Data & AI as Commons
GDPR as a foundation of trust, yet Gaia-X and data spaces are advancing only slowly.
Where does Germany stand, and where does it break?
Europe regulates the handling of data through the General Data Protection Regulation (GDPR, in force since 2018) and complementary legislation (the Data Act and the Data Governance Act). Germany sought to build a federated, sovereign data infrastructure through Gaia-X. The Gaia-X funding programme, commissioned by the Federal Ministry for Economic Affairs and Climate Action (BMWK) and administered by the Federal Network Agency, covered eleven projects worth approximately €132m. It concluded at the end of March 2025 and produced, among other things, data spaces for healthcare, mobility and construction, as well as the multilingual language model Teuken-7B. The commercial base, however, remains in foreign hands. The three US hyperscalers — Amazon, Microsoft and Google — together hold roughly two-thirds of the global cloud infrastructure market, while European providers achieve only single-digit to low double-digit market shares even in their home market. The Federal Office for Information Security (BSI) has repeatedly warned of the sovereignty risks posed by this dependency. Germany and Europe set the rules for data, yet own neither the dominant cloud infrastructure nor the leading AI computing infrastructure on which that data is processed.
In the economy ahead, economic value is created where data is refined with computing power into AI services. Today, this stage of value creation lies largely with non-European corporations. Whoever controls the cloud and the large models also controls the German economy's access to AI. Strict European data regulation protects fundamental rights. Without its own infrastructure, however, it risks becoming a mere barrier to entry for European providers whilst data power remains with the platforms. Gaia-X has established standards and pilot data spaces, but no market-dominant alternative. Without its own computing infrastructure and functioning data spaces, Europe risks writing the rules for a raw material whose value creation accrues to others.
Paths for Germany
The sectoral data spaces that emerged from Gaia-X — covering mobility, healthcare and industry — are being developed from research projects into binding, business-ready infrastructures, with clear rules governing access and remuneration. This is of central importance for Germany, given that its industrial strength rests on machinery and production data that currently lies unused in silos. Shared data spaces could unlock this raw material without surrendering it to the platforms. Feasibility is rated medium, as companies will only share data when the benefits are credible and trust is established.
Trade-off: data sharing versus the protection of proprietary competitive advantages.
Feasibility: The Gaia-X pilots have delivered the standards. What is missing is the leap into productive operation.
Europe is investing jointly in data centres and AI computing capacity — through initiatives such as EuroHPC supercomputers and AI Gigafactories — in order to reduce dependency on US hyperscalers. This matters for Germany because without its own computing power, any claim to data sovereignty remains theoretical and domestic AI providers are placed at a disadvantage. A public baseline provision of computing capacity could function as a commons. Feasibility is low, however, given the enormous investment required and Europe's structural lag in chips and economies of scale.
Trade-off: sovereignty versus enormous investment costs and uncertain competitiveness.
Feasibility: Without this step, European data regulation remains a rulebook without any processing infrastructure to back it up.
The state promotes BSI-certified sovereign cloud offerings and open European language models — such as Teuken-7B, developed out of OpenGPT-X — as alternatives to proprietary US models. This matters for Germany because public administration and critical infrastructure should not remain permanently dependent on individual foreign providers. Open models built to European data protection standards reduce dependency and strengthen control over data flows. Feasibility is moderate, as performance and ecosystem maturity lag behind the market leaders.
Trade-off: sovereignty and data protection versus near-term performance capability and cost.
Feasibility: Teuken-7B and OpenGPT-X demonstrate that European models are achievable. Closing the gap with the frontier remains elusive.
The state makes anonymised public data holdings — covering health, mobility and public administration — available through secure environments for research and domestic enterprises, within the framework of the Data Governance Act and the European Health Data Space. This matters for Germany because large, publicly funded data assets currently go largely unused whilst private platforms continue to extend their data advantage. A governed commons can direct public data towards the benefit of society rather than individual corporations. Feasibility is relatively high, as the legal foundations already exist and implementation is the primary challenge.
Trade-off: open data access versus data protection and the risk of misuse.
Feasibility: The Research Data Centre for Health and the European Health Data Space are the first building blocks of this commons.
Trigger-based phases
Trigger: the expiry of Gaia-X funding and the entry into force of operational obligations under the Data Act and the Data Governance Act.
The pilot data spaces move into permanent operation, sovereign cloud offerings are promoted with BSI certification, and public data holdings are made accessible through protected environments. The aim is to turn standards and prototypes into robust infrastructure.
Trigger: Sustained market dominance by US hyperscalers exceeding two thirds of the market, or a security-critical outage at a dominant provider.
Europe expands its own computing capacity through EuroHPC and AI Gigafactories, promotes open European models as the standard for public administration, and embeds multi-cloud strategies across critical infrastructure. The processing base for European data is thereby strengthened.
Trigger: Demonstrable value creation from shared data spaces making a meaningful contribution to the AI growth potential projected by the IAB (Institut für Arbeitsmarkt- und Berufsforschung).
Data spaces, sovereign cloud and a public data commons together form a coherent ecosystem in which data, as a raw material, serves society and the domestic economy. Europe pairs its rule-setting with its own infrastructure and value creation.
Real-world pilot programmes
Status: Operational; emerged from Gaia-X principles.
A sovereign data space through which transport and mobility data from public and private actors are shared under common rules. It is one of the most concrete examples of how a European data space can function productively.
Status: Under development; anchored in legislation.
The Research Data Centre for Health consolidates claims data from statutory health insurers for research purposes and is being embedded in the European Health Data Space. The pilot demonstrates how publicly held data assets can be made accessible as a commons — and, at the same time, the hurdles posed by data protection requirements.
Status: Published; multilingual European language model.
An open multilingual language model built to European data protection standards, developed by German research institutions within the Gaia-X ecosystem. It demonstrates that Europe can build its own AI models, even as the gap to the global frontier remains considerable.
Status: Established and widely deployed; to be assessed as a model for Germany.
Through its DEPA architecture, India has created a system of user-controlled data sharing in which citizens manage the flow of their data via consent managers. Direct transferability is limited owing to different legal and data protection traditions. Nevertheless, the principle of user-centric data sovereignty is feeding into the European debate.
What we are watching
This share measures who controls the processing infrastructure for European data — and thereby the core of digital sovereignty. If the European share fails to exceed 20% despite Gaia-X and sovereign cloud initiatives, structural dependency will persist. A security-relevant outage at a dominant provider would expose that vulnerability immediately. Synergy Research, Statista, and BSI reports provide the underlying data.
This indicator shows whether Germany's data spaces move from the pilot stage into productive, regular operation. If they remain prototypes without a critical mass of participants, the European data strategy will come to nothing. Germany's industrial data in particular will determine whether an independent ecosystem emerges. Progress can be tracked via the Gaia-X Hub Germany and industry associations.
This metric indicates whether Europe possesses its own sovereign AI model base for sensitive applications. If no European open model enters productive government use by 2030, the public sector will remain dependent on proprietary foreign models. This directly affects control over data flows within public authorities and critical infrastructure. OpenGPT-X, the BSI, and the Bundesnetzagentur provide the monitoring basis.
This indicator measures whether publicly funded data assets genuinely benefit society and domestic innovators. If their use remains hamstrung by data-protection hurdles and federal jurisdictional boundaries, the data advantage of private platforms will continue to grow unchecked. A functioning data commons would serve as a counterweight to private data power. Data Governance Act reports and the Research Data Centre provide the basis for ongoing tracking.