Satellite Uses AI Vision Model to Identify Objects Autonomously
Original: A satellite just learned to find things on its own — here’s what that means
Why This Matters
First orbital AI vision model demonstrates autonomous satellite analysis capabilities
Earth observation satellite Yam-9 successfully used Google DeepMind's Gemma 3 vision-language model to identify areas of interest without human analysts. The milestone demonstration occurred in April 2026, marking first reported orbital use of VLM technology for autonomous satellite analysis.
Loft Orbital's Yam-9 satellite achieved a space industry first by using AI to autonomously identify targets in response to natural language queries. The spacecraft used Google DeepMind's Gemma 3 vision-language model, powered by Nvidia Jetson Orrin AGX GPU, to classify areas where natural environments meet human development and identify railway infrastructure. NASA JPL developed NAVI-Orbital software to enable the VLM operation. The breakthrough could revolutionize satellite operations by performing initial data analysis in orbit, reducing the massive data streams currently requiring ground-based human analysis. Loft's head of AI Paul Lasserre envisions 'always-on patrol layers' enabling satellites to monitor borders and alert operators to suspicious activity through interactive dialogue.