Summary

  • The Address Council is fifteen people, three selected by each of the five regional internet registries, with staggered three-year terms and a chair elected from among its own members. It exists because a memorandum of understanding put it there.
  • Its two headline functions are choosing one ICANN Board director and overseeing the Global Policy Development Process. The institutional pages describing those functions disagree about which Board seat is at stake, and none of the material we obtained states whether "oversees" means to initiate, amend, block, delay, certify or merely forward a proposal.
  • Read carefully, the design is a delegation in two steps: each registry supplies three councillors, and the council supplies one director. Legitimacy therefore travels through five institutional selection chains. What those chains actually require of a selector — an open ballot, a board appointment, a membership vote with a published turnout — is the question that decides whether the council speaks for the address community or merely for its registries.

At the end of January 2026 the world's five regional internet registries held between them the authority to allocate the address space on which every routed network depends. Three of them were in open institutional distress. The Address Supporting Organization, the ICANN supporting organization built to represent that layer of the internet, had a fifteen-person council in place and, on the evidence available, a live Board selection cycle on its own website. Whether that council could have been said to speak for a single operator outside its own five parent institutions is a question its records do not answer.

That is the useful way to read the Address Council. Not as the parliament of the address community, which it is routinely described as being, but as an appointment engine with a policy brief attached. The distinction matters because the two readings imply entirely different accountability structures, and only one of them is visible in the documents.

A body built out of five appointments

The Number Resource Organization's page on composition states the arrangement plainly: the Address Council is composed of fifteen members, with three selected by each of the five regional internet registries (NRO, ASO AC composition). The Address Supporting Organization's own council page adds the mechanics — fifteen members, three from each regional registry's territory, staggered three-year terms with one representative per region refreshed each year, and a chair elected from among the members (ASO, Address Council). A separate roster page is described as the authoritative listing of current membership by regional registry representation, and as listing the fifteen-member composition by region (ASO, AC roster).

The arithmetic is worth sitting with. Five regions, equal weight, three seats each, refreshed on a rotating basis so that no single year changes more than a third of a body whose central act is choosing one person. That is not a deliberative assembly designed for argument. It is a small, stable college designed for a decision — the kind of structure you build when you want continuity in an appointment rather than a debate about one.

The terms reinforce it. Staggered three-year cycles with one seat per region turning over annually mean the council cannot be reconstituted quickly by any external event, including an event that discredits one of its parent registries. Continuity is the design goal, and continuity is exactly what an appointment college needs.

What the composition page does not do is tell you who selects. "Selected by each of the five regional internet registries" is a description of the source of the appointment, not of the procedure. An open election among a registry's members, a vote of its board, and a decision by its chief executive are all consistent with that sentence. Those three procedures produce three very different councillors.

The instrument, and the measure of how little of it we can see

The Address Council exists because of a document, not a statute. The Number Resource Organization describes the ASO Memorandum of Understanding as the foundational document establishing the Address Supporting Organization and its relationship with ICANN, and as defining the organization's structure and governance including the role of the Address Council (NRO, ASO MoU). The captured record also identifies the ICANN Bylaws as the principal legal source for the ASO's place inside ICANN's structure and for the council's authority to select an individual for one Board seat (ICANN Bylaws).

Two instruments, then: a multi-party memorandum among the registries and the corporation, and the corporation's own constitution. Neither was readable in this review. Every page cited in this article was retained as a runtime source record with a null excerpt, because direct retrieval failed for all eight of them. What follows is therefore a reading of what these institutions say about themselves on their own pages, not a reading of the instruments. That distinction is the single most important qualification in this piece, and it belongs beside the argument rather than at the end of it.

It is possible to make a strong institutional claim from a party's own self-description, provided you are honest that this is what you are doing. A registry's statement that a memorandum is foundational tells you what the registry treats as foundational. It does not tell you the amendment rule, who can withdraw, what happens to council seats if a registry's recognition lapses, or whether the corporation can change the arrangement without the registries' consent. Those are the clauses that decide where power sits in a crisis, and they sit in documents nobody outside the five institutions has quoted at us.

Two steps, and the second one is the interesting one

Laid out in sequence, the design reads as a delegation. Each regional registry supplies three councillors (NRO, ASO AC composition). The council supplies one ICANN Board director (ICANN Bylaws). Fifteen people are therefore a transmission belt between five institutional selectorates and one seat in the corporation's highest decision-making body.

