FondationFAM Numérique 5ème cycleNumérique 5ème cycle
DonateDons
Chapitre 2Chapter 2 Part 1 · European digital strategies Partie 1 · Stratégies numériques européennes

Data spaces

Translated from the French, which remains the reference text. This translation is under review.

2.1The context: the capture of the data economy and the systemic risk

The emergence of Data Spaces stems from a geopolitical and macro-economic diagnosis made by European actors at the beginning of the 2010s.

In the economy of the twenty-first century, data appears as a key factor of production. Yet Europe observes that the capture of this resource and its valorisation are asymmetric. The dominant model is that of the centralised cloud and of the platforms ("hyperscalers", mostly American and Chinese). For European industry, sharing its industrial data on these platforms means exposing itself to vendor lock-in (proprietary lock-in) and losing control of its trade secrets. Europe risks becoming a mere "digital colony", supplying the raw material (raw data) to third parties who sell it back the finished product (Artificial Intelligence or analytics).

The European Commission, in its founding text of February 2020 (A European strategy for data - COM(2020) 66 final), formally sets out this diagnosis and calls for the invention of an alternative model, a European « third way »:

"Over the last few years, digital technologies have transformed the economy and society, affecting all sectors of activity and the daily lives of all Europeans. Data is at the centre of this transformation and more is to come. [...] In a society where individuals will generate ever-increasing amounts of data, the way in which the data are collected and used must place the interests of the individual first, in accordance with European values, fundamental rights and rules. [...] In order to release Europe’s potential we have to find our European way, balancing the flow and wide use of data, while preserving high privacy, security, safety and ethical standards."

This context is also fed by a physical and technical constraint, summed up by Gaia-X. The exponential growth of volumes (IoT, industrial sensors) makes the model of centralisation obsolete:

"Data is growing at a meteoric rate; the total amount of data generated by 2025 is set to accelerate exponentially to 175 zettabytes. [...] Today, the challenges associated with computing and the use of the cloud defy standard criteria... Advanced manufacturing processes, supply chain management, new logistics systems, modern software and large data ecosystems require a completely new approach." (Gaia-X brochure).

2.2The genesis: from the Fraunhofer laboratories to continental public policy

Data Spaces come from European applied industrial research. In 2014, to respond to the digitisation of the complex supply chains of the Mittelstand, the German industrial fabric, within the framework of Industrie 4.0, Boris Otto and the researchers of the Fraunhofer Institute ask themselves the founding question: how can data be shared between several companies without storing it with a centralised trusted third party? The research project is then named Industrial Data Space. From its 2017 reference architecture document (Reference Architecture Model for the Industrial Data Space), Fraunhofer sets down the DNA of the concept:

"The Industrial Data Space is a virtual data space leveraging existing standards and technologies, as well as accepted governance models, to facilitate the secure exchange and easy linkage of data in a trusted business ecosystem. It thereby provides a basis for smart service scenarios and innovative business processes, while at the same time ensuring data sovereignty for the participating data owners."

Understanding that this need is not confined to the factory but applies to health, finance or energy, the initiative is renamed International Data Spaces. In 2016, the International Data Spaces Association (IDSA) is created to bring industrial players together and to standardise this work in the form of an open standard.

The real change of scale comes on 19 February 2020, when Ursula von der Leyen and Thierry Breton take up this German-European R&D project to make it the official doctrine of the European Union:

"The European strategy for data aims at creating a single market for data that will ensure Europe's global competitiveness and data sovereignty. This will lead to the creation of Common European Data Spaces. They will ensure that more data becomes available for use in the economy and society, while keeping the companies and individuals who generate the data in control." (European Commission, 2020).

2.3The original vision: decentralisation, sovereignty and federation

The original vision is a radical paradigmatic break with the concept of the centralised "Data Lake". In a Data Space, data does not move by default; it stays at the source. What travels are the usage contracts, the encryption keys and the analytical queries. This vision rests on three conceptual pillars, documented by the three key actors of Data spaces: data sovereignty, the doctrine of the IDSA; federation without a central cloud, the mission of Gaia-X; the canonical definition, laid down by the DSSC.

