Starcloud hits $1.1B valuation building data centers that orbit Earth

Starcloud closed a $170 million Series A at a $1.1 billion valuation 17 months after graduating from Y Combinator, led by Benchmark and EQT Ventures. The Redmond, Washington-based startup is building solar-powered satellites that operate as orbital data centers.

Starcloud launched its first satellite in November 2025 carrying an Nvidia H100 GPU. That spacecraft successfully trained an AI model in orbit and ran a version of Gemini, according to the company. Philip Johnston told TechCrunch the mission provided critical operational data about running state-of-the-art terrestrial chips in space.

"An H100 is probably not the best chip for space, to be honest, but the reason we did it is we wanted to prove that we could run state of the art terrestrial chips in space," Johnston said.

The unit economics problem

Johnston expects orbital data centers will become cost-competitive with terrestrial facilities at power costs around $0.05 per kilowatt-hour, achievable only if commercial launch costs hit roughly $500 per kilogram. Launch economics determine everything.

SpaceX's Starship needs to reach operational cadence for commercial customers first. Johnston estimates that won't happen until 2028 or 2029. The rocket has not yet carried commercial payloads to orbit.

Johnston told TechCrunch that delays wouldn't kill the company. "If it ends up being delayed, we'll just carry on launching the smaller versions on Falcon 9. We're not going to be competitive on energy costs until Starship is flying frequently."

SpaceX's Starlink network produces roughly 200 megawatts of power across 10,000 satellites. Data centers with more than 25 gigawatts of power are under construction in the US alone, according to Cushman and Wakefield. The scale gap between orbital and terrestrial infrastructure is enormous.

Two revenue models

Starcloud sells processing power to other spacecraft on orbit today. The first satellite analyzes data collected by Capella Space's radar satellites. Long-term, Johnston expects the company to serve terrestrial workloads once launch economics improve.

Orbital data centers become viable for compute-intensive workloads currently handled on the ground when launch costs drop. Everything hinges on Starship delivering the unit economics Johnston described.

Starcloud plans to launch Starcloud 2 later this year carrying multiple GPUs including an Nvidia Blackwell chip, an AWS server blade, and bitcoin mining hardware. The satellite will feature the largest deployable radiator flown on a private spacecraft. Johnston expects at least two additional versions before moving to the Starship-class platform.

Starcloud 3 targets Starship deployment

Starcloud 3 will deliver 200 kilowatts from a three-ton platform. SpaceX's Starship will deploy it using the "PEZ dispenser" system, the same mechanism used for Starlink satellites.

Starcloud designed this spacecraft as its first platform optimized for reusable heavy-lift economics. Falcon 9 missions are technology validation, not commercially viable infrastructure.

Technical challenges remain unsolved

Starcloud 2 will test an oversized radiator to address thermal management challenges. Power generation at this scale remains unsolved. Efficiency data doesn't exist yet.

AI training at scale requires hundreds or thousands of GPUs working in tandem. Starcloud would need either massive spacecraft or high-bandwidth laser links between satellites flying in formation to achieve that coordination.

Nvidia shipped an estimated 4 million GPUs to terrestrial hyperscalers in 2025. The number of advanced GPUs on orbit is measured in dozens. Most companies working on orbital compute expect simpler inference tasks will precede training workloads.

Competition from SpaceX and others

Starcloud competes with Aetherflux (reportedly raising a Series B at $2 billion valuation), Google's Project Suncatcher, and Aethero, which launched Nvidia's first space-based Jetson GPU in 2025.

SpaceX filed with the US government for permission to build and operate a million satellites for distributed compute in space. Johnston said the two companies are building for different use cases.

"They are building for a slightly different use case than us," Johnston told TechCrunch. "They're mainly planning on serving Grok and Tesla workloads. It may be at some point that they offer a third-party cloud service, but what I think they are unlikely to do is what we're doing [as] an energy and infrastructure player."

Hardware lessons from launch failures

Starcloud lost an Nvidia A6000 GPU during an earlier launch. Johnston said that failure informed hardware design for subsequent missions. The H100 survived launch and operated successfully, validating the ruggedization approach.

Jensen Huang unveiled Nvidia's Vera Rubin Space-1 chip modules at GTC last week without mentioning that none had been produced or shared with development partners. The space compute sector remains in early validation phases.

Starcloud raised $200 million total since founding. The Series A closed 17 months after the company's Y Combinator demo day, making it one of the fastest graduates to reach unicorn status in the 2025-26 cohort.