The regional registry’s proposed reforms deserve support, but only if they convert litigation, proxy warfare and institutional paralysis into durable checks on every form of capture.

Constitutions are most valuable when trust has failed. When institutions are calm, rules can look ornamental because reasonable people can improvise around ambiguity. Under pressure, the opposite is true. Every undefined power becomes an invitation to conflict, every discretionary procedure a possible weapon, and every reliance on personal goodwill a substitute for governance.

That was the concern behind my decision, during AFRINIC’s first bylaws consultation, not to propose an immediate constitutional settlement. In my earlier essay, “AFRINIC Bylaws Review: Why I Will Not Comment on the New Bylaws”, I argued that a registry emerging from years of litigation, receivership, decision-making paralysis, disputed elections and damaged trust should first recover enough institutional stability to understand the crisis it was constitutionalising. My objection was not to reform. It was to sequence. A rewrite undertaken too early could turn emergency conditions into permanent design.

The second community consultation changes that judgment. AFRINIC now has a detailed draft Constitution, presented article by article, with stated rationales and provisions that can be tested against the failures they are meant to correct. Participating now is not a reversal. It is the logical completion of the earlier position. The first question was whether AFRINIC was ready to redesign itself. The second is whether the proposed design has absorbed the lessons of its breakdown.

My answer is broadly yes. The draft is imperfect and should be amended in important respects. Yet it offers the most serious opportunity in years to convert AFRINIC’s vulnerabilities into durable constitutional restraints. Members should support that direction, engage the text with precision and refuse the false choice between defending an African institution and demanding that it become governable.

From sequencing to substance

AFRINIC is not an ordinary trade association. It is the Regional Internet Registry for Africa and the Indian Ocean region, responsible for administering a layer of technical coordination on which networks, cloud services, universities, governments and businesses depend. Its authority rests neither on territorial sovereignty nor on ownership of the Internet. It rests on recognition, contractual relationships, accurate registries, community-developed policy and confidence that scarce number resources will be administered predictably.

That makes the central principle of this review African stewardship without institutional impunity. Africa needs an effective regional registry capable of resisting extraction, disruption and external capture. But sovereignty cannot mean insulation from scrutiny, protection of internal failure or permission for officeholders to substitute discretion for rules. AFRINIC’s African character is a reason to demand higher standards of due process, financial stewardship, transparency and accountability, not a reason to suspend them.

The Bylaws Review Committee, composed of one representative from each AFRINIC subregion, has produced a draft that addresses several weaknesses exposed by the crisis. Its terms of reference made the committee advisory rather than sovereign, and the final text still requires member approval. That distinction matters. Constitutional legitimacy cannot be delegated to lawyers, directors, receivers, governments or expert committees. They may prepare and test a framework, but the membership must authorise it.

The committee’s independence was itself contested during the first consultation, prompting a public statement defending its impartiality. Such assurances deserve to be heard, but constitutional legitimacy cannot rest on self-description. Independence must be visible in appointment rules, disclosure obligations, consultation records and the treatment of dissenting submissions. The committee’s regional composition gives the process breadth; publication of the reasoning behind accepted and rejected proposals will give it credibility. A review designed to restore trust must leave an auditable trail showing not only what changed, but whose concerns were considered and why.

Several proposed reforms deserve clear support. The draft separates resource membership from the contractual relationship created by the Registration Service Agreement. It states that holding valuable address space does not, by itself, confer superior constitutional status. It defines Good Standing, requires published procedures for suspension and termination, provides notice and an opportunity to respond, and requires a fair, impartial and timely appeal process. The Policy Development Appeal Committee illustrates why review mechanisms are not procedural luxuries in a community where policy decisions can carry large economic consequences.

The draft also introduces term limits, staggered Board rotation and a narrowly framed business-continuity mechanism. Those provisions recognise an institutional truth AFRINIC learned at great cost: continuity without accountability produces entrenchment, while accountability without continuity can produce paralysis. A sub-quorate Board may need temporary authority to organise elections, but not a licence to conduct ordinary governance indefinitely. An acting chief executive may be necessary, but interim leadership should not become a constitutional lifestyle.