Data sovereignty, first, is the watchword. It is the right to technical and legal self-determination:

"Data sovereignty is a central aspect of the Industrial Data Space. It can be defined as a natural person's or corporate entity's capability of being entirely self-determined with regard to its data." (Fraunhofer RAM, 2017).
"Data sovereignty refers to who has control and responsibility over the data along a secure end to end data supply chain – and the rules to be accepted and enforced. Data sovereignty is a spectrum, meaning there can be different degrees of control. [...] We want to pave the way for a data economy in which every company and every person keeps full control over their data treasures." (IDSA, Vision).

The concept of “sovereignty” is thus applied to individuals and to organisations. It is not a matter of “just” repatriating data into a European cloud, as many actors and discourses currently suggest.

To guarantee this sovereignty, next, the underlying infrastructure must be decentralised. This is the essence of the Gaia-X initiative, founded in 2019:

"The architecture of the Industrial Data Space does not require central data storage capabilities. Instead, it follows a decentralized approach, which means that data physically remains with the respective data owner until it is transmitted to a trusted party. Thus, the Industrial Data Space is not a cloud." (Fraunhofer RAM).
"Gaia-X strives for innovation through digital sovereignty. [...] Our intention is that we give the control back to the users by retaining sovereignty over their data. Our outcome will not be a cloud. It is a federated system linking many cloud service providers and users together in a transparent environment that will drive the European data economy of tomorrow." (Gaia-X, General Overview).

The Data Spaces Support Centre (DSSC), finally, the entity mandated by Europe to unify the ecosystem, has laid down the canonical definition, taken from the Blueprint v1.0/2.0:

"A data space is a distributed system defined by a governance framework. It facilitates secure and trustworthy data transactions between participants, emphasising trust and data sovereignty. A data space supports one or more use cases."

2.4The implementation strategy: legislation, architecture and orchestration

To impose this vision in the face of the GAFAM and the free market, Europe has deployed a legal, technical and financial strategy unique in the world: the law to compel sharing (Data Act, DGA), technical standards (IDSA, Gaïa-X), and heavily subsidised support (more than 750 million euros at the very least since 2021). This ambitious arrangement has three parts.

Legislative engineering, first. The EU creates an unprecedented legal framework to unlock the data economy safely:

"Two pieces of legislation have been put in place to protect the rights and interests of citizens while simultaneously fostering industrial and technological development. The Data Governance Act (DGA) is a comprehensive tool designed to oversee the reuse of publicly or protected data across various sectors. It aims to facilitate data sharing by regulating new entities known as data intermediaries... The Data Act entered into force on 11 January 2024 and into application on 12 September 2025. It is a pillar of the European strategy for data. Its main objective is to make Europe a leader in the data economy by harnessing the potential of the ever-increasing amount of industrial data." (European Commission).

The technological matrix, next, is shared between two actors. To avoid redundancies, the technical roles have been shared out, notably through the DSBA alliance (Data Spaces Business Alliance): Gaia-X takes care of the infrastructure, of decentralised identity (wallets) and of the trust framework (Trust Framework); the IDSA takes care of the pure transactional exchange protocols, through the Dataspace Protocol (DSP) and its connectors:

"The IDSA Rulebook lays the foundation for trusted data sharing within data spaces, implemented through the IDS Architecture. It serves as a comprehensive guide, covering all the essential functional, technical, operational, and legal requirements needed to design and operate data spaces based on IDS." (IDSA).

Orchestration, finally, and the common "Blueprint". To prevent the different spaces (mobility, health, energy) from developing incompatible standards, thereby creating new silos, Europe funds the DSSC to draft the "architect's plan" (The Blueprint) and funds SIMPL, an open-source middleware:

"The Blueprint is a consistent and comprehensive set of guidelines to support the development cycle of data spaces. It includes the conceptual model of a data space, data space building blocks, and recommended standards and specifications." (DSSC, Blueprint).
"Simpl is the open-source smart middleware platform that enables cloud-to-edge federations [...]. Simpl is the common software behind common data spaces." (European Commission).

2.5The assessment to date (2024-2025):

The valley of death of "Scaling"

Today, the ecosystem is going through a critical phase: the transition between theory (the Proofs of Concept) and real industrialisation (Scaling).

A validated institutional structure

In its Second Staff Working Document on common European data spaces (January 2024), the Commission validates the creation of more than 14 major thematic spaces funded by the Digital Europe programme:

"This SWD presents the state of play of data-space related EU support actions, including sector- and domain-specific initiatives, the Data Spaces Support Centre (DSSC) and smart open-source middleware (Simpl)..." (EC, January 2024).

