NASA Deploys First AI Foundation Model in Orbit

WASHINGTON — NASA has deployed its Prithvi artificial intelligence model aboard a spacecraft, making it the first AI geospatial foundation model to operate in orbit, according to NASA Science.

The deployment marks a milestone in the use of artificial intelligence for Earth observation, moving AI processing from ground-based systems to space-based platforms where satellite data is collected.

Prithvi, developed by NASA in collaboration with IBM, is a foundation model — a large-scale AI system trained on broad datasets that can be adapted for multiple downstream tasks. The model is designed specifically for geospatial analysis, processing satellite imagery and Earth observation data to support scientific research and environmental monitoring.

Foundation models represent a shift in how AI systems are built. Rather than training specialized models for individual tasks, foundation models learn general representations that can be fine-tuned for specific applications such as flood detection, wildfire monitoring, crop assessment, and climate analysis.

By operating directly in orbit, Prithvi can process Earth observation data closer to where it is collected, potentially reducing the time and bandwidth required to transmit raw satellite imagery to ground stations for analysis. This edge-computing approach could accelerate response times for natural disaster monitoring and other time-sensitive applications.

NASA has positioned the Prithvi model as an open-science initiative. Earlier versions of the model were released as open-source through partnerships with IBM and Hugging Face, making the technology available to researchers worldwide.

The orbital deployment adds to a growing portfolio of U.S. government AI initiatives, as federal agencies increasingly adopt foundation models for mission-critical applications beyond traditional enterprise use cases.