Summary

  • On 27 April 2005, AFRINIC reported that its Board had amended the fee schedule to cut annual membership-only associate membership without allocated resources from US$400 to US$100, and that the AGM unanimously approved the figures. That was a US$300 reduction, or 75%, but the public record does not disclose the motion, voter roll, denominator, item-by-item result or exact effective date.
  • Under the constitution archived shortly before the meeting, an eligible or specially approved Organisation could complete effective membership through the membership process and annual payment, then exercise real internal corporate rights. The US$100 did not buy an address allocation, an ASN, resource registration or sovereign power, and the available evidence does not prove increased participation, revenue neutrality or a subsidy from simultaneous increases in three resource-holder categories.

Analysis

The smallest figure in the room carried the sharpest boundary

On the second day of AFRINIC-2 in Maputo, a line in a fee schedule changed from US$400 to US$100. The event report places the business on 27 April 2005, during a day devoted in part to the member AGM. AFRINIC’s record says the Board made amendments to the schedule. It lists the reduction for “membership only,” “associate membership” and membership “without allocated resources,” then says the AGM unanimously approved “these figures.”

The arithmetic is simple. US$400 minus US$100 is US$300. Relative to the old price, the reduction was 75%. The new standard annual fee was one-quarter of the previous one.

The institutional meaning is not simple, because the price sat at the junction of three different surfaces: admission to a company, participation in that company’s governance, and the provision of Internet number-resource services. Those surfaces can be rhetorically compressed into a vague idea of “membership.” The records instead require them to be separated.

US$100 was the price listed for non-resource membership. It was not a charge for IPv4 or IPv6 space. It did not include an Autonomous System Number. It did not itself create a Registration Service Agreement, control over RPKI, WHOIS or RDAP, reverse-DNS service, or resource tenure. If a member-only organisation later sought number resources, the predecessor and successor schedules say it moved into the applicable regular or resource category.

Nor was US$100 a coin-operated ballot. The constitution archived on 4 April 2005—23 calendar dates before the decision—set an eligibility route, an alternative route through Board discretion, a rule about who held membership, and a payment condition. An applicant still had to be an Organisation that qualified under the ordinary description or was admitted by the Board, complete the membership process and pay the annual fee unless a reduction or waiver applied. Payment was necessary for effective membership. It was not sufficient by itself.

Once effective membership existed, however, the constitutional rights were substantive. Members were entitled to notice of Members’ meetings, attendance and voting. They could elect Board members, determine general policies within the company’s objects, discuss Board activity reports and strategy, examine and adopt accounts, submit proposals, appoint proxies and nominate an eligible Board candidate. Some powers were collective: reviewing Board decisions and amending the constitution required two-thirds of the entire membership, while a special meeting could be requested by one-quarter.

One member paying US$100 did not acquire those threshold powers alone.

That is the core of the 2005 act. AFRINIC materially reduced a recurring price condition for an eligible organisation to become an effective non-resource Member with internal corporate rights. It did not open a market in public sovereignty. It did not sell the registry’s ledger. It did not transfer ownership of African number resources. It did not establish that everyone affected by AFRINIC was represented.

Who could use the US$100 route

The ordinary eligibility language in the pre-meeting constitution was narrower than “anyone interested in the Internet.” Membership was open to an Organisation engaged in using or providing network services based on open-system protocols, geographically based within the African sub-region, and wishing to participate in AFRINIC’s affairs. The Board could admit other Organisations as it deemed appropriate.

The repeated unit is Organisation. Membership was held in the Organisation’s business name. It was not held personally by the owners, representatives or employees associated with that organisation. An individual could act for a Member, but could not turn the organisational status into a private personal entitlement.

The distinction matters because a low price can sound universal when the legal route was not. The reduction lowered one cash barrier for a defined class of potential institutional participants. It did not erase geographic, functional or organisational conditions. It did not eliminate Board discretion at the exceptional-entry layer. It did not remove documentation. The archived 2004 membership materials indicate that the startup workflow used a membership form, a member-only agreement and payment. Those materials provide context for the sequence; they do not prove every term of the final agreement in force on 27 April 2005.