The conflict-of-interest framework is another substantial improvement. Rather than addressing only directors, the proposal extends identification, disclosure and management of actual, potential and perceived conflicts across directors, officers, employees, committee members and others acting for the organisation. It also requires publication of Board minutes, with defensible redactions for legal privilege, personal data and commercially sensitive information. These changes will not eliminate bad judgment. They make bad judgment easier to identify and harder to conceal.

Still, the draft occasionally concentrates power while claiming to distribute it. The Board would appoint all four members of the Nomination Committee and could revoke the committee through a two-thirds vote. The NomCom is declared independent in election matters, but independence is fragile when the same organ appoints, judges, removes and replaces it. The Election Committee would remain composed entirely of senior staff designated by the chief executive. Professional staff should administer elections, but sensitive functions such as voter eligibility, representative verification, disputed ballots and certification should have independent oversight.

The same caution applies to quorum. A nine-member Board should not be able to conduct ordinary business with only three directors merely because an earlier meeting was adjourned. Nor should consequential member decisions depend on a quorum in which directors constitute half of the required presence. Emergency rules should preserve the institution, not create a quieter route around normal participation.

These are arguments for targeted modification, not rejection. Constitutions are not made safer by assuming benevolent officeholders. They are made safer by distributing powers that would be dangerous in hostile hands.

Designing against every form of capture

The most revealing part of the draft concerns representation. A Power of Attorney is not an administrative convenience. It may determine who speaks for a member, who votes, who receives information and who can act legally in that member’s name. Collected at scale, such instruments can alter the internal balance of a membership organisation without changing a single formal voting rule.

That risk ceased to be hypothetical in 2025. The court-appointed Receiver annulled the June Board election after suspicions of irregularities, particularly involving Powers of Attorney, while carefully noting that investigations had not reached final conclusions. ICANN had already called for transparency and fairness, and the Number Resource Organization warned that trust in AFRINIC governance affected the stability and cohesion of the global registry system. The rerun therefore used a direct-voter process without proxies or Powers of Attorney.

The proposed Constitution does not abolish representation. It tries to make it verifiable. It limits proxy concentration, bars candidates from acting as proxies, requires AFRINIC to complete verification before appointments become effective, and requires an original or authenticated Power of Attorney to be submitted at least 14 days before a meeting. These are defensible safeguards. My own investigation of NRS outreach to AFRINIC members showed why offers of coordinated representation deserve scrutiny when they appear amid litigation, elections and a struggle over the future of the registry.

But anti-capture measures can themselves become instruments of capture. If “the Company” has unlimited discretion to decide what evidence is sufficient, management can determine who may exercise membership rights. Verification criteria should therefore be published in advance, applied consistently and accompanied, where practicable, by expedited independent review. The objective is neither unrestricted delegation nor administrative disenfranchisement. It is traceable authority without concentrated voting power.

A coherent anti-capture doctrine must apply in every direction. AFRINIC needs protection from large resource holders converting economic weight into governance power, from strategic aggregation of proxies, from litigation used to exhaust institutional capacity, and from political or governmental control disguised as continental solidarity. It also needs protection from internal concentration: a Board that can neutralise member proposals, a NomCom that can narrow electoral choice, an Election Committee without independent verification, or management that can suspend Good Standing and thereby shape participation.

No constituency should be able to appoint, control, dismiss and replace the body that supervises access to power. That principle is more durable than any argument built around current personalities. Constitutional rules should not be written against Lu Heng, NRS, a particular Board, ICANN or any government. They should be written against the possibility that any future actor could combine economic leverage, procedural authority and institutional incumbency into control.

