Starcloud is trying to move a piece of the data center business off Earth and into orbit. Its new $170 million Series A puts the company at a $1.1 billion valuation and makes it one of the fastest startups to reach unicorn status after graduating from Y Combinator.
The round, led by Benchmark and EQT Ventures, arrives as companies look for ways around the resource and political obstacles that can slow data center development on Earth. But the bet is still early. Starcloud’s plan depends on technology that has not yet been proven at scale and on capital-heavy infrastructure that will be expensive to build.
A funding round built around a big infrastructure bet
Starcloud has now raised a total of $200 million. The company’s first major proof point came in November 2025, when it launched its first satellite carrying an Nvidia H100 GPU.
That satellite is not meant to look like a finished orbital data center. It is a step toward learning whether powerful terrestrial chips can operate in space, where power, heat and reliability constraints are very different from conditions inside a conventional facility.
The next version, Starcloud 2, is planned for later this year. It will carry multiple GPUs, including an Nvidia Blackwell chip and an AWS server blade. It will also include a bitcoin mining computer.
Starcloud is also beginning work on a larger data center spacecraft called Starcloud 3. That spacecraft is designed to launch from Starship, the reusable heavy lift rocket being built by Elon Musk’s SpaceX. Starcloud 3 is planned as a 200 kilowatt, three-ton spacecraft that fits the “PEZ dispenser” system SpaceX designed to deploy its Starlink satellites from Starship.
Why launch costs shape the whole model
CEO and founder Philip Johnston says Starcloud 3 could become the first orbital data center that is cost-competitive with terrestrial data centers. His expectation depends on commercial launch costs reaching around $500 per kilogram.
Johnston said that, under that assumption, costs could be on the order of $.05 per kw/hour of power. That is the core economic promise: if launch prices fall enough, space-based compute could eventually compete with data centers on Earth for some workloads.
The problem is timing. Starship is not flying yet. Johnston says he expects commercial access to open up in 2028 and 2029. Until a new generation of rockets launches often enough, powerful computers in orbit remain costly to deploy.
“If it ends up being delayed, we’ll just carry on launching the smaller versions on Falcon 9,” Johnston said. “We’re not going to be competitive on energy costs until Starship is flying frequently.”
That means Starcloud’s near-term business cannot depend only on replacing terrestrial data centers. The company needs smaller, practical orbital use cases while the larger infrastructure stack matures.
Two markets for compute in orbit
Johnston describes two business models. The first is selling processing power to other spacecraft already in orbit. Starcloud’s first satellite, for example, analyzes data collected by Capella Space’s radar spacecraft.
The second model is more ambitious. If launch costs fall, more powerful distributed data centers in orbit could potentially take work from data centers on Earth. That future would require more hardware, more power and more reliable orbital networking than exists today.
The gap between the idea and the current market is large. When Nvidia CEO Jensen Huang unveiled the company’s Vera Rubin Space-1 chip modules at his company’s annual GPU Technology Conference last week, he did not note that none had been produced or shared with the company’s development partners.
Advanced GPUs in orbit are still counted in the dozens. By contrast, Nvidia is estimated to have sold nearly 4 million to terrestrial hyperscalers in 2025.
The power comparison is just as stark. SpaceX’s Starlink communications network, described as the largest satellite network in orbit with 10,000 spacecraft, produces something around 200 megawatts of energy. Data centers with more than 25 gigawatts of power are currently under construction in the U.S., according to Cushman and Wakefield.
The hard engineering still ahead
Johnston argues that Starcloud is ahead of other space data center efforts because it has deployed the first terrestrial GPU in orbit. According to Starcloud, that GPU was used to train an AI model in orbit and run a version of Gemini.
The company says the result matters beyond performance. It now has data about what it takes to run a powerful chip in space. That includes lessons from a failure: another GPU, an Nvidia A6000, failed during launch.
“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 told TechCrunch.
The technical list is long. Starcloud must deal with efficient power generation, cooling for hot-running chips and the limits of spacecraft design. Starcloud-2 will have the largest deployable radiator flown on a private satellite, and Johnston expects at least two additional versions of that spacecraft to go to orbit.
Synchronization may be even harder. Large data center workloads, especially training, can require hundreds or thousands of GPUs working together. In space, that could mean very large spacecraft or reliable laser links between spacecraft flying in formation. Most companies working on the technology expect simpler inference tasks to happen in orbit before the largest training workloads.
A competitive field with SpaceX in the background
Starcloud is not alone. Aetherflux, Google’s Project Suncatcher and Aethero are all developing space data center businesses. Aethero launched Nvidia’s first space-based Jetson GPU in 2025.
The largest potential rival is SpaceX itself. SpaceX has asked the U.S. government for permission to build and operate a million satellites for distributed compute in space.
Johnston says Starcloud can still coexist with SpaceX because he sees the two companies aiming at different roles.
“They are building for a slightly different use case than us,” he 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.”
That distinction is central to Starcloud’s pitch. The company is not only trying to put chips in orbit. It is trying to build the power, cooling and infrastructure layer that could make orbital compute useful beyond isolated demonstrations.
The Series A shows that investors are willing to fund that possibility. The harder question is whether launch economics, spacecraft design and distributed GPU systems can mature quickly enough to turn space data centers from a technical milestone into a competitive business.