How Community Technology Cooperatives Work
A community technology cooperative is a group of people who jointly own and democratically govern their own internet infrastructure and software instead of renting it from a corporation.
One member, one vote, and the assets belong to the membership.
The technology is the easy part; bylaws, dues, and succession planning are what decide whether it lasts.
A community technology cooperative is a member-owned organisation that builds and runs its own digital infrastructure — network links, servers, the software people actually use — with every member holding one vote in how it is run. The members are not customers of some distant company. They are the owners, and the equipment on the roof or in the basement belongs to them collectively.
That sounds simple and it is not. A co-op is a legal structure, a funding model, and a governance habit at once, and most of the ones that fail do so for boring reasons rather than technical ones. This page covers what the structure means, what such a group runs, how the money and the voting work, where these projects break down, and a realistic first year.
What makes it a cooperative and not just a club
A cooperative is defined by ownership, not by atmosphere. Three things have to be true before the word honestly applies: the members own the assets, the members elect whoever runs day-to-day operations, and each member gets one vote no matter how much money or labour they contributed. That last rule is the line between a co-op and a startup with a friendly tone.
A useful test is to ask who owns the antenna on the roof. In a volunteer collective, the answer is usually whoever bought it, which becomes a problem the day that person moves away. In a co-op, the answer is the membership, written down, with a documented process for what happens to it.
Ownership is what turns a privacy preference into something durable. A group that collectively owns its servers decides what is logged, what is retained, and who is allowed to ask — the practical meaning of data sovereignty at a neighbourhood scale rather than a national one.
The cooperative principles, applied to servers and radios
Cooperatives worldwide work from a shared set of principles first written down by weavers in Rochdale, England, in the 1840s and maintained since by the International Cooperative Alliance. They were designed for grocery stores and credit unions, but they translate to infrastructure almost line for line.
- Open membership. Anyone in the service area can join, not only the people who can read a routing table.
- Democratic control. One member, one vote. The person who donated the switch does not get two.
- Member economic participation. Dues and joining fees fund the thing; any surplus goes back to members or into the network, not to outside investors.
- Autonomy. Do not accept a grant or a sponsorship that hands a vendor control of your DNS, your billing, or your member list.
- Education. Written runbooks and training nights, so knowledge is not hoarded by accident.
- Cooperation among cooperatives. Peer with the co-op two towns over, share spare hardware, copy their bylaws.
- Concern for community. Cover the household that cannot pay full dues.
None of these are decorative. Most co-op disputes turn out to be a violation of one of them that nobody noticed for two years.
Legal structures and how the money actually works
The legal form depends on where you live, and this is one place to pay for advice rather than guess. In the United States the common options are a cooperative corporation under a state co-op statute, an LLC with a cooperative operating agreement, or a nonprofit corporation. Many countries have a dedicated cooperative society registration that is cheaper and simpler.
Tax treatment is a separate question from legal form. The US tax code has an exemption category long used by mutual telephone and electric cooperatives, and charitable exemptions that can fit digital-inclusion work. A charitable structure is governed by a board answering to a public mission rather than by members answering to each other. The WOPR Foundation is a 501(c)(3) nonprofit rather than a cooperative, and that is a real tradeoff: easier to accept donations, harder to say the membership owns the servers.
Income usually comes from a joining fee, recurring dues, capital contributions toward specific equipment, grants, and donated labour. The budget lines that surprise new groups are the unglamorous ones: liability insurance, roof access agreements, upstream transit, an accountant for the annual filing, and a replacement fund for hardware that fails on its own schedule.
What a community technology cooperative actually runs
There are two layers, and a group can start with either. The network layer moves packets between buildings; the service layer is the software members log into. Plenty of co-ops never build a radio network at all and simply run shared services over ordinary internet connections.
For the network layer, the usual kit is rooftop point-to-point radios in unlicensed spectrum, fibre where a landlord or municipality makes it affordable, and open firmware such as OpenWrt or LibreMesh. Routing between nodes is handled by a mesh protocol like Babel or batman-adv, usually with a WireGuard or Nebula overlay for encryption and remote administration. Our guide on how mesh networks work covers the routing side in detail. For very low-bandwidth links such as sensor readings and messages when the grid is down, LoRa radios running Meshtastic are cheap and slow; Blackout Labs builds off-grid LoRa mesh hardware for that job.
The service layer is where members feel the benefit. Nextcloud replaces shared drives, Matrix or a fediverse server such as Mastodon or Akkoma replaces group chat, Jitsi handles calls, Vaultwarden handles passwords, and an identity server like Authentik or Keycloak gives everyone one login. Underneath sits something dull and reliable: Debian, Docker or Proxmox, PostgreSQL, Caddy or nginx, and Ansible so the build is reproducible. Our step-by-step de-Googling guide maps the same tools onto a single household.
Governance is the part that decides whether it survives
Bylaws are boring and they are the product. They set who may join, what dues are, what makes a quorum, how the board is elected, and how a decision is recorded so nobody relitigates it later. Groups that skip this do not avoid governance; they do it informally, which means the loudest person governs.
Voting deserves attention because a co-op vote has to be trusted by people who were not in the room. Small groups manage with a show of hands and honest minutes. Once a vote decides money, you want a verifiable process where the count can be checked without exposing individual votes. Folkmoot is verifiable voting software built for exactly this kind of member decision.
