TLDR
- Starcloud’s satellite with Nvidia H100 GPU ran Google’s Gemma AI model in space for the first time in November 2025
- The orbital chip delivers 100 times more processing power than any previous space-based GPU and trained NanoGPT using Shakespeare
- Company targets 5-gigawatt space data center powered by uninterrupted solar energy to ease Earth’s infrastructure pressure
- October 2026 launch will deploy multiple H100 chips and Nvidia Blackwell platform for expanded capabilities
- Google Project Suncatcher, Lonestar Data Holdings, and Aetherflux pursuing competing orbital data center projects
Starcloud has made history by training an artificial intelligence model in space. The Washington-based company launched a satellite with an Nvidia H100 graphics processing unit in early November 2025.
The Starcloud-1 satellite currently operates Google’s Gemma large language model in orbit. This marks the first successful operation of an LLM on a high-powered GPU outside Earth’s atmosphere.
The H100 chip provides 100 times more computing power than previous space-based GPUs. Gemma sent its initial transmission saying “Greetings, Earthlings” and describing its mission to observe and analyze from space.
CEO Philip Johnston said the test demonstrates that data centers can function in orbit. The company also trained NanoGPT, developed by OpenAI founding member Andrej Karpathy, using the complete works of Shakespeare.
Space Data Centers Target Energy Problems
Terrestrial data centers face growing energy and environmental challenges. These facilities strain power grids and use billions of gallons of water annually while generating substantial greenhouse gas emissions.
The International Energy Agency forecasts data center electricity use will more than double by 2030. Johnston stated that Starcloud’s orbital operations will cost 10 times less in energy than ground-based centers.
The company plans a 5-gigawatt orbital data center with solar panels spanning roughly 4 kilometers in each direction. Space-based facilities capture constant solar energy without interruption from weather or Earth’s rotation.
Johnston explained the motivation stems from energy limitations on Earth. “Anything you can do in a terrestrial data center, I’m expecting to be able to be done in space,” he said.
Starcloud satellites will operate for five years based on Nvidia chip longevity. The startup belongs to Nvidia Inception program and completed Y Combinator and Google for Startups Cloud AI Accelerator programs.
Practical Uses for Orbital AI Systems
The satellite responds to real-time queries about its position and operational data. It can report its current location over continents and predict future positioning over different regions.
Starcloud is building commercial applications with Capella Space satellite imagery. The technology could locate lifeboats from ship accidents and identify wildfire ignition points instantly.
Johnston said these systems provide immediate intelligence for emergency response teams. The satellite tracks telemetry data including altitude, orientation, geographic position, and velocity.
Users can interact with the AI model about the satellite experience. The system provides detailed responses using the high-powered model’s capabilities.
Competition Heats Up in Space Computing
Multiple companies have launched orbital data center initiatives. Google announced Project Suncatcher on November 4 to deploy solar-powered satellites with tensor processing units.
Lonestar Data Holdings develops a commercial data center for the lunar surface. Aetherflux targets a first-quarter 2027 launch for its orbital data center satellite.
Morgan Stanley analysts noted challenges including radiation damage, repair complexity, debris risks, and regulatory hurdles. Companies pursue the technology for unlimited solar access and larger-scale operations.
Starcloud’s October 2026 launch will carry several Nvidia H100 chips plus Nvidia’s Blackwell platform. The satellite will include a Crusoe cloud platform module letting customers run AI workloads from orbit.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
