Summary

  • APNIC Member organizations each receive one vote in the 2026 NRO NC election. Their online window runs from 27 August until 14:30 in Mumbai on 10 September.
  • Eligible non-Member conference attendees also receive one vote, but only after in-person check-in at APNIC 62 and verified attendance at at least one of APNIC 54–61. Their window lasts five hours on 10 September.
  • APNIC’s procedure requires candidate totals, valid and invalid ballot counts, disputes and scrutineer communications. It does not expressly require the public result to separate Member and attendee votes or publish either eligible denominator.
  • The 2025 public BigPulse result displayed 599 candidate votes as one total. That is a prior publication shape, not evidence of 2026 turnout or outcome.

One ballot opened on two clocks

The first voter could enter on 27 August. APNIC instructed an authorized contact to log in to MyAPNIC with multifactor authentication, follow the Vote Now link and reach the BigPulse interface. The entitlement belonged to the Member organization: one organization, one vote, exercised by a Corporate Contact or another contact to whom voting rights had been assigned.

The second voter cannot enter yet. A conference attendee must be physically registered and checked in at APNIC 62 in Mumbai by 10:00 on 10 September. That person must also have attended, either in person or online, at least one of the eight preceding APRICOT or standalone APNIC conferences, from APNIC 54 through APNIC 61. Only then does the individual receive one vote. The attendee window opens at 09:30 and closes at 14:30 local time on the same day.

Both ballots count toward the same candidate contest. Both are intentionally recognized by the published rules. Yet they are not the same kind of authorization.

A Member ballot is an organizational act. Its principal is a corporate member, and its human operator acts through a contact-rights record. An attendee ballot is an individual act. Its qualification comes from current physical presence plus a historical conference record. One derives from membership; the other from recurring participation. Giving each one vote does not erase that difference. It makes the difference part of the arithmetic.

This is not an argument that one class is real and the other decorative. APNIC has chosen a hybrid community electorate for two of its three regional NRO NC seats. The question is narrower: when two kinds of voter produce one result, what must the public record preserve so a reader can understand the result without learning who cast a ballot?

The declaration contract stops at the total

APNIC’s general NRO NC election procedure is not silent about result integrity. The Executive Council appoints an Election Chair. The Chair appoints independent scrutineers from other RIRs, ICANN and ISOC staff who are present at the meeting. Election Officers administer voting, retrieve online reports and report the result. Qualified voters and candidates have a bounded route to lodge a dispute before declaration.

The declaration must include the name and total vote count for each candidate, the number of valid and invalid ballots, disputes and their resolution, and any communication from the scrutineers about the count. Those are useful fields. They make a final number more than a winner’s name.

But the required list does not expressly say how many Member organizations were eligible, how many attendee entitlements were verified, how many votes each class cast, or how each candidate’s total was composed. A candidate can receive 300 votes and the procedure can publish that exact number while leaving readers unable to tell whether it came mostly from organizations, mostly from conference participants or evenly from both.

That ambiguity is not fraud. It is compression. The election rules create two source rows, then the result contract keeps only the sum.

The distinction matters because the two rows answer different questions. An organizational vote asks what an APNIC Member, acting through an authorized contact, chose. An attendee vote asks what a person with a current and recent conference relationship chose. A single total answers who won under the combined rule. It does not answer what the combined rule was made of when it ran.

The 599-vote precedent cannot supply its own denominator

The 2025 election provides a concrete public artifact. APNIC’s conference report named Devesh Tyagi as the successful candidate and linked to the BigPulse result. BigPulse displayed four candidate totals: 411, 87, 53 and 48. They sum to 599, which the page reported as total votes.

The page did not split those votes by voter class. It did not state the number of eligible Member organizations, verified attendee voters, ballots accessed or unused entitlements in either category. The total was exact but its composition was not visible.

APNIC’s conference report published other numbers: 451 in-person attendees, 60 online-only participants, 40 economies and 127 represented Member organizations. They are not the missing denominators. A person counted as an attendee may not have met the prior-conference rule, may not have checked in by the voting deadline or may have chosen not to vote. A represented Member organization may not have been eligible, may have been represented by several people but held one vote, or may not have voted. The event table and the ballot ledger describe different populations.

