Nvidia Corp (NASDAQ:NVDA, XETRA:NVD)-backed startup Starcloud has achieved a milestone by training the first artificial intelligence model in space, according to a CNBC report.
The company launched its Starcloud-1 satellite in early November 2025, equipped with an Nvidia H100 GPU, which the company says is 100 times more powerful than any graphics processor previously flown in orbit.
According to Starcloud CEO Philip Johnston, the satellite successfully ran Google’s open large language model Gemma and trained NanoGPT on Shakespeare’s works, generating responses in the style of the playwright.
“Anything you can do in a terrestrial data center, I’m expecting to be able to be done in space,” Johnston told CNBC.
He added that orbital data centers could reduce energy costs by up to 10 times compared with Earth-based facilities.
Starcloud-1 also integrates satellite telemetry for real-time queries, enabling AI to analyze orbital data such as imagery for wildfire detection or maritime rescue.
The company said the project demonstrates the potential for space-based AI to alleviate terrestrial infrastructure constraints, including electricity consumption, water use, and greenhouse gas emissions from data centers.
Looking ahead, Starcloud plans to launch a 5 GW orbital cluster with Nvidia H200 chips and Crusoe cloud integration by October 2026, aiming to provide large-scale AI computing powered entirely by solar energy.
There is increasing interest in orbital data centers, with initiatives including Google’s Project Suncatcher and lunar projects by Lonestar.