The Board also had a constitutional power to reduce or waive fees for deserving organisations and entities. That complicates any attempt to compare the scheduled US$400 and US$100 prices with what every organisation actually paid. The published standard fell by US$300, but no waiver ledger has been found. It is unknown how often discretion had lowered the effective burden before the general cut, who received relief, or whether the standard reduction altered waiver practice.

A further condition lay after admission. Annual payment was a condition precedent to effective membership, and the constitution allowed membership to terminate when fees remained unpaid for three months after the due date. The fee was therefore not merely a one-time application charge. It was a recurring condition connected to good standing and the continued corporate-rights surface. Because the category was defined by the absence of allocated resources, the nonpayment consequence did not by itself supply a power to revoke resources that the member did not hold through that category.

This precise sequence prevents two opposite errors. The first is to say that US$100 simply bought a vote. That ignores eligibility, admission and documentation. The second is to say the payment bought nothing meaningful because AFRINIC was not a government. That ignores the real rights an effective Member held inside the company. A disciplined reading accepts both propositions: entry was conditional, and the internal rights were genuine.

What effective membership meant in April 2005

The most consequential interpretive choice is the date. Current labels cannot be pasted onto an earlier constitution merely because the same word, “associate,” appears in both periods.

In April 2005, the fee records used a shifting cluster of labels. The 2004 predecessor schedule referred to “Special membership” and “Member-Only Fees (Honour members).” The AFRINIC-2 report referred to membership-only associate membership without allocated resources. The later 2005/2006 fee page called the line “Member-Only Fees (Associate Members).” Yet the contemporaneous constitution did not define Associate Member as a separate corporate class. It spoke of Members and Organisations.

The safest method is therefore not to infer rights from the fee label. It is to ask what the constitution attached to effective Member status at that date. The archived April 2005 text answers that question directly.

Article 7.5, as captured, attached notice, attendance and voting rights at Members’ meetings. The subsequent provisions supplied a wider bundle. Members elected the Members serving on the Board of Trustees. They had a role in determining general policies for the company’s objects. They could discuss Board activity reports and strategic planning, examine and adopt accounts, and address other questions necessary to the company’s functioning.

They also had specified methods to challenge, amend, convene and participate: collective review of Board decisions, constitutional amendment, special-meeting requests, meeting proposals, proxy representation and nomination of one eligible Board candidate.

Those rights were not equal in their mode of exercise. Notice, attendance and a meeting vote attached to effective membership. A proposal, proxy or nomination could be exercised according to its own constitutional mechanics. Review, amendment and special-meeting powers depended on collective thresholds. Calling all of these “rights” is accurate only if their individual and collective forms remain visible.

The constitution did not state that a Member paying a larger resource-category fee received more votes. That supports a limited institutional observation: nominal meeting rights followed Member status rather than the size of the fee line. A non-resource Member at US$100 could hold the same stated meeting-vote right as a resource-holding Member whose category fee was much larger. But the record does not reveal the 2005 voting roll, proxy pattern, attendance by class or actual influence. Formal equality in the text is not proof of equal practical power.

The date also protects the analysis against backdating later observer rules. In 2007, AFRINIC’s chief executive described a proposed major constitutional change that would create Full and Associate membership levels. That is affirmative evidence of a version boundary. Later constitutional arrangements and current associate pages may describe different rights, including observer participation, but they cannot establish what the US$100 status carried in April 2005. The 2005 question must be answered with the 2005 instrument.

Nor can the archived 2004 draft agreement silently supply every contractual term. It offers useful context: an annual term, renewal tied to payment, compliance duties, communication and training obligations, treatment of requests and an appeal route to the Board. Yet no final executed version governing April 2005 was located. Those draft terms cannot be stated as settled terms of the post-cut membership bargain. The constitutional rights and fee boundary are much firmer than the surrounding agreement details.

The evidence proves an institutional report, not a reconstructed ballot

Official records carry weight, but only for what they record. The AFRINIC-2 report is primary evidence that AFRINIC documented a Board amendment, the old and new prices, and unanimous AGM approval of the figures. It is not a substitute for the underlying resolution or ballot ledger.

