Orbital AI Compute Expansion
Starcloud Raises $250 Million to Advance Orbital Data Centers and AI Hardware
Key Points
Starcloud flew Nvidia's H100 GPU to orbit on Starcloud-1 in November 2025, marking Nvidia's first H100 deployment in space.
Starcloud is collaborating with Nvidia on the Space-1 Vera Rubin module, an AI hardware system to be flown in space with Nvidia developing the first purpose-built space GPU; the company aims to fly the Vera Rubin Space-1 GPU by late 2028.
Launch capacity remains the binding constraint: SpaceX plans to phase out Falcon 9 by 2028 in favor of Starship (which is still unproven), while other rockets (Blue Origin, ULA, Rocket Lab) are not yet flying regularly; Starcloud has also sought FCC permission to operate up to 88,000 satellites.
The $250 million extension is earmarked to expand manufacturing capacity, fund Nvidia collaboration on space GPUs, and secure future launch allocations—emphasizing launch capacity as a critical prerequisite for its orbital compute roadmap.
New investors Goanna Capital and Standard Capital joined the extension round, expanding the backer base alongside Cedar Capital, 776, Benchmark, EQT, Soma, and NFX.
Starcloud, an orbital data center startup, raised $250 million in a Series A extension that values the company at $2.3 billion. CityBiz reports the funding will help Starcloud expand manufacturing and deploy advanced satellites that could cut data center energy costs by up to 90% compared to ground facilities. The round was led by Manhattan West Ventures and includes major backers like Nvidia, Cisco Investments, and Benchmark.
The $250 million extension brings Starcloud's total Series A funding to $420 million. GeekWire notes that Nvidia contributed $25 million and is collaborating with Starcloud on space-ready AI chips. The company has already flown Nvidia's H100 GPU to orbit and plans to deploy its flagship Starcloud-3 spacecraft on SpaceX's Starship, but faces a critical bottleneck: securing enough launch capacity.
Nvidia Partnership and Space-Ready GPU Hardware
Starcloud and Nvidia are collaborating on the Space-1 Vera Rubin module, a GPU system designed to work in orbit. GeekWire states that Nvidia is developing the first purpose-built space GPU, with plans to fly the Vera Rubin Space-1 GPU by late 2028. Starcloud already made history in November 2025 by flying Nvidia's H100 GPU to space on its Starcloud-1 satellite—the first time Nvidia's top AI chip reached orbit.
The partnership addresses a key challenge for orbital computing. Space GPUs must withstand extreme temperatures and radiation that would destroy earth-based chips. Yahoo Finance reports that this collaboration is central to Starcloud's roadmap to enable AI inference—processing and analysis—directly in space where data centers can run 24/7 powered by continuous sunlight.
Launch Capacity: The Binding Constraint on Growth
Starcloud's biggest challenge is not technology or funding—it is access to rockets. The company has sought FCC approval to operate up to 88,000 satellites but cannot launch them without reliable access to space. SpaceX plans to stop flying its Falcon 9 rocket by 2028 and switch entirely to the Starship, which has never completed a full mission. Blue Origin, ULA, and Rocket Lab are not yet flying regularly enough to fill the gap.
This launch crunch is why Starcloud is emphasizing the importance of securing long-term launch agreements. Yahoo Finance reports that much of the $250 million extension is earmarked to fund manufacturing expansion and secure future launch allocations. Without steady launch capacity, even a well-funded startup cannot deploy the satellites needed to build out its orbital data center network.
How Orbital Data Centers Cut Energy Costs by 90%
Starcloud's core pitch is that data centers in space can be far more efficient than ground-based facilities. Continuous sunlight in orbit means solar panels generate power around the clock. Radiative cooling—heat radiating directly into space—eliminates the need for expensive air conditioning systems that consume vast amounts of electricity on Earth.
CityBiz reports that these physics advantages could reduce energy costs by up to 90% versus traditional data centers. As AI demand drives up electricity consumption globally, orbital computing offers a potential solution. However, the model only works if Starcloud can reliably and repeatedly launch satellites to scale its constellation—which depends on solving the launch capacity problem.
Investor Confidence and Expansion Plans
The $250 million extension reflects strong investor confidence in Starcloud's vision. New backers Goanna Capital and Standard Capital joined the round alongside returning investors Cedar Capital, 776, Benchmark, EQT, Soma, and NFX. Yahoo Finance notes that this diverse backer group includes venture firms and strategic investors like Nvidia and Cisco who see orbital computing as a next-generation infrastructure opportunity.
With $420 million in Series A funding, Starcloud plans to expand manufacturing capacity to build more satellites. The company will deploy Starcloud-2 and Starcloud-3 spacecraft, with Starcloud-3 targeted for launch on SpaceX's Starship. Each satellite will carry advanced AI chips and serve as a node in a global orbital data center network—if launch capacity permits.
Sentiment
Mostly PositivePublishers
12
Articles
29
Reach
41