Colossal projects of 250 M€ such as Catena-X (the automotive industry, which connects thousands of SMEs to Volkswagen or BMW) or the EHDS (European Health Data Space, which was the subject of a specific regulation in 2024) prove the viability of the concept. The IDSA's connectors now go through strict certification phases (Technology Compatibility Kit).

The diagnosis from the field: the barriers to adoption (DSSC)

However, mass adoption comes up against the roughness of economic reality. The DSSC, in its analyses of the maturity of the initiatives, lists in black and white the current systemic barriers:

"European data spaces are therefore increasingly moving from concept to operational reality. Across sectors, work on technical infrastructure, interoperability, and governance is moving from pilots to deployment... While robust infrastructures and common standards are essential, the true economic and societal potential of European data spaces depends on the ecosystem's ability to scale." (DSSC).

The barriers identified by the DSSC are, in its words, six:

“Fragmentation across legal, technical, governance, and business domains, creating silos and undermining interoperability. Differences in national legislation, inconsistent technical standards, and fragmented governance models complicate cross-sector and cross-border collaboration, eroding trust and slowing adoption. Inflexible and immature infrastructures, which result in high entry and integration costs. Many data spaces rely on legacy systems, early-stage technologies, or rigid organizational models, making onboarding and adaptation difficult, particularly for SMEs and smaller public actors. Unclear incentives for participation, which limit stakeholder engagement. Without tangible, immediate benefits — or aligned long-term value propositions — actors may hesitate to contribute or share data, stalling the creation of critical mass necessary for effective collaboration. Trust and data sovereignty concerns, which affect willingness to share and reuse data. Legal uncertainty, complex licensing, and unclear liability allocation amplify perceptions of risk, while gaps in governance and technical safeguards may further inhibit confidence in data-sharing arrangements. Uncertain evolution pathways, where unclear strategies for scaling, sustaining, and adapting data spaces increase operational and investment risks. Initiatives often depend on public subsidies or pilot funding, and ambiguity around future technical, governance, and regulatory developments makes long-term planning difficult. Market barriers, including structural challenges that limit the entry, growth, or sustainability of new or smaller actors. These challenges encompass not only economic factors but also legal, technical, and cultural dimensions that can discourage diverse participation and reduce ecosystem inclusivity.”

The Digital New Deal report, presented below, draws the figures from it:

"Recent maturity evidence is sobering. In its January 2024 Maturity assessment report, the Data Spaces Support Centre (DSSC) reports that [...] under strict development-cycle criteria, 64% data space initiatives were still preparatory, 22% exploratory, and 14% implementation, with 0% classified as operational or scaling."

The uncompromising report of the Digital New Deal

The finding set out by the Digital New Deal report, currently being published, entitled Data ecosystems interoperability : Building union from fragmentation - Turning Data Space islands into a European Data & AI continent is unequivocal: Europe has generated a massive impetus (momentum), but suffers from a cruel absence of articulation. The European Union is financing sectoral silos that are very well regulated, but incapable of communicating with one another fluidly, and of scaling up.

The paradox of proliferation without scale (0% scaling up)

The figures are a warning. Although Europe counts more than 145 initiatives (on the IDSA's radar), almost none has reached the stage of industrialisation. The report points to a false indicator of success: the multiplication of initiatives is not synonymous with the creation of a single market for data.

"Europe has reached a new phase: the problem is no longer initiating data spaces; it is scaling them into something that behaves like a 'single market' infrastructure. [...] Proliferation is not scale. A large number of initiatives can still produce a small number of operational, reusable, cross-border data transactions..."

The archipelago trap ("The Archipelago Trap") and the tax on SMEs

As things stand, Data spaces are not interoperable. The European Data Spaces develop as isolated islands. For each new connection to a different ecosystem, companies (and above all SMEs) must pay an "integration tax": re-signing contracts, redoing semantic mappings, re-integrating new connectors.

"If that interconnection does not happen fast enough, the outcome is not a 'plural Europe'. It is an archipelago: many small, internally coherent networks that remain externally costly to use together."

The consequence is direct for the European economic fabric:

"For SMEs, the interoperability question is existential. [...] Fragmentation is a tax on participation. If each data space, sector, or Member State requires bespoke onboarding, bespoke contracts, bespoke semantics and bespoke technical interfaces, the result is predictable: only the biggest actors can afford the integration burden and overhead, everyone else becomes a spectator."

