Horizon Europe Space 2026 – Scientific Analysis & Exploitation of Space Data: The Algorithm-as-a-Service Mandate

1. The Exploitation Gap – Why Europe’s Copernicus and Galileo Data Are Underutilised

Europe has built the world’s most sophisticated civil space infrastructure. The Copernicus programme operates the Sentinel fleet, generating over 20 petabytes of data annually. Galileo provides unmatched centimetre-level positioning accuracy. By 2026, the European space data archive will exceed 150 petabytes.

Yet a 2025 study by the European Space Policy Institute (ESPI) found that only 17% of Copernicus data downloaded is used in any analytical workflow beyond a single visualisation. Less than 4% is combined with other data sources. The remaining 96% of analytical effort is duplicated – the same basic processing is performed thousands of times by different research groups. This is the exploitation gap.

The Horizon Europe Space 2026 – Scientific Analysis & Exploitation of Space Data call directly targets this gap. With a total budget of €78 million across four topics, the call shifts funding from data generation to data exploitation. The call introduces three structural innovations: mandatory algorithm-as-a-service delivery, federated processing, and reproducibility-by-design. A technically brilliant algorithm that requires a PhD in remote sensing to operate will score poorly. An algorithm packaged as a REST API that a climate economist can call from Python will score highly.

2. 2026 Call Architecture: Four Topics, One Mandatory Exploitation Focus

All topics share a mandatory adherence to the Copernicus Data Access and Exploitation Framework, requiring that all funded software be deployable on the Copernicus Data Space Ecosystem (CDSE) cloud platform.

Critical requirement for 2026: All Topic 01-01 and 01-02 proposals must include a data processing budget that runs on the CDSE. The Commission will no longer fund proposals that download large datasets to local infrastructure for processing.

The call also introduces a 9-level Exploitation Readiness Level (ERL) scale. Your proposal must specify starting ERL and target ERL (e.g., minimum 7 for Topics 01-01). Service usage by external groups defines higher ERLs.

3. The Algorithm-as-a-Service Mandate: From Notebook to API

Previous space data exploitation calls funded algorithms as scientific software (notebooks/Git repos). The 2026 call funds services. An Algorithm-as-a-Service (AaaS) offering must include:

• Element 1: A publicly accessible API endpoint accepting standard geospatial queries (GeoJSON) and returning standard formats.
• Element 2: Interactive API documentation (Swagger UI or Redoc) and client libraries (Python, R).
• Element 3: Automatic logging of API calls and user feedback loops.
• Element 4: A sustainability plan for hosting beyond the project life.

4. Mini Case Study: Project SENSEIVER (2024-2026)

How a river monitoring API turned Copernicus data into an operational water quality service

Project SENSEIVER was funded to develop an automated service estimating water quality parameters from Sentinel-2/3 data for European rivers.

The processing pipeline ran entirely on CDSE infrastructure. The output API accepted a river reach identifier and returned time-series parameters. The algorithm was served as a REST API via FastAPI, deployed on a CDSE Kubernetes cluster. Python and R client libraries were open-sourced on PyPI and GitHub.

Results:

• Reduced time for water authorities to obtain a water quality estimate from 2-3 weeks to 8 seconds (>99% reduction).
• Increased the number of repeatedly monitored river reaches by 1,740%.
• Used by 34 external user organisations generating over 14,000 monthly API calls.

The critical learning: SENSEIVER initially deployed a complex OAuth2 authentication system. Water authorities failed to integrate it due to strict IT policies. The team pivoted to a simpler API key model with IP whitelisting, increasing adoption dramatically.

5. Mandatory Data Fusion Requirements: Copernicus + Galileo + In-Situ

Topic 01-02 requires fusion of at least two of the three data families. The call specifies three non-negotiable fusion requirements:

1. Spatiotemporal alignment at the data level: Fusion must occur at the pixel/measurement level, not by overlaying pre-processed maps.
2. Uncertainty propagation: Your fusion algorithm must propagate input uncertainties to a final confidence interval.
3. Missing data handling: Your algorithm must explicitly handle missing data (clouds blocking optical sensors, urban canyons blocking signals).

6. Reproducibility-by-Design: The New Standard for Topic 01-04

Topic 01-04 is mandated to develop platforms that make space data analyses reproducible by default. The Commission requires platforms to enable one-click replication of published results within 5 minutes, executable computational notebooks, versioned data access tied to CDSE persistent identifiers, and a reproducibility badge system.

Designing a Horizon Europe Space 2026 proposal that satisfies the algorithm-as-a-service mandate, includes a credible uncertainty quantification work package, and achieves a target Exploitation Readiness Level of 7 or above is a specialised skill. Ensure your user onboarding work package is robustly designed to score top marks under the 2026 call rubric.

---

Strategic Verification for 2026

This analysis has been cross-referenced with the Intelligent PS Strategic Framework. It is intended for organizations seeking high-performance bid assistance. For technical inquiries or partnership opportunities, visit Intelligent PS Corporate.