Europe has built world-leading Earth Observation (EO) infrastructures, but today’s exploitation paradigm remains fragmented and focused on data pipelines rather than integrated intelligence. As a result, transforming EO data into actionable insights is slow, difficult to scale, and poorly aligned with the needs of non-space industries. Meanwhile, global digital leaders are advancing rapidly with geospatial foundation models, world models and agentic AI systems that reason across heterogeneous data. Without a comparable European capability, much of the value generated from European EO assets risks being externalised.
To address this, the proposed activity aims to prototype a World Agentic Engine (WAE) – an agentic, AI driven capability able to orchestrate models, data, and services using a shared semantic layer. Built around multi-mission geospatial embeddings, GFMs, and world models, it will support reasoning, prediction, scenario exploration, and cross domain decision making. Rather than duplicating data, the WAE will interface with existing infrastructures and integrate both institutional and commercial actors.
A key objective is to develop and to validate the economic model of the World Agentic Engine, which will federate institutional and commercial players within a European-centric ecosystem. The system will use a usage-based transaction economical model, where each information/action/knowledge unit invocation is recorded and feeds a revenue-sharing scheme among unit owners, service operators, compute and data providers, including commercial and restricted datasets (e.g., national constellations), while taking in account the individual national data policies.
By doing so, the World Agentic Engine will strengthen Europe’s positioning in the Earth Intelligence domain, moving out from a plethora of isolated EO applications and services towards an integrated and open ecosystem for Earth Intelligence, based on European data, capable of supporting scientific discovery, commercial services, enabling policy monitoring and validation in a coherent and scalable manner, able to capture the collective value of European’s investments.
The activity will demonstrate technical feasibility and economic viability through five development pillars:
- agentic orchestration,
- system integration,
- multi-model support,
- commercialisation strategy,
- pilot use cases. A primary pilot will support FAO and national forest monitoring systems, showcasing AI driven reporting and climate finance applications. Additional pilots will explore domains where agentic EO intelligence delivers operational impact.
ESA is planning to issue an industrial Invitation To Tender (ITT) for World Agentic Engine (WAE) to demonstrate Earth Intelligence, moving from isolated applications to an integrated, scalable, sovereign European ecosystem.
Interested industrial entities and institutions are invited to participate in this Information Day.
Date: 11 May 2026, 10:00 – 11:00 CEST, online
Registration: follow the instructions on this link.