Subtracting 127 from 599 would therefore manufacture a fact. So would treating 451 as the number of attendee voters. The public sources do not establish either operation. The honest conclusion is only that the result was aggregated and the source classes cannot be reconstructed from the published totals.

Nor can the 2025 shape be projected into 2026. This year’s election remains open. Three nominees are listed. No vote total, invalid-ballot count, dispute, category composition or winner exists in the frozen record. APNIC may publish more detail this time. A past output records a prior publication state; it does not promise the next one.

A hybrid electorate is a governance choice, not a natural fact

The NRO NC has fifteen members, three from each RIR region. In the APNIC region, two are elected by the community and one is appointed by the Executive Council. The elected member in this contest will serve for two years and participate in the body that performs the role of the ICANN ASO Address Council.

The published duties include work on global number-policy development, recommendations concerning recognition of new RIRs, selection procedures for other ICANN bodies and advice to the ICANN Board on number-resource allocation policy. The role is institutional and bounded. Winning an APNIC election does not create a personal mandate to speak for every network, company, government or user in Asia Pacific.

That limit makes the ballot composition more, not less, important. “The community elected” is an easy phrase. The actual procedure is more precise: Member organizations and qualified conference attendees elected under a combined rule. The first phrase converts a mechanism into a constituency. The second preserves the actors the rule really admitted.

The hybrid design has a defensible logic. Members finance and formally govern APNIC, while experienced conference participants may bring operational and policy knowledge from beyond the legal membership. One category anchors corporate participation; the other widens the field. A combined ballot can keep the NRO NC seat connected to both.

The strongest defense is therefore not that the two groups are identical. It is that APNIC intentionally gives their individual ballots equal weight in this contest. Accepting that defense still leaves the accounting question untouched. Equal weight at the moment of counting does not require anonymous provenance at the moment of reporting.

The receipt can be thin without being vague

A useful result receipt would begin before voting. It would freeze the rules and eligible denominators by class at a declared time. For Members, it would record the number of eligible organizations and the rule that prevents several authorized contacts from creating several organizational votes. For attendees, it would record the number verified against current check-in and the preceding-eight-conference requirement.

The next line would show the funnel by class: entitlements created, ballots accessed, ballots submitted, ballots accepted, ballots rejected or invalid, and unused entitlements. Reason families could distinguish a late submission, failed eligibility verification, duplicate attempt blocked by the system or another bounded procedural state. No name, login identifier, authentication factor or device data belongs in the public object.

The result line would retain the combined totals required by the current procedure. It should also show candidate totals by class where the group size is large enough to protect privacy. If a class is too small, APNIC could delay, group or threshold the split while still publishing the aggregate class turnout. The objective is to reveal institutional composition, not individual preference.

Disputes and scrutineer observations already belong in the declaration contract. The receipt should bind them to a versioned result and preserve later corrections additively. If a total changes, the old value, reason, authority and replacement time should remain reconstructible rather than disappear behind an edited page.

BigPulse’s security configuration, voter authentication logs and anti-abuse methods can remain protected. Auditability does not require a tutorial for evading controls. It requires enough public state to distinguish an entitlement, an attempted vote, an accepted ballot and a counted result.

What the mix would—and would not—prove

A split result would show whether the two admitted constituencies behaved similarly, whether participation came mainly from one class, and whether the winning margin depended arithmetically on the hybrid design. It would help APNIC decide whether outreach reaches both populations and help future reviewers compare elections without guessing from conference attendance.

It would not prove that one class is wiser, more representative or more legitimate. It would not tell why a person voted as they did. It would not convert the NRO NC into a parliament. It would not show that a different electorate should have produced a different seat.

Even a category-level margin must be interpreted carefully. Member organizations vary enormously in size and network role but receive one vote here. Conference attendees can be employees of Members, governments, vendors, civil society groups or independent participants. The categories identify qualification paths, not pure interests.

That is why the receipt should describe the election rather than moralize it. The rule combines two populations. The record should preserve the combination. Readers can then argue about design using the same facts instead of treating the word “community” as a denominator.

The ballot is already open. The smallest improvement is available before it closes: declare now which class fields the result will retain, then publish them with the existing totals on 10 September. One ballot can still choose one winner. It should not need to pretend it had one source.

Sources