The report does not provide a signed motion. It does not say whether the cut and the three increases were one bundled proposition or four separate propositions. It gives no quorum, voting denominator, voter names, classes, proxy list, count, abstention record or item-level result. More than 100 people reportedly attended the broader meeting, but event attendance cannot be converted into an AGM electorate. No turnout or unanimity percentage can responsibly be calculated from that audience figure.

“Unanimously approved” should therefore remain attributed: AFRINIC’s report says the AGM unanimously approved the figures. There is no basis to call the report false merely because supporting mechanics are unavailable. There is also no basis to embellish the sentence into a fully audited claim that every eligible Member voted, every class participated, or the associate reduction had a separate unanimous ballot.

That distinction is not pedantry. If “these figures” referred to a package, a Member might have approved the overall fee schedule while holding a different view about one line. If there were four separate motions, each price would have a cleaner authorization trail. Both structures fit the short report. Neither has been established.

It is also unknown whether any member-only organisation voted on the change to its own category. The constitution gave effective Members voting rights, and earlier startup records show that paid member-only organisations could participate in a founding election. But no class-specific roll for AFRINIC-2 has been found. A potential right and an observed exercise of that right are not the same fact.

The exact effective date is similarly open. The report records the decision on 27 April 2005. An archived official fee page from 16 January 2006 publishes the resulting US$100 annual price. Between those points, the checked record does not supply an implementation notice, first invoice or transition rule for an existing US$400 member. It is reasonable to say the later schedule reflects the decision. It is not reasonable to invent the billing day.

Even the pre-event constitution has a bounded caveat. The official page was captured 23 elapsed days before the meeting. No intervening amendment was found, and the meeting report does not mention one. The capture is the best located contemporaneous text. A signed company-registry copy dated exactly 27 April and a complete amendment ledger would close the remaining low-to-medium uncertainty.

Official publication thus establishes recorded words and acts. It does not prove unrecorded motives, participation, implementation details or authority beyond the instruments’ lawful scope. That rule applies equally to favourable and sceptical narratives.

A lower barrier, not proven democratisation

The most attractive effect claim is also the one that requires the most restraint. Reducing a recurring fee from US$400 to US$100 could expand the set of eligible organisations able or willing to join. A fixed annual amount weighs differently on institutions with different resources. At one-quarter of the old price, non-resource membership became materially cheaper in cash terms.

Heng Lu’s poverty-incidence doctrine sharpens the mechanism. Fixed institutional overhead can fall disproportionately on smaller or poorer operators and organisations. A reduction of 75% plausibly eases that burden. But the 2005 line was a non-resource participation price, not a demonstrated compulsory levy on every resource holder. No data show which organisations regarded US$400 as prohibitive, which could afford US$100, or whether the Board’s waiver power already provided relief.

Accordingly, “lowered the price barrier” is a factually grounded conclusion. “Democratised AFRINIC” is not. The first compares two scheduled amounts attached to the same route. The second claims an observed change in who joined, who voted, whose interests were represented or how influence was distributed. The checked record supplies no applicant counts, approval or rejection data, renewal series, associate-member roll, turnout record, income distribution or representation measure.

The difference between opportunity and outcome should remain explicit. The decision expanded the opportunity set in price terms: an otherwise qualifying organisation faced a US$100 standard annual fee rather than US$400. Whether that changed behaviour is unknown.

A counterfactual illustrates the limit. If the fee had stayed at US$400, the same constitutional rights would have carried a standard cash condition four times as large. The likely direction is a higher entry barrier. Yet the magnitude of any lost participation cannot be computed without knowing applicants’ resources, willingness to pay, available waivers and alternative forms of engagement.

The Board’s waiver power complicates the comparison further. If waivers had been common at the old price, the effective reduction for “deserving organisations” could have been smaller than the schedule implies. If waivers had been rare or difficult to obtain, the general cut could have reduced dependence on discretionary relief. Neither state is documented.

