Periodic Reporting for period 2 - SEDIMARK (SEcure Decentralised Intelligent Data MARKetplace)
Período documentado: 2024-04-01 hasta 2025-09-30
Context and Objectives
The European data economy is rapidly expanding and is expected to exceed €800 billion by 2025, driven by IoT, digital services, and AI. Data has become a strategic asset for innovation, competitiveness, and the digital and green transitions. Data marketplaces, projected to reach €100 billion, are key enablers but remain largely ineffective due to centralised architectures that limit data sovereignty, raise privacy and trust concerns, provide weak guarantees on data quality and interoperability, and exclude many organisations.
An analysis of 155 European data-sharing initiatives shows that only 15% have viable revenue models, with most relying on unstable public funding. Data providers face compliance risks, consumers struggle to find trustworthy datasets, and the lack of standardised connectors and quality assurance hinders adoption.
At the same time, EU initiatives such as the European Data Strategy, Data Governance Act, Data Act, AI Act, and Common European Data Spaces call for decentralised, interoperable, privacy-preserving, and trustworthy data sharing. This creates an urgent need for advanced platforms supporting federated AI, decentralised trust, and transparent governance.
SEDIMARK addresses this need by designing and validating a secure, decentralised marketplace for data and AI services, combining Distributed Ledger Technologies (DLT), Self-Sovereign Identity (SSI), semantic interoperability, and energy-efficient Green AI.
The European data economy is rapidly expanding and is expected to exceed €800 billion by 2025, driven by IoT, digital services, and AI. Data has become a strategic asset for innovation, competitiveness, and the digital and green transitions. Data marketplaces, projected to reach €100 billion, are key enablers but remain largely ineffective due to centralised architectures that limit data sovereignty, raise privacy and trust concerns, provide weak guarantees on data quality and interoperability, and exclude many organisations.
An analysis of 155 European data-sharing initiatives shows that only 15% have viable revenue models, with most relying on unstable public funding. Data providers face compliance risks, consumers struggle to find trustworthy datasets, and the lack of standardised connectors and quality assurance hinders adoption.
At the same time, EU initiatives such as the European Data Strategy, Data Governance Act, Data Act, AI Act, and Common European Data Spaces call for decentralised, interoperable, privacy-preserving, and trustworthy data sharing. This creates an urgent need for advanced platforms supporting federated AI, decentralised trust, and transparent governance.
SEDIMARK addresses this need by designing and validating a secure, decentralised marketplace for data and AI services, combining Distributed Ledger Technologies (DLT), Self-Sovereign Identity (SSI), semantic interoperability, and energy-efficient Green AI.
Vision and Technical Contribution
SEDIMARK replaces data centralisation with a distributed registry where data and AI services remain locally stored at the edge. Assets are cleaned, validated, enriched, labelled, certified, and anonymised before sharing, ensuring privacy-by-design and reduced data movement.
The platform enables discovery and reuse of heterogeneous datasets following the FAIR principles, leveraging Smart Data Models and NGSI-LD for cross-domain semantic interoperability. DLT provides tamper-proof logging and transparent governance, while Decentralised Identifiers (DIDs) and Verifiable Credentials (VCs) enable trusted onboarding.
Green AI techniques automate data quality management and distributed analytics, achieving 20–30% reductions in energy consumption without degrading model performance.
The system builds on consortium technologies starting at TRL 5 and was validated through four operational pilots:
* Mobility Digital Twin (Helsinki) for urban traffic modelling and planning
* Urban Bicycle Mobility (Santander) for privacy-preserving transport analytics
* Energy Consumption Analytics (Greece) using federated learning on sensitive customer data
* Water Data Valorisation (France) for environmental monitoring and data monetisation
By project end, SEDIMARK reached TRL 7, demonstrating readiness for European Data Spaces.
Main Achievements
SEDIMARK delivered:
* A decentralised, GAIA-X and IDSA-compliant marketplace architecture integrating catalogue, connector, offering manager, and DLT trust services
* A common ontology and AI-driven toolset for data profiling, enrichment, synthetic data generation, and quality annotation
* A production-ready data pipeline with automated quality assessment and Green AI optimisation
* Distributed AI frameworks enabling federated and gossip learning while preserving data sovereignty
* A secure access and governance framework based on SSI, smart contracts, anonymisation, and GDPR-compliant controls
* A ready-to-market platform with monetisation models, white-label deployment, and open-source adoption pathways
All components were fully integrated into a unified platform and validated across the four pilots.
SEDIMARK replaces data centralisation with a distributed registry where data and AI services remain locally stored at the edge. Assets are cleaned, validated, enriched, labelled, certified, and anonymised before sharing, ensuring privacy-by-design and reduced data movement.
The platform enables discovery and reuse of heterogeneous datasets following the FAIR principles, leveraging Smart Data Models and NGSI-LD for cross-domain semantic interoperability. DLT provides tamper-proof logging and transparent governance, while Decentralised Identifiers (DIDs) and Verifiable Credentials (VCs) enable trusted onboarding.
Green AI techniques automate data quality management and distributed analytics, achieving 20–30% reductions in energy consumption without degrading model performance.
The system builds on consortium technologies starting at TRL 5 and was validated through four operational pilots:
* Mobility Digital Twin (Helsinki) for urban traffic modelling and planning
* Urban Bicycle Mobility (Santander) for privacy-preserving transport analytics
* Energy Consumption Analytics (Greece) using federated learning on sensitive customer data
* Water Data Valorisation (France) for environmental monitoring and data monetisation
By project end, SEDIMARK reached TRL 7, demonstrating readiness for European Data Spaces.
Main Achievements
SEDIMARK delivered:
* A decentralised, GAIA-X and IDSA-compliant marketplace architecture integrating catalogue, connector, offering manager, and DLT trust services
* A common ontology and AI-driven toolset for data profiling, enrichment, synthetic data generation, and quality annotation
* A production-ready data pipeline with automated quality assessment and Green AI optimisation
* Distributed AI frameworks enabling federated and gossip learning while preserving data sovereignty
* A secure access and governance framework based on SSI, smart contracts, anonymisation, and GDPR-compliant controls
* A ready-to-market platform with monetisation models, white-label deployment, and open-source adoption pathways
All components were fully integrated into a unified platform and validated across the four pilots.
SEDIMARK advances the state of the art by delivering the first fully decentralised, feeless, and standards-compliant data marketplace combining DLT-based trust, semantic interoperability, and Green AI at scale. Key breakthroughs include:
* Operational privacy-preserving AI with federated learning on 5,000+ real users
* Automated semantic onboarding reducing marketplace entry from hours to minutes
* 20–30% energy savings in AI processing across pilots
* Edge-native AI enabling local processing, reduced bandwidth, and extended IoT device lifetime
These results address critical barriers to data space adoption: trust, interoperability, quality, energy efficiency, and monetisation.
* Operational privacy-preserving AI with federated learning on 5,000+ real users
* Automated semantic onboarding reducing marketplace entry from hours to minutes
* 20–30% energy savings in AI processing across pilots
* Edge-native AI enabling local processing, reduced bandwidth, and extended IoT device lifetime
These results address critical barriers to data space adoption: trust, interoperability, quality, energy efficiency, and monetisation.