When decentralised networking is discussed in policy circles, the conversation usually starts in the wrong place. It begins with routing, encryption, interoperability and the familiar promise that peer-to-peer systems can route around failure. All of that matters. But by mid-2026 the more revealing question is simpler: where, exactly, may the node sit. On whose roof. Attached to which pole. Powered by whose meter. Connected across which wall. The internet’s most centralised layer is often the pavement.
This is not a rhetorical flourish. Any mesh, however ingenious its protocol, must still inhabit buildings, street furniture and airwaves. It must tolerate walls, foliage, weather, maintenance schedules, landlord objections and municipal permit queues. It must cross the legal geography of property rights and the administrative geography of local government. A decentralised topology can therefore be politically fragile even when it is technically elegant. Resilience begins to look less like a protocol property and more like an easement problem.
Decentralisation meets the curb
The common image of a mesh network is abstract: a cloud of nodes relaying packets in many directions, immune to a single point of failure. But real deployments acquire shape from the built environment. Roof height determines line of sight. Facade materials attenuate signals. Building density creates canyons and dead zones. Lamp posts offer mounting points, but only if a municipality permits use and supplies power. Multi-dwelling buildings can concentrate users, yet a single managing agent may veto installations that would have connected hundreds of flats.
NIST’s work on radio propagation through building materials remains a useful corrective to software-centric thinking. Concrete, metalised glass and dense interiors are not minor annoyances; they are part of the network architecture. In crowded urban settings, even the movement of people can alter signal behaviour. The technical literature therefore points towards an inconvenient truth for decentralisation advocates: physical placement is not implementation detail but constitutional structure.
The quiet power of landlords
If one wanted to identify the least glamorous but most consequential actors in decentralised connectivity, landlords would rank high. A permissive owner can enable rooftop backhaul, shared internal cabling and courtyard relays. A hostile or merely indifferent owner can extinguish a neighbourhood scheme without touching a packet. The power is especially acute in apartment blocks, social housing estates, university accommodation and commercial property where large populations sit behind a single property-management decision.
Telecoms regulation has long recognised that in-building access matters, but conventional rules were written with established operators and standard fixed deployments in mind. Mesh and community-led models fit awkwardly into those categories. They often rely on small devices, volunteer maintenance and hybrid arrangements between private residents, civic groups and public institutions. The legal system, by contrast, prefers tidy binaries: public or private, operator or end-user, temporary or permanent fixture. Decentralised practice lives in the blur between them.
A network that no single actor can switch off still depends on physical corridors that many actors can obstruct.
European reforms help, but only partially
The internet’s most centralised layer is often the pavement.
Europe has not ignored the infrastructure question. The European Electronic Communications Code and, more recently, the Gigabit Infrastructure Act are designed to reduce deployment costs, improve access to physical infrastructure and streamline permit procedures. Their logic is sound. Digging roads repeatedly is wasteful; access to ducts, poles and in-building infrastructure can lower barriers; permit uncertainty deters investment.
Yet these instruments are still shaped by a worldview in which network deployment is undertaken by recognisable providers pursuing conventional business cases. They make it easier to install infrastructure, but they do not fully resolve the status of hyper-local, decentralised, intermittently funded or collectively governed networks. The gap is subtle. A city may simplify permit procedures while still lacking a framework for granting low-cost, standardised access to street furniture for small civic deployments. A housing authority may support digital inclusion while forbidding resident-installed relay equipment. In short, the administrative state may become more efficient without becoming more plural.
Why rights of way now matter more
The significance of rights of way has grown for three reasons. First, geopolitics has made resilience a planning priority. European governments now speak more openly about redundancy, continuity and the need to withstand outages, sabotage or political coercion. Second, urban electrification and sensorisation have made street furniture newly valuable. Poles, cabinets and public buildings are no longer passive fixtures but contested platforms for connectivity, energy management and surveillance. Third, denser wireless environments have increased the premium on carefully chosen mounting locations rather than mere access to spectrum.
That changes the economics of decentralisation. In earlier phases of wireless networking, access to unlicensed spectrum could appear to be the great liberator. It remains important. But in many urban areas the scarce resource is no longer simply spectrum as such. It is well-sited, legally secure, low-friction occupancy of small physical points in the city. The real contest is over who may occupy small pieces of shared physical space, on what terms, and under whose supervision.
Unlicensed does not mean unobstructed
There is a persistent misconception that unlicensed or lightly licensed wireless bands automatically favour decentralised networks. They certainly lower one barrier to entry. They do not dissolve the rest. Congestion, interference management, power limits and propagation constraints still apply. More importantly, the ability to lawfully transmit says nothing about the ability to mount, power and maintain equipment where it needs to be.
OECD and ITU work on broadband repeatedly shows that infrastructure bottlenecks and deployment costs shape connectivity outcomes as much as service-layer competition does. That conclusion travels well into the mesh context. A neighbourhood may have users, devices and technical expertise, yet fail because no viable path exists for rooftop permissions or because each street crossing requires a separate bureaucratic negotiation. In such settings, decentralisation is less a triumph of self-organisation than a test of whether the local governance system can tolerate many small acts of infrastructure making.