The result is a more modest but still important finding. AFRINIC did not prove that participation widened; it altered the conditions under which widening could occur. It made formal, voting-capable organisational membership possible at a much lower listed price without requiring the organisation first to hold number resources.

That separation could matter to institutional legitimacy. A registry’s resource customers are not the only organisations with a reason to scrutinise governance, accounts, Board conduct or policy. A non-resource route can permit another set of eligible organisations to become internal principals. Yet their legitimacy remains bounded to the rights and scope of the company. A larger or cheaper membership is not automatically representative of every affected network, user, country or state.

The revenue equation has one known variable and several missing ones

A 75% cut has an immediate price-only implication. At an unchanged membership count and before any other effects, gross standard-fee revenue from this category would fall by 75%. Four annual memberships at US$100 produce the same US$400 gross amount as one annual membership at US$400. Relative to a base of one old-price member, three additional new-price members would be required for price-only revenue neutrality.

That arithmetic is a scenario, not a description of AFRINIC’s accounts. No category membership count, collection cost, waiver record, arrears series or revenue line was located. It is unknown how many organisations were paying US$400, how many renewed at US$100, how many joined, or how much money was actually received. The currency printed in the schedules was USD or US$, but no invoice specimen establishes payment currency, bank charges or tax treatment for an individual payer.

If membership quadrupled after the cut, gross standard category revenue per old-price baseline member would match before administration, collection and waiver effects. More than fourfold growth would raise that gross amount; less than fourfold growth would reduce it. If no new organisation joined and renewals were unchanged, the access price would still have fallen, but the empirical claim of wider participation would fail and category revenue would decline.

The public record does not state which outcome occurred. It also does not publish a forecast showing which outcome the Board expected. The cut may have been designed to expand participation, rationalise a startup price, distinguish governance from resource administration, or serve another objective. Those are hypotheses, not recorded reasons.

The strongest favourable case is coherent. AFRINIC could have treated US$100 as a modest screen for serious organisational commitment while ensuring that formal voice was not reserved for address holders. The substantive constitutional rights would make the payment a real membership product rather than a donation. Simultaneously higher prices in three resource categories could fit a design in which direct registry users carried more financial load.

But coherence is not evidence. No Board paper, AGM debate, budget worksheet or cost study was located. The favourable interpretation should be presented as a steelman, not as AFRINIC’s proven intent.

Three increases stood beside the cut, but adjacency is not a ledger

The same AFRINIC-2 paragraph lists three upward price movements. The medium resource category increased from US$5,000 to US$6,500. The large category increased from US$7,000 to US$13,000. The extra-large category increased from US$15,000 to US$20,000.

Those lines are necessary context because they show the associate reduction was part of a broader fee-schedule amendment. They are not invitations to write three additional articles inside this one.

The medium increase was US$1,500, or 30%. The large increase was US$6,000, approximately 85.714%. The extra-large increase was US$5,000, approximately 33.333%. Those calculations describe the sibling lines. They do not establish who bore the cost, why the numbers were chosen, whether the categories reflected service cost, or how the changes affected particular operators.

Most importantly, the simultaneous movements do not prove cross-subsidy. The record does not say money raised from medium, large or extra-large members funded the US$300 associate reduction. It does not say the higher resource fees offset expected category revenue loss. It does not say all four prices were derived from one redistribution model.

A causal claim would need an accounting or policy link: a Board paper, budget model, motion, revenue forecast or statement of earmarking. None was located in the checked record. The changes may have been connected in intent, or merely revised together as parts of a schedule. The evidence closes temporal and documentary adjacency, not financing causation.

This boundary also prevents a rhetorical transformation of fees into taxes. AFRINIC’s Board exercised fee-review power under a private corporate constitution, and Members reportedly approved the figures. The US$100 line was a voluntary non-resource membership price for an organisational status. The sibling lines were prices for separate resource-holder categories. Whatever their incidence, the instruments do not establish sovereign taxing power.

Heng Lu’s scope discipline remains useful here. Registry charges should be evaluated against essential coordination functions, optional activities, transparency and cost. Yet a modern critique cannot supply a missing 2005 budget. It can identify the question—what exactly did each fee fund?—without pretending the historical answer is known.

