Summary

• OVHcloud and CGI have secured a 30-month ESA contract to build the Digital EO information system.
• ESA expects its Earth-observation holdings to exceed 500 petabytes by 2035, increasing demand for distributed storage, compute, and AI.
• Initial infrastructure will operate across France, Germany, and Italy, with support from Poland and links to national and European systems.

OVHcloud and CGI have been selected by the European Space Agency to build a new digital foundation for Earth-observation data, putting European cloud infrastructure underneath a satellite-data estate that ESA expects to exceed 500 petabytes by 2035.

The multi-million-euro contract runs for 30 months and covers Digital EO, a decentralised information system intended to bring together storage, processing, computing, artificial intelligence, and services used across ESA’s Earth-observation programmes. CGI France Défense et Spatial will act as architect and systems integrator, while OVHcloud will provide the underlying compute, network, storage, and managed-cloud capacity.

Initial infrastructure will be deployed in France, Germany, and Italy, with support capabilities operated from Poland. The architecture is intended to connect with national and European systems and retain the option of extending into other European countries and Canada, serving scientists and public bodies as well as companies building services from satellite information.

The contract arrives as Europe builds more of the infrastructure required to use space data rather than concentrating only on launching and operating satellites. More processing is already moving closer to satellites themselves, yet information that reaches the ground still has to be stored, catalogued, combined, analysed, and made accessible across a fragmented collection of public programmes and commercial services.

Satellite capacity becomes a data problem

Earth observation creates an awkward infrastructure challenge because individual missions generate large, continuously expanding datasets that can remain valuable for years. Climate monitoring, agriculture, mapping, emergency response, environmental enforcement, infrastructure planning, insurance, and defence applications frequently depend not on one image but on comparing observations over long periods or combining them with other geospatial datasets.

That creates demand for storage, but also for compute located close enough to the data to avoid repeatedly moving enormous files between systems. AI adds another layer because image classification, object detection, change analysis, forecasting, and other machine-learning workloads need access to current observations alongside deep historical archives.

Digital EO is intended to provide a common backbone rather than forcing every programme or user to assemble those capabilities separately. OVHcloud will provide the cloud layer, while CGI’s role covers architecture, user communities, services, and continuous development across the wider system.

The scale matters commercially as well as technically. Satellite companies have spent years arguing that Earth observation can support industries far beyond the traditional space sector, but that argument becomes useful only when data are discoverable and computationally accessible without every customer becoming a specialist in satellite-ground infrastructure.

A manufacturer monitoring supply sites, an insurer assessing flood exposure, or an energy company examining vegetation around infrastructure does not necessarily need raw satellite files. It needs a service capable of turning several sources into an operational output, which makes the cloud and software layers between spacecraft and end users increasingly important to the economics of the sector.

Sovereignty reaches the application layer

The choice of European providers extends the continent’s sovereignty debate beyond government email or conventional public cloud. Earth-observation information supports environmental policy and scientific work, while some datasets and derived services also touch critical infrastructure, resilience, security, and strategic planning, giving governments a reason to care about where the computing layer sits and who controls its operation.

ESA describes Digital EO as a sovereign system, while OVHcloud has spent much of the past year building a commercial proposition around trusted European cloud services. Its SecNumCloud qualification has strengthened that route into sensitive European workloads, although winning strategic contracts still requires European providers to match the functionality, scale, and economics available from much larger global cloud companies.

Digital EO is therefore a useful test of whether sovereignty can be embedded into infrastructure without isolating the resulting system. ESA wants a decentralised architecture capable of federating with other European and national services rather than another closed repository built around a single programme.

That interoperability will determine how much value the project creates beyond ESA itself. Europe already produces large quantities of public and scientific information, but making high-value datasets easier to inspect and reuse has become a wider policy objective because their economic value depends heavily on whether businesses can actually build services on top of them.

The first measure of success will be less glamorous than the satellites feeding the system: whether Digital EO can move enormous and heterogeneous datasets between storage, compute, AI tools, and users without creating another difficult-to-navigate institutional silo. If that foundation works, Europe’s expanding Earth-observation capacity gains a clearer route from orbit into public services, research, and commercial software.