OVHcloud and CGI to build ESA Earth data hub

The European Space Agency (ESA) has selected OVHcloud and IT services firm CGI to construct a cloud-based platform for managing Europe’s Earth observation data. The 30-month contract, valued at several million euros, tasks OVHcloud with providing the underlying infrastructure while CGI handles system design and integration. Neither company has disclosed specific capacity figures, contract value, or projected operating costs for the project, known as Digital EO.

The platform aims to consolidate storage, computing, and AI capabilities for researchers, public agencies, and commercial users. Rather than replacing existing national or European data repositories, Digital EO will interconnect them, requiring robust interoperability to allow seamless data discovery, transfer, and analysis. ESA anticipates its Earth observation holdings will exceed 500 petabytes by 2035, though this figure encompasses its broader data assets, not the storage directly tied to the current contract. The scale underscores the need for efficient data handling, as satellite archives only deliver value when users can locate and process relevant datasets quickly and affordably.

Infrastructure and operational challenges

Digital EO will operate across data centers in Italy, France, and Germany, with support functions based in Poland. This geographic distribution aligns with OVHcloud’s role in European public-sector cloud initiatives, where jurisdictional control and infrastructure sovereignty are priorities. However, spreading the platform across multiple countries introduces complexity in access policies, network performance, resilience, and operational accountability. The companies have not detailed how data will be replicated, which workloads will run in which locations, or what recovery time objectives will apply.

CGI’s responsibilities as system architect and integrator are equally critical. While storage and compute form the foundation, the platform must also bridge existing facilities and present usable services to external users. Inconsistent interfaces or access rules between repositories could undermine the benefits of additional capacity. Developers and institutional users will likely seek clarity on APIs, identity management, data formats, and pricing before committing to build applications on Digital EO.

Strategic context and industry implications

The contract reflects Europe’s broader push to host strategic public workloads with regional cloud providers. It tests whether these providers can combine jurisdictional control with the performance and scalability required for large, distributed datasets. Once operational, Digital EO’s success will hinge on data accessibility, throughput, reliability, and cost—metrics that remain unspecified in public disclosures. Researchers and agencies will evaluate the platform’s ability to reduce transfer bottlenecks, though the actual savings will depend on workload patterns and the final system design.

For OVHcloud, the project offers a high-profile opportunity to demonstrate its capabilities in a sector where sovereignty and compliance are key considerations. CGI, meanwhile, faces the challenge of integrating disparate systems while ensuring the platform meets the needs of diverse users, from academic researchers to government agencies. The lack of public details on technical benchmarks leaves open questions about how the system will perform under real-world conditions.

What to watch

As the project progresses, industry observers will monitor several areas. First, the technical implementation—how effectively the platform connects existing repositories and handles cross-border data flows. Second, the operational model—whether the distributed setup introduces latency or complicates compliance. Finally, user adoption—whether researchers and agencies embrace Digital EO as a viable alternative to existing solutions, particularly if pricing or performance falls short of expectations. The 30-month timeline suggests initial deployments could begin within the next year, though full-scale rollout may take longer.