A Startup Wants to Put Data Centers in Space — and Just Raised $250 Million to Do It

Starcloud has raised $250 million in new funding, valuing the startup at $2.3 billion. Manhattan West Ventures led the round, with Nvidia contributing $25 million and Cisco joining alongside Benchmark, EQT, Soma, NFX, 776, Cedar Capital, Goanna Capital, and Standard Capital (TechCrunch).

The extension comes just over five months after Starcloud's initial $170 million funding round, which valued the company at $1.1 billion and was led by Benchmark (SpaceNews). That March round made Starcloud the fastest startup in Y Combinator history to reach a $1 billion valuation, doing so 17 months after its demo day (Y Combinator). The new money will fund a larger manufacturing facility and the development of Starcloud-3, the company's largest spacecraft, intended to fly on SpaceX's forthcoming Starship rocket (TechCrunch).

Starcloud, led by CEO Philip Johnston, is building satellites that can run AI workloads in space. AI systems have two main phases: training, where the system learns from large amounts of data, and inference, where the trained system produces results from new data. Starcloud has done both in orbit. It is the only company currently running an Nvidia H100 — a powerful data center chip — in space, and the first to train a model using it (TechCrunch). The company's next spacecraft, Starcloud-2, is scheduled for launch in 2027. Each satellite will provide 8 kilowatts of computing power for AI tasks in orbit (TechCrunch; SpaceNews). Customers for Starcloud-2's services include US government agencies (TechCrunch).

The company is thinking far bigger than a few satellites. Starcloud has filed a request with the FCC to operate 88,000 spacecraft (TechCrunch). Its website promotes space-based data centers as offering 90% lower electricity costs and round-the-clock solar power (Starcloud.com), positioning orbital computing as an answer to the growing energy demands of data centers on Earth (Y Combinator).

The challenge is that getting satellites into space is becoming harder. SpaceX plans to retire its Falcon 9 rocket by 2028 as it transitions to the newer Starship (TechCrunch). Rival rockets from Blue Origin and ULA are not yet flying on a regular schedule, and Rocket Lab's Neutron has not yet launched (TechCrunch). SpaceX CEO Elon Musk said the company will delay an attempt to catch a returning Starship rocket by a few months and will try to re-fly the vehicle for the first time at the end of 2026 or early 2027 (TechCrunch).

To work around limited launch options, Starcloud is considering buying a dedicated Falcon 9 launch and signing contracts with other rocket providers for future missions (TechCrunch). Starcloud-3 depends on Starship, a rocket that has not yet proven it can launch, land, and relight reliably — let alone fly often enough to support large-scale infrastructure in space.

Nvidia's involvement is worth noting beyond the money. The $25 million is a relatively small investment for Nvidia, but it suggests the leading maker of AI chips sees enough promise in space-based computing to take a financial stake. Starcloud's ability to run an H100 in orbit and train a model on it gives that investment a concrete basis rather than a purely speculative bet.

The appeal of putting computers in space comes down to two advantages. First, satellites in the right orbit get continuous sunlight, meaning uninterrupted solar power without clouds or darkness. Second, if the data is collected in space — by Earth-observation satellites, for example — processing it in orbit avoids the delay of sending it back to Earth first. Starcloud's government customers for Starcloud-2 suggest early demand is concentrated in exactly those cases where the data does not need to come back to Earth in raw form.

The 88,000-satellite filing points to Starcloud's long-term vision: a vast network of data centers in orbit. That request will go through a regulatory review process that has historically taken years, even for far smaller constellations. The nearer-term test will be the 2027 Starcloud-2 launches, which will show whether orbital computing can serve paying customers at a price that makes business sense.

The open question is whether the total cost of launching, powering, cooling, and maintaining computers in space can get close to what it costs to run a data center on Earth, even with the energy advantages Starcloud claims. The company has the funding, a partnership with the leading AI chipmaker, and a proven orbital GPU. The next 18 months will show whether that is enough to turn space-based data centers from an experiment into real infrastructure.