the first time. This experiment differs from most other spaceborne graphics processing units, which are designed to process data at the edge.
Starcloud is sharing its results with Nvidia as the company develops a space-specific chip called Vera Rubin Space-1, and Starcloud hopes to send it into orbit in late 2028. Design decisions under consideration include the relationship between operating temperature and radiator size, the location of radiation shielding, and the degree to which the chips must be reinforced to withstand the violence of launch.
Starcloud currently employs 25 people and is developing production lines in a 100,000-square-foot facility in Woodinville, Washington.