Delegation is not a criticism. Most of the internet's governance is delegated, and a federated design that maps onto the institutions actually holding the resource is defensible on its face. The question a delegation raises is not whether it is one, but whether the chain terminates somewhere accountable. A registry is accountable to its members, its service region and, in the end, its own legal jurisdiction. A council is accountable to the five registries that filled its seats. An ICANN director is accountable to the corporation's bylaws and to the multistakeholder apparatus around it. Each link is individually intelligible.

It is the joints that nobody has published.

Consider what is missing at the joint between registry and councillor. Whether a councillor may be recalled, and by whom. Whether a councillor must report back to the selecting registry, and in what form. Whether a councillor sitting on a Board selection vote has an obligation to disclose a preference formed before appointment. Whether the three councillors from one region vote as a bloc or are expected to vote independently. None of these is exotic governance detail. Each one changes whether the fifteen behave as fifteen individuals or as five delegations of three.

At the other joint, between council and Board, the design concentrates an appointment decision in fifteen people. What that decision is worth depends on facts the captured record does not supply: the size of the Board, the confirmation or ratification steps attached to a selected director, and whether the seat carries a term that outlasts the council that filled it. A single director among a larger board is not control of the corporation. It is a seat, and seats matter in proportion to how often the room divides evenly.

The seat with two numbers

Here the record becomes genuinely strange, and worth reporting precisely because it is strange.

The Number Resource Organization's page on the selection of the ICANN Board seat describes the Address Council as responsible for selecting individuals to fill Board Seat 9 (NRO, Board seat selection). The Address Supporting Organization publishes a page for the 2026 ICANN Board Seat 10 selection process (ASO, 2026 Board Seat 10 process). The captured ICANN Bylaws record points the same way, describing the council's authority to select an individual for Board Seat 10 (ICANN Bylaws).

One of these pages is probably behind a revision and one is probably current, or the council fills a seat whose numbering changed, or two different seats have been filled at different times through the same mechanism. All three explanations are consistent with the material we hold. None of them is established by it. We did not obtain the current Bylaws text, the 2026 process page's contents, or a dated announcement naming the seat and the selected director.

So the honest formulation is narrow: the council is reported to select an individual for one ICANN Board seat, and the institutional records we captured do not agree on its number. That is a small fact with a large implication. A body whose own parent organizations publish divergent seat numbers for its most consequential single act is a body whose public documentation is not maintained as a legal record. It is maintained as institutional communication, updated when convenient, and read by almost nobody who is not already inside the tent.

For an operator trying to work out who on the ICANN Board answers to the address community, that ambiguity is not academic. It is the difference between being able to name your representative and being told to check two websites that disagree.

The verb that carries the authority

The second function is larger and vaguer. The Number Resource Organization states that the ASO oversees the Global Policy Development Process for the development of policies relating to the management of global IP address space (NRO, Global Policy Development Process). The council page uses the same framing (ASO, Address Council).

Oversight is a word with a wide skirt. In different institutional settings it has meant, at minimum: the power to initiate a proposal; the power to review one that originated elsewhere; the power to amend; the power to return a proposal for further work; the power to delay by inaction; the power to certify that a proposal met procedural requirements; the power to block; or the modest power to receive a completed proposal and forward it. These are not fine distinctions. A body that can block is a veto player. A body that can only forward is a postbox with a letterhead.

None of the material we captured states which of these the Address Council holds. Not one page describes the council's role at any individual stage of the global policy process — initiation, public comment, regional adoption, global convergence, ratification, implementation. The word "oversees" is doing all of the work in every description we have, and it is doing it without specification.

This is where the doctrine lens earns its place. Institutional authority has layers, and they are not interchangeable. There is formal competence, which comes from the constituting documents. There is representational legitimacy, which comes from how the officeholder was chosen. There is functional legitimacy, which comes from executing a bounded mandate transparently. And there is operational legitimacy, which comes from the acceptance of the parties who must live with the outcome. A body can be strong on one layer and hollow on another. The Address Council's formal competence is described as real.

Its functional mandate, at least in public description, is one verb. Its representational chain is five unpublished procedures. Its operational acceptance by the operators whose addresses are governed is untested in anything we found.

That is the finding, and it is worth stating without embroidery: on the public record available to us, the Address Council's policy authority is asserted rather than specified.

What a regional selection chain would have to show

The most consequential unknown is not at ICANN's end. It is at the registries' end, multiplied by five.