The technological "War of the Stacks" (Stack Wars)

Despite the efforts of standardisation, the ecosystem has divided into technological chapels (Eclipse EDC, FIWARE, Prometheus-X, Pontus-X...). Even though many of them use the same connectors, the adaptations they make cause the implementations to diverge and break interoperability. The actors optimise the inside of their Data Space, but neglect the bridges to the outside.

"Data spaces increasingly behave like competing ecosystems. Competing stacks, competing models, and competing vocabularies create 'data spaces stack war', with too few live cross-data spaces use cases, and too much optimisation inside each individual data space."

The strategic danger is identified: if the European solutions refuse to interconnect, actors will turn by default to the dominant Cloud players, which offer integrated "turnkey" solutions.

"If Europe settles for 'siloed data spaces', foreign cloud-data-AI platforms will supply the missing integration layer and capture network effects."

The limit of regulation by Law

Europe believed, wrongly, that the Data Act and the Data Governance Act would be enough to harmonise the market. This point will be confirmed in the second and third parts of the report: if the law creates trust, it is not enough to create operational technology standards (the move from "PDF compliance" to "compliance-as-code").

"Regulation helps but does not converge implementations. [...] EU law increasingly frames what 'trustworthy data sharing' should look like, but it does not automatically make implementations converge. Even where regulation provides strong building blocks [...] the legal layer rarely dictates a single technical profile, onboarding journey, credential model, or reference implementation."

2.6The future of Data Spaces: Artificial Intelligence and a world standard

The recent publications (2024-2025) of the Commission and of the standardisation bodies trace three vital strategic horizons for Data Spaces: artificial intelligence, the single market for data, the world standard.

The unavoidable fuel of Artificial Intelligence

It is the pivot of the decade. Europe has acknowledged that its ambitions in AI (and in sovereign generative AI models) will depend on its capacity to train these algorithms on high-quality industrial data (B2B), sheltered from copyright problems (unlike the wild "scraping" of the Web). Data Spaces are the only secure channel to feed European AI.

"Specifically in the rapidly evolving field of generative AI, the availability of high-quality data is paramount for fostering innovation and ensuring Europe's competitiveness in the world. Foundation models [...] depend on a vast amount of data to train themselves. [...] In this light, common European data spaces play a critical role in providing a steady supply of data, including language data, to economic actors, thereby creating a fundament for AI innovation in Europe." (European Commission, Staff Working Document 2024).

In this framework, the EC is currently deploying its future Data Union Strategy, whose explicit aim is "to bolster AI innovation by enhancing access to high-quality, curated data across industries".

The Single Market for Data (cross-sector interoperability)

The future lies not in "sectoral silos", but in their fusion. An Energy Data Space must be able to communicate seamlessly with a Mobility Data Space in order, for example, to optimise the charging of electric vehicles on the scale of a continent.

"Together, the data spaces will gradually be interconnected to form the single market for data." (European Commission).

The quest for world hegemony (the ISO ambition)

The final objective of the IDSA and of Gaia-X is to replicate what Europe achieved with the GDPR ("the Brussels Effect"), but on the technical plane. The ambition is to make the Dataspace Protocol (DSP) the world ISO norm for data transfer, in the same way that TCP/IP or HTTP became it for the Internet of information.

"As the DSP [Dataspace Protocol] moves towards becoming an international standard, it promises to revolutionize data sharing by ensuring interoperability and bridging diverse systems. [...] IDSA's goal is to make the DSP an ISO Standard. By working with the Eclipse Foundation, IDSA accesses a vast network of open-source developers, accelerating the development of data spaces and contributing to a global scale of trusted data sharing and digital sovereignty." (IDSA).

According to the Digital New deal

Faced with this critical assessment, the future of Data Spaces no longer runs through grand declarations of intent, but through resolute market engineering. The document traces a pragmatic roadmap for the coming decade.

The indispensable fuel of Agentic Artificial Intelligence

The future of Data Spaces is inseparable from AI, but not just any AI: Agentic AI (systems capable of acting, planning and negotiating autonomously). Yet an agentic AI cannot read PDF contracts; it needs governance encoded in computer form (machine-readable).