Operational authority is a separate question from democratic authority, and conflating them is a common failure. An outage is not a democratic event. The workable pattern: the membership votes on policy, budget, and who holds the keys, and a small operations group acts immediately within that policy and reports afterwards.
Write down the unpleasant procedures while everyone still likes each other: how root access is granted and revoked, how credentials rotate when someone leaves, and how a member can be removed.
Where community technology cooperatives fall down
The most common failure is the bus factor. One capable volunteer holds every credential, knows where the DNS is registered, and is the only person who has ever restored a backup. When that person burns out or moves, the co-op discovers it was never really collective. The fix is unglamorous: two people with root, a documented restore someone else has actually performed, and credentials in a shared vault rather than one laptop.
The second is support expectations. Members who pay dues expect help, and volunteers do not want to answer the phone at two in the morning. Either the co-op says plainly that service is best-effort and repairs happen on weekends, or it budgets to pay someone.
There are also things this model is not good at. If you need a contractual uptime guarantee, a co-op run by neighbours is the wrong tool. Nothing life-safety-critical should depend on it as the only path. Regulated data such as medical records brings compliance duties a volunteer board should not take on casually. And if your backhaul is thin, members used to instant high-definition streaming will be disappointed.
For a household that wants privacy and no meetings, the honest answer is that self-hosting alone, or paying a small privacy-respecting provider, fits better than starting a cooperative.
A realistic first year
The sequence below assumes a handful of committed people rather than a mailing list of enthusiasts. Five to ten who will still be there in a year matter more than a hundred who signed up once.
- Months one to three. Write a one-page purpose statement: who this serves, what it provides, what it does not. Agree on the legal form. Find out what registration costs where you live.
- Months three to six. Incorporate, adopt bylaws, open a bank account in the organisation's name, and set dues. Money in a personal account is how co-ops end up in court.
- Months six to nine. Launch exactly one service. File storage or group chat is a good first choice because the failure modes are visible and forgiving. Write the runbook as you build, and make sure a second person can restore it.
- Months nine to twelve. Add the network layer if connectivity is the actual need, or a second service if it is not. Hold a real budget review with the membership and publish the numbers.
One rule is worth adopting from day one: nothing goes into production until a second person has restored it successfully. It slows the first months, and it is the practice that most reliably separates co-ops that reach year three from those that quietly stop.
How it compares to the other models
Cooperatives are one of several ways to run shared infrastructure, and not automatically the best one. The differences show up most clearly in who owns the equipment and what happens when the founder leaves.
| Model | Who owns the assets | Who decides | When the founder leaves |
|---|---|---|---|
| Cooperative | The members, collectively | Members, one vote each | Bylaws already say who takes over |
| Nonprofit | The organisation, held for a public mission | A board, answering to that mission | Board appoints a successor |
| Volunteer collective | Whoever paid for it | Whoever shows up | Often collapses |
| Commercial provider | Shareholders | Management | Nothing changes for you until the terms do |
A cooperative is right when the people using the infrastructure are also the people who should control it, and there are enough of them to share the load. A nonprofit fits better when the work serves people who are not members and runs on donations or grants. A volunteer collective is fine for experiments and bad for anything anyone depends on.
These models also mix. A nonprofit can incubate a co-op, hold the initial equipment, and transfer it once the membership can carry it. More guides on the underlying pieces are in our learn section.
Frequently asked questions
How many people do you need to start a technology cooperative?
Legally, most jurisdictions require only a handful of incorporators, sometimes as few as three. Practically, aim for five to ten people who will still be involved in a year, including at least two who can administer a server. Below that, the workload lands on one person and the cooperative is a cooperative in name only.
Do cooperative members need technical skills?
No, and a co-op made only of technical people usually struggles. You need someone who can keep books, someone who can chair a meeting without it running three hours, and someone who will answer member questions patiently. Two or three people who can run servers is enough, provided the knowledge is written down rather than held in one head.
How is a community technology cooperative different from a nonprofit?
A cooperative is owned and controlled by its members, who are also the people it serves, and each member votes. A nonprofit is governed by a board accountable to a public mission, and donors get no ownership or vote. Cooperatives suit member-funded shared infrastructure; nonprofits suit work that serves people beyond the membership.
What does it cost to run a community technology cooperative?
Costs vary too much by location to quote honestly, but the categories are consistent: incorporation and annual filings, liability insurance, a business-class internet connection or transit, server hardware or hosting, domain and certificate costs, and a replacement fund. Budget for accounting help and for hardware failing on its own schedule rather than yours.
Can a cooperative legally provide internet service to its members?
In many places yes, but the rules vary widely. Unlicensed spectrum generally needs no radio licence if you respect power limits and use certified equipment, while acting as a service provider can trigger registration, tax, or reporting duties. Some regions restrict community networks outright. Check your national regulator and local rules before advertising service.
What happens to the equipment if the cooperative shuts down?
Your bylaws should say, and this clause is worth writing early. Cooperatives typically settle debts first, then return members' capital contributions, then distribute or donate what remains, often to another cooperative. Nonprofit rules are stricter and usually require remaining assets to go to another mission-aligned organisation rather than to individuals.