Beyond Extraction: Do Community Data Utilities offer a viable path forward?

Every day, our purchases, travels, and our presence in shared spaces become rendered into data points, reflecting what we want or need, what we can and can't afford, and what we aspire towards. These patterns become visible when our data is pooled together and then become signals for what gets built, priced, stocked, targeted and automated. The challenge is that the institutions currently able to capture this value are not communities themselves, but large tech platforms who benefit from technical capacity, data infrastructure, capital, and scale. This asymmetry in who gets to understand and make use of this data is further complicated by the fact that data often becomes more valuable in aggregate. For instance, a single household's shopping habits are worth almost nothing to anyone whereas a whole town's data, if pooled, would be worth a great deal. Yet today that collective value is fairly scattered and unrealised, while the platforms' advantage and access to this data continues to compound. 

Jens Molbak, the founder of New Impact, describes this as a structural market failure: enormous potential value is created widely, yet captured narrowly, with no obvious mechanism for the market to correct itself. In the past, failures of this kind gave rise to a range of decentralised and democratic structures like cooperatives, mutuals, building societies, and credit unions - each giving ordinary people a way to pool purchasing power, risk and trust. Today, it appears that the AI economy is presenting a similar problem at a pace we haven't seen before and without an equivalent institution to counterbalance it yet. The Community Data Utility (CDU) is a new type of data intermediary that has been floated as one solution to fill that gap. 

On July 3rd 2026, The Centre for Data Futures at Kings College London along with NewImpact, Project Liberty Institute and Lucy Blythe hosted an in-person workshop on Community Data Utilities. The CDU draws on the logic of a public utility: the idea that data, like water or electricity, might be governed as essential infrastructure, in the public interest. The Data Empowerment Clinic, one of the Centre for Data Future’s key workstreams, aims to create a space to explore, examine and early stage concepts like these and we see CDUs emerging as a promising idea that must continue to be tested, along other approaches like data cooperatives, data trusts, and other intermediary structures for governing data in the interest of the people. The workshop gathered over 40 participants from diverse stakeholder groups across academia, philanthropy, government and industry to further examine the CDU concept: what it involves, what it would take to build one and how this might translate across sectors and specific use-cases.

To get an in-depth take into the insights we uncovered across each sector, our partners at NewImpact have compiled a full set of resources including additional findings, sectoral insights, open questions, and areas of convergence in a readback you can explore.

What is a Community Data Utility?

The Community Data Utility, proposed by New Impact, is envisioned as a community governed institution through which people could exercise their individual and collective agency over data concerning them by setting the terms of its use, and ensuring that the value it generates benefits the people and communities the data came from.

Operating within a defined geographic territory, a CDU would act as a fiduciary steward for residents' data, aggregate community data and purchasing power with informed consent, negotiate services on behalf of residents, and ensure economic value flows back to communities. The idea would be that every resident in a specific geography is automatically qualified, but must opt in to join. Once they are in, their collective demand is aggregated. The CDU then uses that demand to negotiate better terms with suppliers, and returns the value to the community through lower prices, better services, or local investment. A more detailed whitepaper, outlines this framework and points to possible use-cases.

What did we workshop?

We began by testing the conditions under which a CDU could work. Participants were asked: what would need to be true legally, in governance, and in terms of its business model for a CDU to be viable, trusted, accountable and sustainable? This question was explored through three lenses: community needs and governance (who is a CDU accountable to, and how is that accountability built in from the start); incentives and the business model (what makes participation worthwhile for communities, and what makes the model sustainable without simply recreating platform dynamics at a smaller scale); and legal mechanisms and structures (what legal form a CDU would need to take, and what protections and obligations that form would carry).

Participants reacted to the CDU model and raised several critical questions: First, on framing: does a "data utility" translate meaningfully in a UK context (or does it need a re-brand?), how might it relate to existing models like data trusts, and is a CDU best understood as a set of people, a set of processes, a set of technical standards or all three at once?

Second, on testing: which domains are most promising for a first pilot, what would make a CDU regulatable and testable in practice, what scale is necessary for it to function, and what can be learned from earlier attempts at similar models in the US.

Third, on growing: where would initial investment come from, what role should philanthropy play in funding proof of concept work before a CDU makes a profit, and what would the government need to put in place (through policy or legal infrastructure) to make this kind of institution possible at all.

While the idea of a CDU was received with considerable interest across the room,there was clear consensus that no single, one-size-fits-all approach will work. The reality is that different contexts will likely call for different models in design, in governance, and in how they're deployed. 

Below are our top three takeaways to take this work forward:

Co-design is non-negotiable in shaping how CDUs will be implemented.

Diverse communities of stakeholders should be able to shape a CDU's structure directly which will be critical to establish its legitimacy and build early trust. A significant piece of this design work will be to establish data governance and participation mechanisms that address questions such as: who operates or governs the CDU; whether that institution is genuinely part of the community it serves; how people understand and can change their consent; how data leakage and misuse would be prevented; how decisions and transactions would be scrutinised; what happens when interests conflict; whether members can challenge decisions or leave; how commitments survive changes in leadership, ownership, or governance; and whether the benefits are sufficient to justify the risks.

The CDU must be tested in a bounded, real-world context, and iterated as needed. 

Considering the model has been envisioned as place-based, the next step will be to work through four questions with key stakeholders:

1. Who is the community to be empowered (which could be defined geographically, culturally, religiously, or as a diasporic group)?

2. What challenges or challenges might they want to solve (e.g food insecurity)?

3. Where the CDU could be piloted and run - including who leads the effort, who might act as champions, and whether it might serve well to be anchored by an existing civil society collective/group?

4. What specific goals might it seek to empower communities around? A CDU will only earn trust and engagement if it delivers tangible benefits communities actually value - this could range from convenience, better access to services, lower costs, improved health outcomes, stronger negotiating power, or better contractual terms?

The CDU can only function within a policy and legal environment that grants it real authority.

Piloting and testing a CDU will not happen in a vacuum, but rather it will need a policy and legislative context that actively supports it. For a CDU to function, it must be recognised as an authority in its own right, with the legal standing to make decisions and hold fiduciary duties in the way the law provides for. The UK government has already acknowledged the need to remove barriers for a healthy network of data intermediaries to thrive, and we see it as a necessary condition for models like CDUs to be tested, scaled, and iteratively improved.

What's next?

Our first workshop aimed to open dialogue, surface what's most relevant to a UK context, test which assumptions need re-examining, and identify the building blocks necessary to translate the model into a pilot. Many participants also pointed to the need and the opportunity to test CDUs in other regions and contexts, some of which may have legal or policy environments even more conducive to piloting the model.

We aim to build on this momentum with a second working session held virtually, where we'll narrow the focus from broad exploration toward one or two bounded use cases, the minimum functions a pilot would need, and the assumptions we’ll be testing directly. If you'd like to follow this conversation and get involved, do sign up to be updated on the next event.