The right doctrine: a ledger can coordinate without ruling

The article’s most important authority limit comes from Heng Lu’s doctrine. An RIR records and coordinates uniqueness. It is a ledger and coordination-service institution, not a government, legislature or sovereign. Its authority arises from corporate instruments, contracts, consent and the practical coordination surface; it does not acquire a general licence to punish operating networks or command states.

Applied to April 2005, the doctrine does two jobs at once.

First, it protects real Member rights from being dismissed. Effective Members were not merely an audience labelled “the community.” Within AFRINIC’s corporate structure, they were principals for the powers the constitution actually granted. They received notice. They could attend and vote. They could elect Board members, review reports and accounts, submit proposals, use proxies and join threshold-based collective action. The AGM’s approval, as reported, had internal significance.

Second, the doctrine stops that corporate significance at its lawful boundary. A member vote could authorize AFRINIC’s internal act within the company’s objects and applicable law. It could not transform the company into the sovereign of African Internet infrastructure. It could not bind absent operators simply because they were affected. It could not confer geographic ownership of number resources, create punishment power or convert meeting attendance into a continental mandate.

Heng Lu’s distinction between stakeholder and principal is particularly useful. A stakeholder is affected. A principal authorizes within a defined relationship. The more than 100 attendees at AFRINIC-2 were not necessarily Members or voters. Event attendance did not make them an electorate. Effective Members were corporate principals for the constitutional rights they held, but they were not the public-law principals of every African network, user or state.

This is why the language of “community sovereignty” obscures more than it explains. If “community” means everyone present, it erases the formal membership boundary. If it means paying Members, it inflates a private electorate into a public one. If it means everyone affected by registry coordination, it confuses stakeholders with principals. The 2005 records are clearer: an Organisation could qualify, be admitted, complete the process, pay and acquire defined corporate rights. That chain is meaningful without being sovereign.

The same distinction keeps the fee from becoming a tax. A tax rests on public authority and compulsion of a different legal order. The US$100 charge belonged to an optional non-resource membership route. An organisation did not have to buy this status merely to exist as an African operator. The payment maintained an internal corporate relationship. That does not immunise the fee from scrutiny, but it defines the scrutiny correctly: price, access, rights, governance, transparency and consent, rather than taxation by a state.

LARUS’s modern explanation that RIR enforcement rests on contracts and member cooperation rather than sovereignty reinforces the role boundary. NRS’s modern advocacy about fees and member accountability shows why money and voting rights remain consequential. Neither supplies the missing 2005 motion, roll or revenue ledger. AFRINIC’s records remain the best evidence of what AFRINIC documented and published. Heng Lu’s doctrine controls how far those institutional words can be stretched.

The strongest case for the cut—and why it remains a case

A fair analysis should state the best institutional argument in full.

AFRINIC was building a regional coordination institution. If formal membership had remained tied in practice to holding resources or paying US$400 for a non-resource role, organisational voice could have been concentrated among direct registry customers and wealthier participants. Reducing the non-resource price to US$100 could create a wider route into governance without giving away resource service. A meaningful but modest fee could signal commitment and support administration. The constitution’s notice, voting, election, accounts and review rights made the product substantive.

Higher charges for medium, large and extra-large resource categories could reflect the heavier direct service relationship or greater capacity of those organisations. The AGM’s reported unanimous approval could indicate a shared judgment that the redesigned schedule was acceptable.

Every element of that case has some support as structure. The category was without resources. Member rights were meaningful. The cut was large. Three resource-holder categories rose. The Board could waive fees. AFRINIC reported unanimity.

What remains missing is the connective evidence. No document says the purpose was to diversify membership. No record says US$100 screened seriousness. No study says the old price excluded smaller organisations. No budget says larger resource holders were paying for wider governance access. No member series shows increased uptake. No voter roll shows non-resource Members exercising the franchise at the decision. No revenue ledger shows whether the model was sustainable.