Before anyone can call this body community-elected, captured, independent or democratic, five questions need answers per region. Who nominates. Who votes. Whether the final step is an election or an appointment by a board or executive. What the eligibility rules are. What turnout looks like, counted against an eligible population that someone actually defines. And what the removal or recall path is if a councillor loses the confidence of the constituency that was supposed to have chosen them.

To those five, add the disclosure layer. Whether sitting councillors must declare conflicts when the council acts on a matter touching their own registry's interests. Whether votes on Board selection are recorded and published, and if so at what granularity. Whether minutes name who argued what.

None of this is unavailable in principle. Registries publish candidate statements and election results for other bodies. The absence here is an absence in what we could obtain, not proof of absence from the world — and it is exactly the kind of absence that lets a fifteen-person body be described in two incompatible ways by two reasonable observers, both of whom are working from the same four sentences of public description.

One asymmetry is worth naming, because it shapes the incentives. Every regional registry is an institution with a membership base, a fee structure, a policy forum and a stake in the global rules that govern transfers, waiting lists, recovery procedures and the treatment of legacy holdings. Converting five separate channels of influence inside ICANN into one collective interface produces something valuable to those five institutions: a single legible counterpart for the corporation, and a set of fifteen durable relationships in the room where the corporation decides. Nothing in that observation requires anyone to have acted in bad faith.

Structures shape behaviour without needing conspirators.

Why an operator should care

If you run a network, hold an allocation, lease address space, or sit on the governance side of an organisation whose continuity depends on registration records, the Address Council is upstream of you in two ways at once.

It is upstream through the Board seat. One director chosen by fifteen people participates in decisions about the corporation's budget, its strategic plan, its relationship with the registries and, occasionally, its posture toward the registries when the registries disagree among themselves. A director chosen by a body of registry appointees is a director whose instincts will be informed by the institutions that appointed the body. That is not a scandal. It is a structural fact, and it is the same structural fact that any federated appointment system carries.

It is upstream through the policy process. Global address policy determines whether a transfer is straightforward or obstructed, whether recovery procedures exist and who invokes them, how legacy registrations are treated when the registering institution changes hands. If the Address Council can shape, delay or stop that process, then its internal governance is a commercial variable. If it can only forward completed proposals, then it is not, and the energy currently spent discussing the council would be better spent on the registries' own forums.

Which of those is true is precisely what the public record does not settle. That is why the verb matters more than the seat number.

What would change this reading

The reading above is falsifiable, and it should be tested against specific evidence rather than defended.

If each registry publishes nomination rules, an electorate definition, vote counts and turnout, and if each of those procedures is contestable by the registry's members, then the council is a genuinely representative body with an unusually small chamber, and the delegation frame weakens considerably. If the council additionally publishes its own operating procedures, its quorum rules, its recusal rules and its voting records, the functional layer fills in.

If a current Bylaws section, cited by number and quoted, states the council's powers at each stage of the global policy process, then "oversees" resolves into something an operator can plan around, and the article's central complaint — that the mandate is described rather than specified — evaporates for the policy function.

If an announcement names the selected director, the seat and the date, the numbering ambiguity becomes a documentation footnote rather than a symptom.

And if the records show the council has, at some point, exercised a substantive power — refused, delayed, returned or amended a global proposal — then the body is a real gatekeeper and deserves to be analysed as one, with all the legitimacy demands a gatekeeper attracts.

Any one of those findings would move the conclusion. Several of them together would overturn it. That is the standard the argument should be held to.

The consequence

Strip the acronyms away and the arrangement is easy to state. Five institutions hold the world's regional address registries. Each sends three people to a fifteen-person body. That body chooses one director of the corporation that coordinates the internet's names and numbers, and is described as overseeing the rules those institutions apply.

Any operator who depends on those rules has a legitimate interest in knowing how the three people from their region got there, what they were told to do, and what they can be removed for. Today, on the record we could verify, none of that is public in a form a reader can check. The council's authority is real enough to have a page, a roster and a seat. It is not yet documented well enough for anyone outside the five institutions to say what it authorises.

That is a fixable problem, and it is fixed with publication rather than reform. Rules for selecting councillors, published by each registry. A definition of the council's powers at each stage of the policy process, published by the council. A single agreed seat number, published everywhere. None of those requires amending a memorandum. All three would convert a body described in one verb into an institution an operator can actually hold to account — see the Address Council's entry in the BTW directory Only for Strategic Circle for the surrounding institutional record.