Starcloud’s latest funding round values the startup at $1.1 billion, making it one of the fastest to reach unicorn status after graduating from Y Combinator.
The company’s Series A, which closed 17 months after its demo day, was led by Benchmark and EQT Ventures. Another sign of interest in outsourcing data centers to orbit as infrastructure and political barriers slow their growth on Earth, but the business model relies on unproven technology and high capital costs.
Starcloud has now raised a total of 200 million dollars, and launched its first satellite with an Nvidia H100 GPU in November 2025. The company will launch a more powerful version, Starcloud 2, later this year with more GPUs, including an Nvidia Blackwell chip and an AWS server blade, and a bitcoin mining computer.
The company will also begin building a data center spacecraft designed to launch on Starship, the reusable heavy-lift rocket being built by Elon Musk’s SpaceX. Starcloud 3, as the spacecraft is called, will be a 200-kilowatt, three-ton “pez dispenser” system SpaceX designed to launch its Starlink satellites from Starship.
CEO and founder Philip Johnston said he expects that will be the first orbital data center with costs competitive with terrestrial data centers, with costs of the order of $ .05 per kw / hour of energy – if commercial implementation costs the earth about $ 500 per kilogram.
The challenge is that the Starship doesn’t fly yet; Johnston says he expects commercial access to open in 2028 and 2029. That’s the reality facing all major space data center projects: powerful space computers will be expensive until a new generation of rockets begins to launch into high-performance space, something that may not happen until the 2030s.
“If it ends up being delayed, we will continue to launch smaller versions of the Falcon 9,” Johnston said. “We won’t be competitive on energy costs until the Starship is flying regularly.”
Techcrunch event
San Francisco, CA
|
October 13-15, 2026
“There’s kind of two types of business,” Johnston explained: One is selling processing power to other spacecraft in orbit; the company’s first satellite, for example, is analyzing data collected by Capella Space’s radar spacecraft. Then, in the future when startup costs come down, more powerful distributed data centers may pull work from their terrestrial counterparts.
That shows how new this field is. When Nvidia CEO Jensen Huang presented the company’s Vera Rubin Space-1 modules at his company’s annual GPU Technology conference last week, he didn’t note that none were manufactured or shared with the company’s development partners.
In fact, the number of advanced GPUs in orbit is estimated at twenty, while Nvidia is estimated to have sold about 4 million hyperscalers on Earth by 2025.
Or consider that SpaceX’s Starlink communications network, a massive network of orbiting satellites with 10,000 spacecraft, produces about 200 something. megawatts of energy, while more than 25 data centers gigawatts electric power is currently under construction in the US, according to Cushman and Wakefield.
Johnston says his company is ahead of the competition, with the first terrestrial GPU installed in orbit. It was used to train an AI model in orbit, the first, according to Starcloud, to use the Gemini version. Beyond performance, Johnston says Starcloud now has valuable data about what it takes to run a powerful chip in space.
“The 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 can use state-of-the-art chips in space,” he told TechCrunch. That hard-earned knowledge — another GPU, the Nvidia A6000, failed at launch — will influence future designs.
There is a laundry list of technical challenges to be solved, including generating efficient power and cooling thermally active chips. Starcloud-2 will have the largest radiator launched by a private satellite; he expects at least two more versions of that spacecraft to orbit, Johnston said.
Then there is the challenge of synchronization. Large data center workloads, typically for training, require hundreds or thousands of GPUs to run in parallel. Doing that in space would require incredibly large spacecraft, or powerful and reliable laser links between spacecraft flying in formation. Most of the companies working on this technology expect those payloads to arrive long after simple targeting tasks have taken place in orbit.
In addition to Starcloud, Aetherflux, Google’s Project Suncatcher, and Aethero — which launched Nvidia’s first Jetson GPU for space in 2025 — are all developing space data center businesses.
The elephant in the room is SpaceX itself, which has requested permission from the US government to build and operate a million satellites to launch into space.
Taking on SpaceX is a daunting task for any entrepreneur, but Johnston sees room for coexistence.
“They’re building a slightly different use case than we are,” he told TechCrunch. “Especially they plan to do a lot of work for Grok and Tesla. It’s possible that at some point they’ll offer a third-party cloud service, but I think it’s unlikely they’ll do what we’re doing. [as] energy and infrastructure player.”