The strongest conclusion is therefore narrower than the strongest case. AFRINIC made the standard non-resource membership price materially cheaper while preserving a constitutional route to real corporate rights. It separated that membership from number-resource service. Those design choices could widen formal access. The public record does not prove that they did, why the Board chose them, or who paid for the resulting revenue effect.

That restraint is not hostility to AFRINIC. It is the same evidentiary discipline that prevents an unsupported accusation. One should not call the vote fabricated, the fee extortionate, the resource increases a tax transfer or the organisation sovereign. Each claim would outrun the sealed record. The record is strong enough to support a more exact and more interesting account.

Twelve gaps define the edge of the story

The first gap is motion structure. The phrase “these figures” could describe a package or several items. The available report does not decide.

The second is the electorate. There is no quorum, denominator, voter identity, class roster, proxy record or count. More than 100 event participants are not a substitute.

The third is participation by non-resource Members. Their constitutional voting right is established; their exercise of it on 27 April is not.

The fourth is implementation timing. The decision date and later published schedule are known. The effective date, first invoice and transition treatment are not.

The fifth is membership response. There are no located counts for eligibility, applications, approvals, rejections, waivers, renewals or departures.

The sixth is revenue. The unit-price arithmetic is certain, but collections, arrears, waivers, costs and total category receipts are missing.

The seventh is rationale. The Board’s recorded reason for selecting US$100 has not been found.

The eighth is financing. No document proves that the three resource-category increases caused, funded or offset the associate cut.

The ninth is the final agreement. A 2004 draft and workflow survive, but the executed form governing April 2005 has not been located.

The tenth is terminology. “Special,” “honour,” “member-only” and “associate” appear around the same non-resource category, while the April constitution defines Members rather than an Associate class. A contemporaneous definitions instrument reconciling the labels is missing.

The eleventh is constitutional version certainty. The 4 April capture is close and no intervening change was found, but a signed 27 April instrument and full amendment ledger would be stronger.

The twelfth is payer-level currency treatment. Schedules state US dollars, but no invoice establishes settlement currency, bank charges or tax treatment.

These gaps are not all equal. The absent ballot mechanics, member counts, rationale and revenue are material to claims about authorization, access and economics. Currency processing is less material to the institutional act. The constitution-capture interval is a bounded version uncertainty, not a reason to discard the best contemporaneous text.

Together, the gaps establish a useful rule: uncertainty should narrow the conclusion, not dissolve the known facts. The price change, eligibility route, payment condition, express Member rights and non-resource boundary are well supported. Outcome, motive, incidence and some mechanics remain open.

What the US$100 decision means

The decision matters because it demonstrates that governance access and resource service can be separated. An organisation did not need an allocated address block or ASN to hold effective membership. At the new standard price, the cash condition for that status was much lower than before. The status carried real internal rights under the constitution then in force.

That architecture can be assessed without mythical language. The Board reviewed and amended a fee schedule. The AGM reportedly approved the figures. Organisational criteria and Board discretion governed admission. Documents and payment completed the path to effective membership. The constitution defined rights. Separate qualification, agreements and fees governed number resources.

The architecture also reveals its own limits. The Board’s discretion could shape admission outside the ordinary route. Collective rights required coordination among Members. A lower fee could expand eligibility in practice, but could also reduce category revenue. Formal voting equality did not guarantee practical influence. The public record did not expose the electorate or economic rationale well enough to measure those effects.

US$100 therefore bought neither African Internet governance nor nothing at all. For an admitted and compliant Organisation, it completed a low-priced annual route to membership with notice, attendance, a meeting vote and a broader accountability bundle. For everyone else, it created no entitlement. For resource administration, it was the wrong category. For sovereignty, it was irrelevant.

The durable lesson is one of institutional precision. A registry can maintain a ledger and coordinate uniqueness. A company can charge for membership and give Members meaningful governance rights. A member electorate can authorize internal acts. None of those propositions requires the fiction that AFRINIC rules a continent.

The 2005 act widened the formal opportunity in one measurable dimension: price. Whether it widened the membership, electorate or representation is a question the surviving public record does not answer.