StarCloud’s latest funding round values the space computing company at $1.1 billion, making it one of the first startups to reach unicorn status after graduating from Y Combinator.
The company’s Series A, which closed 17 months after its Demo Day presentation, was led by Benchmark and EQT Ventures. This is another sign of interest in outsourcing data centers in orbit, as resource and political obstacles have slowed the development of data centers on Earth, but the business model relies on unproven technology and large capital investments.
Starcloud has now raised a total of $200 million and will launch its first satellite powered by Nvidia H100 GPUs in November 2025. Later this year, the company plans to launch a more powerful version, Starcloud 2, with multiple GPUs and a Bitcoin mining computer, including Nvidia Blackwell chips and AWS server blades.
The company also plans to begin developing a data center spacecraft designed to launch from Starship, a reusable heavy-lift rocket manufactured by Elon Musk’s SpaceX. StarCloud 3, as its name suggests, is a 200-kilowatt, 3-ton spacecraft that fits into the “Pez Dispenser” system designed by SpaceX to deploy Starlink satellites from Starship.
CEO and founder Philip Johnston said he expects this to be the first in-orbit data center to be cost-competitive with ground-based data centers, with costs around $0.05 per kilowatt hour per power hour, given commercial launch costs of about $500 per kilogram.
The challenge is that Starship isn’t flying yet. Johnston said he expects commercial access to open in 2028 and 2029. This is the reality facing all large-scale space data center projects. Powerful space computers will be cost-prohibitive until new generations of rockets can be launched at high operational cadences, which may not happen until the 2030s.
“If it’s delayed, we’ll just continue to launch a smaller version of the Falcon 9,” Johnston said. “Until Starship flies frequently, we can’t be competitive on energy costs.”
tech crunch event
San Francisco, California
|
October 13-15, 2026
“There are two business models,” Johnston explains. One is to sell processing power to other spacecraft in orbit. For example, the company’s first satellite will analyze data collected by Capella Space’s radar spacecraft. And in the future, more powerful distributed data centers could take work from ground-based data centers as launch costs come down.
This shows how new this industry is. When Nvidia CEO Jensen Huang announced the company’s Vera Rubin Space-1 chip modules at the company’s annual GPU technology conference last week, he did not mention that none were manufactured or shared with the company’s development partners.
In fact, the number of advanced GPUs in orbit is in the dozens, and Nvidia is estimated to have sold nearly 4 million to hyperscalers on the ground in 2025.
Or consider that SpaceX’s Starlink communications network, the largest satellite network in orbit with 10,000 spacecraft, generates about 200 megawatts of energy, while data centers with more than 25 gigawatts of power are currently under construction in the United States, according to Cushman and Wakefield.
Johnston claims the company is far ahead of its competitors by deploying the first GPU on Earth in orbit. According to StarCloud, it was used for the first time to train an AI model in orbit and run a version of Gemini. Johnston says StarCloud has yielded valuable data about not just performance, but what it takes to run powerful chips in space.
“Honestly, the H100 is probably not the best chip for space, but the reason we did this was because we wanted to prove that we could run cutting-edge ground-based chips in space,” he told TechCrunch. The hard-earned knowledge (another GPU, the Nvidia A6000, failed during launch) will influence future designs.
There are many technical challenges to be solved, such as efficient power generation and cooling of hot running chips. StarCloud 2 will have the largest deployable radiator ever flown on a commercial satellite. Johnston said he expects at least two additional versions of the spacecraft to head into orbit.
Then there’s the challenge of synchronization. The largest datacenter workloads (often for training) require hundreds or thousands of GPUs working together. Doing that in space requires either very large spacecraft or powerful, reliable laser links between spacecraft flying in formation. Most companies working on this technology expect these workloads to occur long after simpler inference tasks are in orbit.
Besides Starcloud, Aetherflux, Google’s Project Suncatcher, and Aethero, which launched Nvidia’s first space-based Jetson GPU in 2025, are all developing space data center businesses.
The elephant in the room is SpaceX itself, which is seeking permission from the U.S. government to build and operate 1 million satellites for distributed computing in space.
Taking on SpaceX would be a difficult task for any entrepreneur, but Johnston believes there is room for coexistence.
“They’re building for a slightly different use case than we are,” he told TechCrunch. “They plan to primarily service Grok and Tesla workloads. They may eventually offer third-party cloud services, but what I don’t think they’ll do is the same thing we do as an energy and infrastructure player.”
