Europe in the Coordination Age · 4 of 7
The Institutional Layer — The Decompression Chamber
On the data cooperative as the fiduciary vehicle at the interface between networks and institutions.
This series of essays offers a civilian perspective on the challenges of the intelligence abundance age — a companion reading to Examining Altman’s “Moore’s Law for Everything”.
“The cooperative is the one institutional form that can convene the table nobody else can.”
The previous essays traced the instability emerging at the interface between institutions and networks: platforms centralizing coordination, decentralized systems struggling to sustain governance, and states building digital public infrastructure without trusted intermediaries capable of operating between citizens, markets, and sovereign systems. The gap they identified is structural. What follows is a proposal for the institutional form designed to occupy it.
We suggest that the interface between institutions and networks requires two distinct components with different origins. Institutional mediation extends from the institution’s side — the fiduciary structure, the legal and institutional substrate, the arbitration layer reaching toward network logic. Relational accountability is native to the network’s side — it emerges the way reputation does in a professional community: through the traces actors leave behind, through the consequences of defection becoming visible to those who matter, and through signals that only sustained human relationships can carry and that no algorithm can fully replicate. Neither is sufficient alone. This article addresses the first. The next addresses the second.
The data cooperative is one of the most promising institutional candidates to occupy that gap — and the EU Data Governance Act has already created legal space for neutral data intermediation services: entities designed to facilitate voluntary data sharing while remaining structurally separate from the actors using the data. That is not yet a full cooperative architecture, and it should not be overstated. But it is the beginning of the legal category this argument requires: an intermediary whose role is to make data sharing possible without allowing the intermediary itself to become the extractor.
What remains to be built is the institutional form that fully inhabits that space. A data cooperative extends that logic into institutional form. It is member-owned, democratically governed, and designed around fiduciary-like obligations toward those whose data it stewards — closer to the structural obligation a lawyer or doctor carries toward a client than to a platform’s relationship with its users. Its purpose is not to maximize data extraction but to make cooperation possible among actors who cannot safely share directly.
The concept has roots in research communities — including MIT’s work on social physics, data stewardship, and pooled operational data — and its intellectual lineage runs through platform cooperativism, mutual credit systems, and further back to institutional forms that predate any ideological association with the word cooperative.
The New York Stock Exchange began as a member-governed association: twenty-four brokers signing the Buttonwood Agreement in 1792 because no single actor could be trusted to govern a shared market. Dee Hock built Visa International on what he called “chaordic” principles — combining decentralized emergence with institutional order — because coordinating thousands of competing banks required a structure that aligned incentives without concentrating control. What unites these precedents is not ideology but structural logic: such institutions are among the few civic intermediaries capable of making coordination rational when no single actor can be trusted to govern the shared resource.
Data cooperatives, stewardship models, and data intermediaries are increasingly appearing in G7, G20, OECD, UN, and multilateral policy discussions — including explicitly in UN Statistical Commission working groups on data governance. Not as mature solutions, but as institutional directions increasingly taken seriously under conditions where existing governance categories appear incomplete.
Building one remains hard. The data cooperative combines the execution difficulty of a startup with the governance burden of a nonprofit — and adds regulatory uncertainty, financial sustainability challenges, technical complexity, and legal risks. A challenge that is rarely named but consistently underestimated: data literacy. Members must understand what they are sharing, why, and under what conditions. Governance without comprehension quickly becomes nominal. Cooperative incubation must include literacy infrastructure alongside legal templates and governance toolkits — otherwise participation fatigue accelerates precisely because the system feels opaque rather than consequential.
Niche experiments are underway across multiple domains. Bauform eG operates in the German construction sector. The Aapti Institute in India has developed frameworks for community data stewardship in low-resource contexts. Driver cooperatives in urban mobility, patient data cooperatives in healthcare, and agricultural data pools in precision farming each represent different applications of the same structural logic. Each is difficult. Each points in the same direction.
India’s DEPA framework — its Data Empowerment and Protection Architecture — offers a related proof of concept, though it is not a data cooperative. Its Account Aggregator system operates through networks of consent managers acting as fiduciary intermediaries between sovereign infrastructure and citizen data. As of March 2026, the system had enabled interoperability across roughly 2.88 billion financial accounts, with hundreds of millions linked by users through consent-based data sharing. The significance is not only scale but institutional direction: the emergence of intermediary governance layers between individuals, platforms, and state infrastructure. But DEPA also illustrates the limits: legal and technical architecture alone do not automatically generate durable legitimacy or relational accountability.
Europe may also possess unusually fertile institutional ground for such experimentation. Existing consortium and coordination ecosystems — including initiatives such as GAIA-X, EGI Foundation, and broader European research infrastructure networks — already operate across federated governance environments where interoperability, shared stewardship, and distributed coordination are practical rather than purely theoretical concerns.
Get Gael Van Weyenbergh’s stories in your inbox
Join Medium for free to get updates from this writer.