"Agentic AI changes the technical requirement: 'AI that acts' needs machine-readable governance. [...] Agentic AI does not 'consume datasets' occasionally. It needs continuous, reliable access across many actors. In practice, agentic AI requires: machine-readable discovery, machine-to-machine negotiation, machine-enforceable permissions, traceability by default."

The interoperability of Data Spaces becomes the sine qua non condition of European sovereignty in AI:

"No scalable AI without data space interoperability. [...] There is no trustworthy AI without trustworthy data, and no competitive and agentic AI without interoperable, actionable, capable and governable data."

The "GSM Moment" of data: imposing an interoperability core

It is the central concept of the report for answering American domination (Hyperscalers, Databricks, Palantir...) and Chinese acceleration (where data is now capitalised on the balance sheets of companies). To survive, the data economy must take its inspiration from mobile telephony (the GSM norm of the 1990s): to oblige all actors, technically and politically, to use a minimal common standard.

"Europe needs a 'GSM moment' for data & AI: a political and industrial commitment to a minimal, enforceable interoperability core that is common, measurable and reusable across sectors and Member States."

To get there, the report introduces the doctrine of "MUST vs SHOULD" and of the coopetition pact ("Cooperate on rails, compete on services"):

"The coopetition deal: cooperate on rails, compete on services. [...] The bargain: avoid fragmentation without creating a monopoly. [...] Rails are the parts of the system where fragmentation destroys scale and raises entry barriers (identity, discovery, policies, trust protocols). [...] Services are where competition is healthy and value is created (UX, performance, AI tooling)."
"MUST is the anti-fragmentation minimum, not a unification project. [...] The core should be small enough to be adopted, but strong enough to make a market."

Convergence from the base: the DSIF and the Hypercatalogue

Faced with the slowness of top-down standardisation (coming from above), the immediate future is being written bottom-up (from the field). The report highlights the DSIF (Data Space Interoperability Framework) initiative, in which the very creators of the competing technologies (EDC, Prometheus-X, FIWARE, Simpl...) have allied to build concrete bridges.

"The most concrete answer that has emerged at ecosystem level to the fragmentation [...] is the Data Space Interoperability Framework (DSIF) a community-driven initiative whose stated purpose is to make Europe’s heterogeneous data space stacks interoperate in production, rather than coexist on paper."

This convergence is embodied notably in the creation of a Hypercatalogue to decompartmentalise the search for data on a European scale:

"A user, a developer or a public buyer wanting to assemble a service today must navigate each marketplace separately; the same query, run across the aggregated offer, returns a markedly richer set of options. [...] The right answer is federation: each data space catalogue should be able to surface offerings published in the others..."

The "Social Network" of Data Spaces: human interoperability

It is the most innovative proposal of the report. The ecosystem has become so complex that technology is no longer enough. If the actors do not communicate, the standards will never be adopted. A "knowledge graph" must be created, linking humans, use cases and legal rules.

"Data spaces are not only made of protocols. They are still made by people. [...] The final bold proposal of this report is simple: make the data space community interoperable with itself."
"Europe should collectively build the social network and knowledge hub of the data space community. [...] A shared, open, federated, machine-readable social and knowledge graph for data spaces: a common rail connecting practitioners, organisations, projects, federations, standards, rulebooks, rolebooks, components..."

The ultimate horizon: the metamorphosis into "Smart Data Ecosystems"

In conclusion, the report acknowledges that the very term "Data Space" (which echoes too closely a mere space of storage or of technical exchange) is reaching its limits. The future belongs to a more encompassing concept, merging data, algorithms (AI), computing infrastructures and human governance:

"From Data Spaces to Smart Data Ecosystems. In practice, what Europe is building is not merely a federation of data infrastructures. It is the foundation of a new generation of Smart Data Ecosystems. The term data space may now have reached its limits: it does not fully capture AI, governance, trust, business models or the human cooperation needed to make the system work. [...] The second season of data spaces is not a technical adjustment. It may become the second act of Europe’s digital project."
Mission carried by the Fondation FAM. Mission portée par la Fondation FAM. ← Back to mission homeRetour à l'accueil de la mission www.fondation-fam.org — Chapter 2 — Data spaces — published 2026-07-10 · a chapter of its own since 2026-10-14 · updated 2026-10-29 www.fondation-fam.org — Chapitre 2 — Les data spaces — publié le 10 juillet 2026 · chapitre à part entière depuis le 14 octobre 2026 · mis à jour le 29 octobre 2026