The Redmond, Washington startup, which already has an Nvidia H100 GPU operating in orbit and has trained the first AI model in space, is now building a Starship-class spacecraft designed to be the first orbital data center cost-competitive with ground-based facilities.
star cloud has raised $170 million in a Series A round led by Benchmark and EQT Ventures, valuing the Redmond, Washington company at $1.1 billion. The round closed 17 months after Starcloud’s Y Combinator demo day presentation, making it one of the fastest startups in YC history to achieve unicorn status.
Total funding now stands at $200 million. The company builds data centers in space, starting with GPU computing for other satellites and working toward a long-term vision of orbital infrastructure capable of handling the same workloads as terrestrial hyperscale facilities.
The central proposal is simple to describe and tremendously difficult to execute. In orbit, solar energy is essentially unlimited and free once a satellite is deployed. The cooling is passive: waste heat is radiated into deep space, which operates at around -270°C, without the need for water.
The 💜 of EU technology
The latest rumors from the EU tech scene, a story from our wise founder Boris and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
There are no planning permissions, no grid connections, and no land acquisition battles. Starcloud CEO and co-founder Philip Johnston argues that these structural advantages will make orbital data centers cost-competitive with terrestrial ones, once launch costs drop enough. The problem, as Johnston openly acknowledges, is that the enabling technology is not yet operational.
However, Starcloud is further ahead than any of its competitors. In November 2025, just 21 months after its founding, it launched Starcloud-1: a 60kg satellite carrying an Nvidia H100 GPU, making it approximately 100 times the most powerful GPU ever operated in space.
The satellite subsequently became the first to train an AI model in orbit, specifically NanoGPT, trained on the complete works of Shakespeare, and the first to run a version of Gemini. The company is now processing data from Capella Space’s radar satellites in orbit, the first commercial use case. An Nvidia A6000 GPU failed during launch, a technical reality that will influence future hardware choices.
Series A will finance three things. First, Starcloud-2, launching in October 2026: a more powerful satellite with multiple GPUs, including a Nvidia Blackwell Chipan AWS blade server and a bitcoin mining computer.
It will carry the largest deployable radiator ever flown on a private satellite. Second, the company will begin developing Starcloud-3: a 200-kilowatt, three-ton spacecraft designed to fit the ‘fish dispenser’ deployment system that SpaceX built to launch Starlink satellites from Starship.
Johnston says Starcloud-3 should be the first orbital data center that can genuinely compete with ground-based facilities on cost, with projected prices of around $0.05 per kilowatt-hour, but only if Starship’s commercial launch costs reach roughly $500 per kilogram.
It expects commercial access to Starship to open in 2028 or 2029. If Starship is delayed, the company will continue launching smaller satellites on Falcon 9.
The strategic context is saturated and increasingly so. spacex, which acquired Elon Musk’s artificial intelligence company xAI In February 2026, it requested permission from the US government to build and operate a distributed computing network of one million satellites in orbit. Blue Origin has expressed similar ambitions.
Google has Project Suncatcher; Aethero launched Nvidia’s first space Jetson GPU in 2025; Aetherflux recently reported gaining a valuation of $2 billion.
The scale gap between what exists in orbit and what exists on Earth remains extraordinary: SpaceX’s Starlink network, with 10,000 satellites, generates around 200 megawatts of power, while data centers with more than 25 gigawatts of combined capacity are currently being built in the United States alone.
Long-term plans, Starcloud’s includes an eventual constellation of 88,000 satellites, represent a future that may not arrive in time.
