CGI’s Insula platform processes and harmonizes large volumes of Φsat-2 satellite data, enabling AI-driven trend detection, anomaly monitoring and predictive modeling for Earth observation. Since June 2025, over 3,000 users have accessed more than 10,000 multispectral images through Insula, accelerating AI model training and analytics.
CGI developed Insula in Italy to process and harmonize large volumes of satellite data, enabling advanced analytics such as trend detection, anomaly monitoring and predictive modeling. The platform integrates calibrated, geolocated datasets for scientific and commercial AI applications.
ESA’s Φsat-2 uses onboard AI for real-time cloud detection, automatic image filtering, maritime vessel classification, street map generation and image compression. Insula receives and distributes calibrated multispectral data to external developers and research teams.
Since its public launch in June 2025, Insula has onboarded over 3,000 users and delivered more than 10,000 multispectral images. This volume of real-world data supports faster AI model training and innovation across environmental monitoring and emergency response scenarios.
The collaboration reinforces CGI’s position in space-based AI services and consulting, opening potential revenue streams from data licensing and platform subscriptions. Success in Φsat-2 may lead to similar partnerships with other space agencies and commercial satellite operators.