Municipalities are the hidden governors of resilience
National governments often talk about sovereign capability, but municipalities decide whether many forms of local resilience are practical. They control or influence access to lamp posts, bus shelters, schools, libraries, social housing, planning approvals and public works coordination. They can standardise permits or leave applicants in departmental mazes. They can publish clear technical rules for power draw and mounting safety or insist on bespoke negotiations for each site.
Resilience begins to look less like a protocol property and more like an easement problem.
This makes local government an unexpectedly strategic layer in decentralised networking. A municipality need not run a network to shape its destiny. It need only make access predictable, proportionate and transparent. Conversely, a city that celebrates digital inclusion while treating every small node as an exceptional encroachment on public space will push decentralisation back towards informality or illegality. The issue is not only cost. It is legal legibility.
Informality is a brittle substitute for policy
Many successful grassroots networks began informally: neighbours shared roof access, volunteers climbed ladders, institutions offered a cupboard and a power socket. Such improvisation can be admirable and efficient. It is also brittle. Informal permissions vanish when a caretaker changes, a building is sold, an insurer objects or a resident complains. What looked decentralised may in fact rest on undocumented favours extended by a handful of gatekeepers.
APC’s work on community networks has long documented the importance of local trust and social organisation. By 2026, the lesson is sharper: social capital is necessary but insufficient. Durable decentralisation requires public rules that convert ad hoc tolerance into repeatable entitlement. Without that shift, decentralised connectivity remains vulnerable not only to censorship or market concentration, but to routine property management.
The real contest is over who may occupy small pieces of shared physical space, on what terms, and under whose supervision.
The architecture of buildings is becoming a network policy issue
One underappreciated implication is that construction and renovation policy now intersects directly with decentralised connectivity. New insulation standards, energy-efficient glazing and thicker materials can worsen radio penetration. Building retrofits driven by climate policy may therefore alter the viability of indoor-to-outdoor relay patterns. Conversely, requiring pathways, risers, rooftop access points or neutral in-building connectivity spaces can lower long-run barriers to many networking models, centralised and decentralised alike.
Europe’s infrastructure rules already move in this direction for high-capacity networks. But the policy vocabulary remains oriented towards fibre-ready premises and conventional installations. There is a case for widening the frame. If a building is expected to support emergency communications, digital inclusion and resilient local connectivity over several decades, then design choices about rooftop access, internal conduits and equipment cupboards are no more incidental than choices about fire exits or meter rooms.
Security arguments cut both ways
Critics of decentralised networking often invoke security, and not always cynically. Shared street furniture, resident-managed nodes and distributed maintenance can create genuine concerns about tampering, liability and accountability. Public authorities worry about unsafe installations or devices interfering with critical systems. Property owners worry about insurance and responsibility if equipment falls, overheats or is misused.
The real contest is over who may occupy small pieces of shared physical space, on what terms, and under whose supervision.
These concerns are real, but they are sometimes used as a universal solvent for difficult governance questions. The answer should not be blanket prohibition any more than it should be laissez-faire. It should be standardisation: clear mounting specifications, power budgets, inspection rules, revocation procedures and model access agreements. Security is strongest when obligations are explicit. A decentralised network becomes governable not by becoming centralised, but by becoming legible to institutions that must coexist with it.
Redundancy requires plural physical hosts
The strategic value of decentralised networks lies in redundancy. They provide alternate paths when conventional routes fail, become congested or are deliberately disrupted. But redundancy at the logical layer is hollow if too many nodes depend on the same class of host: the same landlord portfolio, the same public authority, the same utility poles, the same maintenance contractor. Genuine resilience requires heterogeneity not only in routing but in custody.
That is why rights of way and access rules deserve attention from those concerned with national and civic robustness. A city in which schools, religious buildings, co-operatives, private homes, libraries and small businesses can all host interoperable equipment under clear rules is structurally different from one where only a narrow set of entities can lawfully participate. The former distributes veto power. The latter merely disguises centralisation behind wireless hops.
What a mature decentralised policy would recognise
A more mature policy settlement would start from an unromantic premise: decentralised networks are neither a rebellious overlay nor a complete substitute for conventional infrastructure. They are a complementary layer whose public value depends on boring institutional details. Those details include standard access to public assets, proportionate wayleave rules, building codes attentive to future connectivity, and municipal procedures that treat small installations as routine rather than exotic.
It would also recognise that the central policy question is not whether decentralisation is ideologically attractive. It is whether a polity wants optionality. In a world of cyber incidents, natural hazards, political pressure and ordinary service failures, optionality has become a governance asset. Mesh and peer-to-peer systems can provide some of that optionality, but only if the law allows them to touch the city.
From protocol politics to urban constitutionalism
The deepest mistake in this field has been to imagine that decentralisation is settled primarily by code. Code matters; standards matter; cryptography matters. But once networks leave the white paper and enter neighbourhoods, they are shaped by planning departments, cadastral maps, tenancy agreements and utility maintenance schedules. The struggle over the future of decentralised connectivity is therefore less a battle of protocols than a quiet form of urban constitutionalism.
Seen this way, the next stage of mesh networking will not be won by the most elegant claim about censorship resistance. It will be won, if that is the word, by those jurisdictions that can reconcile safety, property, competition and civic resilience in the management of tiny physical footholds. That may sound prosaic compared with the mythology of the unstoppable network. But prosaic systems often determine whether ideals survive contact with reality. In 2026, the fate of decentralisation runs through roofs, poles and pavements.