Data trusts, data unions, personal data stores, and fiduciary agents each address different parts of the same emerging governance gap. The cooperative is not a replacement for existing markets, platforms, or regulatory systems. It is a governed coordination layer: a space where data that cannot safely or legitimately be aggregated through existing structures can nonetheless become collectively usable under conditions participants consider trustworthy. Think of it the way organic certification coexists with conventional food production: a different standard, a different guarantee, serving a different need. Multiple forms can coexist. The cooperative addresses the specific challenge of pre-competitive coordination among distrustful actors at scale.
Its strength lies in layered governance: member ownership, democratic participation, fiduciary-like alignment, portability through the right to fork, and privacy-preserving computation through OPAL — Open Algorithms — which enforces privacy through architecture rather than policy. The algorithm moves to the data rather than centralizing the data itself. The boundary is mathematical, not merely contractual.
The cooperative does not eliminate the tension between institutions and networks. It mediates it. We identify two immediate functions that the data cooperative performs at this interface.
The first is an institutional airlock: a pre-competitive space where actors who cannot trust each other directly can coordinate under institutional protection. The pressure becomes especially visible in sectors where coordination is economically necessary but direct transparency remains competitively dangerous. Bauform eG illustrates the logic: contractors in the construction sector often need to share supply-chain intelligence — pricing fluctuations, logistics delays, supplier reliability — while avoiding direct exposure to competitors. Shared openly, that intelligence becomes a competitive vulnerability. Through the cooperative’s airlock, firms deposit operational data under institutional protection. Rivals see only aggregated sector intelligence — collective patterns, not individual exposures. The pressure differential between complete exposure and total secrecy is neutralized.
Each actor faces what economists call a first-mover penalty: whoever opens their data first bears the full exposure risk while others free-ride on the intelligence gained. The result is collective paralysis — everyone would benefit from sharing, no one has the incentive to go first. The cooperative dissolves this trap. No single actor bears the cost of opening first. They open together, under fiduciary protection.
And for the first time, members gain access to collective foresight — patterns, reliability signals, and coordination capacity that emerge only when the data of many actors is combined under governed conditions. Nobody can convene this table unilaterally — not the state, not the private sector, not civil society, because each distrusts the others as convenors. The cooperative is one of the few institutional forms designed to hold that space.
The second function is economic activation: transforming dormant capital into live collateral. A logistics firm with ten years of operational reliability, a community health worker with a decade of verified care records, a small manufacturer with an established supplier network — these actors possess real economic value that the formal system cannot see because it has no institutional mechanism to verify it. The cooperative renders operational relationships legible, transforming coordination-enabling data from isolated records into a governed asset class — dormant capital made institutionally legible and collectively owned. It is market creation — from value that already exists but cannot be mobilized.
A question the cooperative’s architecture must answer explicitly: who governs it?
The Dutch water boards — the waterschappen — offer the most durable historical precedent. They did not emerge from shared values or idealism. They emerged from shared exposure: the farmers, cities, and industries whose land would flood if the dikes failed had no choice but to govern together. Those who bore the consequences governed the system. The same principle applies here.
A data cooperative’s multistakeholder board should represent those whose data is in the system, whose operations are affected by its outputs, and whose communities depend on its integrity. In practice this means: workers and labor representatives, small and medium enterprises, civic organizations and community groups, public health or sector regulators, and state representatives — each with defined and bounded roles, none with veto power over the others. No single interest can dominate. Representation is calibrated to exposure, not capital. Eight centuries of operational continuity suggest the model is robust.
Data cooperatives are not immune to governance failures. Tight networks can become insider systems. Democratic participation can fatigue until a small group captures decision-making by default. Institutions created to protect members can gradually begin optimizing for institutional survival instead.
The question is therefore not whether failure becomes impossible, but whether failure remains visible, contestable, and reversible before institutional drift hardens into permanent extraction.
Several structural mechanisms matter here. Member ownership aligns institutional incentives toward participants rather than external shareholders. Democratic governance gives participants standing within the system rather than dependency on it. The right to fork — the ability to withdraw data with its provenance intact and form alternative institutions if governance fails — introduces competitive pressure into governance itself. Cooperatives that serve members well retain them; those that drift toward capture risk losing them to better alternatives. None of these mechanisms eliminate risk. They change its structure: capture becomes more difficult to conceal, more costly to sustain, and easier to exit.
The cooperative does not solve trust by itself. It creates the institutional conditions under which trust can become actionable. Institutional trust scales through procedure; relational trust scales through consequence. The institutional layer provides one side of the interface — but the chamber only works because real people inside it stake their reputations, relationships, and reciprocal standing on its integrity. The relational architecture that makes this possible — and why it becomes the load-bearing layer of any human-centered digital system — is the subject of the next article.
→ Next: The Relational Layer — The Load-bearing Layer· 5 of 7
→ The historical interface pattern in depth: The Handshake That Didn’t Happen Yet · When Institutions Meet a Newcomer, the Interface Is What Survives
→ The full framework: Trust as Infrastructure
Written with AI assistance. Ideas and arguments by the author.