This is also where corporate governance must be separated from Internet number-resource governance. AFRINIC is a company limited by guarantee, and its members possess statutory and constitutional rights. But the Policy Development Process is not the private legislature of corporations holding address blocks. The draft correctly defines the PDP as bottom-up, open and transparent, with all stakeholders able to participate. Elsewhere, however, it uses the expression “member driven” when describing allocation guidelines. That phrase should be corrected. Members govern the company; the wider Internet community develops resource policy.

This separation is foundational to the RIR model. Not every operational decision requires a new policy. Registries must run services, secure systems, correct records and manage contracts every day. But substantive rules governing eligibility, allocation, transfer and reclamation should not be invented through administrative practice when they belong in a transparent community process. Board authority, operational competence and bottom-up policy are complementary only when their boundaries are visible.

The recent RPD mailing-list debate over “operational” decisions and community policy exposed the danger of collapsing either category into the other. Requiring a policy proposal for every technical act would make the registry unmanageable. Treating every contested rule as operations would make the PDP ornamental. The right test is functional: routine implementation belongs to staff; generally applicable standards that determine who may obtain, transfer or lose number resources require community authority. That boundary should be stated clearly enough that neither administrators nor organised participants can redefine it opportunistically after a dispute arises.

Stewardship after scarcity

AFRINIC’s crisis cannot be understood as a legal or electoral dispute alone. It is also a conflict over the political economy of IPv4 scarcity. The original IPv4 address space is finite. As free pools were exhausted, transfer markets expanded and address blocks acquired measurable commercial value. A regulated transfer market can improve utilisation and allow networks to obtain resources they need. The question is whether scarcity converts registry-administered rights into absolute private property and reduces the RIR to a passive ledger.

The answer should be no. AFRINIC’s ratified transfer framework seeks to permit controlled transfers while maintaining registration accuracy and reducing informal or undocumented transactions. The underlying policy text reflects years of community debate. The applicable membership agreements and registry policies matter because number resources are allocated and administered through a contractual and institutional framework; they are not simply inventory detached from the conditions under which they were issued. In a post-exhaustion environment, the registry’s task increasingly includes preserving reliable WHOIS and RDAP data, a function reinforced by AFRINIC’s resource verification process.

This is the context in which the dispute involving Lu Heng, Cloud Innovation and approximately 6.2 million IPv4 addresses must be approached. Precision is essential. According to ARIN’s 2021 account, AFRINIC issued four blocks to Cloud Innovation between July 2013 and September 2016, totalling 6,291,456 addresses. AFRINIC later stated that the addresses had been allocated following needs expressed for the region and that their use was governed by the Registration Service Agreement.

AFRINIC alleged that Cloud Innovation had breached that agreement and that contractual procedures were therefore justified. ARIN said its own routing analysis found the overwhelming majority of the addresses in use outside Africa and concluded that AFRINIC appeared to have reclamation rights under the RSA. Yet ARIN also said the substantive issues were for the Mauritian courts and cautioned the community against prejudging them.

Cloud Innovation disputes AFRINIC’s case. It has argued that it committed no policy violation, that AFRINIC misunderstood BGP routing, and that demands to monitor downstream users were impracticable. It has also publicised interlocutory court rulings in its favour. Those rulings are relevant, but procedural or interim victories do not necessarily resolve the underlying contractual merits. Lu Heng has separately described an RIR as essentially a neutral address book rather than an owner of the Internet, a view that captures the philosophical divide but does not itself answer the contractual questions.

None of this should be converted into guilt by nationality. Lu Heng’s nationality proves neither state direction nor contractual wrongdoing. Nor should the resources be described loosely as property “belonging to Africa” without reference to the RSA and applicable policy. The legally relevant questions concern the representations made, the contractual conditions accepted, the policies in force, the actual use of the resources and the remedies recognised by Mauritian law.

The new Board cannot, however, treat constitutional review as the completion of institutional reconstruction. It must pursue the 6.2 million-address dossier to a lawful, factual and institutionally defensible resolution. AFRINIC’s public court tracker shows that litigation remains active, and its May 2026 court update confirms that the dispute continues to generate consequential proceedings. The Board should use the RSA, applicable policy and Mauritian law to establish the facts and enforce any obligations that are legally demonstrated. If breach is proven, it should seek restitution, revocation, regularisation or another available remedy. If the evidence does not sustain AFRINIC’s allegations, the institution must accept that result.

AFRINIC must be strong enough to enforce its rules, and disciplined enough to enforce only its rules.

That standard rejects both impunity and vengeance. No member should become untouchable because its resources are valuable or because it can impose extraordinary litigation costs on the registry. No member should be sanctioned because it is unpopular, foreign or commercially powerful. Due process is not an obstacle to stewardship. It is what separates stewardship from arbitrary control.

The dispute has already consumed institutional attention, money and credibility. Some independent analysts, including the Internet Governance Project, have emphasised price arbitrage and governance failure rather than a simple morality play. That critique should be heard. Scarcity magnifies incentives, and weak governance invites both exploitation and overreach. The constitutional answer is not to deny markets or to surrender to them, but to make the rules governing market participation legitimate, predictable and enforceable.

A constitution for the next crisis

Serious objections to the review remain. The process may still feel fast for an institution whose Receiver’s formal discharge was pending when the current Board began its work. Some provisions appear shaped by recent adversaries. Verification powers could exclude legitimate representatives. The Board retains excessive influence over nominations, emergency policies and meeting procedure. Most importantly, Article 24 contains no actual governing-law or dispute-resolution framework because external legal advice is still pending.

These objections justify amendment, not abandonment. Article 24 should not be inserted after consultation as a technical legal addendum. Governing law, jurisdiction, internal remedies, arbitration and access to courts define how constitutional rights can be enforced. The completed language must return to the community before adoption. Likewise, safeguards designed in response to a crisis should be stated in general terms and applied symmetrically, so that they constrain tomorrow’s Board as effectively as today’s litigant.

The same principle should govern every unresolved legal question in the draft, including the scope of Board authority over termination. External counsel can identify what Mauritian law permits, but legal permissibility does not settle institutional wisdom. The community must still decide which powers should exist, which safeguards should accompany them and which decisions require independent review.

The membership should also reject an all-or-nothing approval process. Approximately 23 special resolutions, each corresponding to an article or coherent section, would allow members to adopt broadly supported reforms while returning contested provisions for improvement. Voting on every editorial change separately would fragment the Constitution; replacing the entire text through one resolution would turn constitutional consent into a forced package deal. Section-by-section approval is the defensible middle course.

The responsibility now extends beyond resource members. Network operators, engineers, civil-society organisations, researchers, governments, lawyers and other African stakeholders should read the draft, submit precise comments and support provisions that improve legitimacy. Government participation can contribute public-interest expertise, but it must remain participation rather than command. Technical expertise should inform policy without becoming technocratic insulation. Commercial interests are legitimate, but they should not purchase constitutional authority.

Abstention is not neutral. When African operators and experts withdraw from governance because the process is exhausting, contentious or imperfect, the resulting space is occupied by actors with the greatest financial incentives, legal resources or mobilisation capacity. Multistakeholder governance does not guarantee wise outcomes. It ensures only that outcomes need not be surrendered in advance to the best organised minority.

AFRINIC’s constitutional review should therefore be supported as an opportunity, not celebrated as a cure. The aim is not merely to rescue the registry from its past. It is to build an institution capable of surviving its next crisis without courts becoming its ordinary decision-making forum, without representation becoming a vehicle for capture, and without sovereignty becoming an excuse for opacity.

Africa needs a registry that can protect African stewardship without protecting its own failures, resist external concentration without reproducing internal domination, and administer scarce digital resources without abandoning due process. A Constitution cannot guarantee that future leaders will be prudent. It can make their imprudence visible, contestable and reversible. That is enough reason for the community to engage this draft, improve it and give AFRINIC a stronger constitutional foundation than the one its crisis exposed.