Introduction

AFRINIC is usually introduced in the vocabulary of internet plumbing: a regional internet registry, a member organisation, a distributor of IPv4 and IPv6 addresses, and a keeper of autonomous-system numbers for Africa and parts of the Indian Ocean. The description is true, but it now undersells the institution's economic weight. AFRINIC sits where technical uniqueness, corporate reliance, scarcity rents, court supervision, member voting and global coordination meet. From one angle it is a bookkeeper. From another, especially when a live network depends on a scarce IPv4 block, it can feel like a gatekeeper to economic continuity.

The ledger-versus-gatekeeper frame is an analytical lens, not a finding of law. A ledger records who holds what, keeps the database coherent, publishes reliable registration data, and executes community policy with procedural restraint. A gatekeeper decides who may continue using a scarce input, which post-allocation uses remain legitimate, whether a business model is tolerable, and when a holder's commercial expectations can be interrupted. Every registry has to do some gatekeeping. The danger starts when the boundary is vague, because discretion then grows faster than the procedures built to restrain it.

AFRINIC is a useful case because the facts are concrete. AFRINIC itself says it is a nonprofit, member-based organisation registered in Mauritius, entrusted with distributing and managing internet number resources including IP address space and ASNs. Its policy manual describes a bottom-up policy process, open participation and community consensus. The Number Resource Organization, in a 2023 statement on receivership, described the receiver's task as preserving continuity while elections and management were restored. The Register, KrebsOnSecurity and the Internet Governance Project have documented disputes, elections, corruption allegations, court proceedings and commercial claims around that formal structure.

Those exhibits do not all point in one direction. KrebsOnSecurity reported in 2019 on allegations that AFRINIC address records had been manipulated and that a former executive had ties to companies selling address blocks. AFRINIC's then chief executive told Krebs that an investigation was under way. IGP, writing in 2021, treated the later Cloud Innovation dispute as a political economy problem shaped by IPv4 scarcity, below-market allocation and aggressive enforcement. The Register's 2025 and 2026 reporting then tracked receivership, attempted elections, annulment, a later board election, renewed litigation and ICANN intervention.

The resulting story is not a morality play about a heroic institution or a villainous member. It is a problem of institutional economics. IPv4 addresses are not ordinary property in registry doctrine, but they are commercially relied upon and traded or leased in practice. AFRINIC's official fee schedule charges administrative membership and allocation fees; the market treats usable IPv4 as a scarce input with a price. A registry built to allocate uniqueness is therefore being asked, in public and in court, to explain how far its control over that uniqueness can reach. The question is how to preserve the ledger without pretending that ledger entries have no economic consequences.

The answer matters beyond one registry. RIR governance works because networks accept a shared allocation system instead of creating conflicting address claims. That acceptance depends on accurate records, predictable process and a belief that no participant can capture the rulebook after others have built reliance around it. If AFRINIC becomes a symbol of institutional fragility, other registries will face pressure for emergency rules. If emergency rules convert registries into political gatekeepers, the cure may weaken the very legitimacy it seeks to save.

The registry as ledger

The registry function begins with a simple technical premise: globally routed networks need unique identifiers. IPv4 addresses, IPv6 prefixes and ASNs must not be allocated in conflicting ways if the public internet is to remain coherent. A regional internet registry therefore maintains records, evaluates requests under policy, publishes registration information, supports reverse DNS and related services, and allows operators to show counterparties what resources they are authorised to use. In that role, the registry is closer to a land registry than to a ministry of industry, though even that analogy can mislead.

AFRINIC's own policy manual gives the ledger view strong support. It defines an internet registry as an organisation responsible for distributing address space and registering those addresses. It describes a hierarchy in which IANA or PTI allocates number resources to AFRINIC, and AFRINIC redistributes them to members and delegates authority to make assignments or sub-allocations where appropriate. It also distinguishes policy from general business procedures, placing number-resource rules inside a bottom-up process rather than inside unchecked staff discretion. That architecture is meant to keep the registry predictable.

A ledger is not passive. It checks eligibility, requires documentation, expects accurate WHOIS or related records, and refuses requests that do not meet policy. AFRINIC's exhaustion page says requests are evaluated under the Consolidated Policy Manual and that complete applications go through hostmaster review and approval procedures. The fee page sets categories and charges by resource size. These are real controls, but their purpose is registration, uniqueness, conservation and fairness. They are not a mandate for the registry to develop its own economic theory of how a holder should earn revenue from already-recorded resources.

The distinction is important because registry decisions have strong downstream effects. A route object, reverse-DNS delegation, RPKI certificate or WHOIS record is not merely clerical to the operator that depends on it. Banks, cloud firms, hosting companies, access providers and content networks treat IP resources as operational inputs. Customers may never know the registry exists, but their service continuity can depend on the stability of the registry's records. When the record changes abruptly, the effect can look less like a filing correction and more like removal of productive capacity.

That is why the ledger model requires restraint as much as authority. A registry that finds inaccurate contact data, unused assignments, fraud or policy evasion must be able to respond. Yet it should do so through clear criteria, proportionate remedies, independent appeal and enough notice to protect innocent downstream users. The value of the ledger is that it turns institutional trust into operational predictability. If the registry's discretion becomes open-ended, the ledger begins to resemble a licence that can be re-priced, reinterpreted or withdrawn after reliance has formed.

AFRINIC's dispute history shows how thin that boundary is. In ordinary times, a policy manual and fee schedule look technical. In scarcity, they help decide who can keep using a commercially valuable input. The more valuable the input, the more every clause becomes a bargaining instrument. The more litigation surrounds the institution, the more every ambiguous bylaw, membership category and voting rule becomes a possible weapon. Legal clarity is not an ornament for the ledger. It is part of the infrastructure.

How scarcity changed incentives

IPv4 scarcity is the economic background to nearly every modern RIR dispute. The protocol's address space is finite, and the adoption of IPv6 has not made IPv4 irrelevant because the two systems are not simply interchangeable in the short run. Networks still need IPv4 reachability for customers, legacy systems, hosting, fraud controls, reputation systems and commercial compatibility. As long as that remains true, an address that once looked like an administrative identifier can acquire the characteristics of a scarce production input.

IGP's 2021 analysis of the AFRINIC crisis made this point with unusual clarity. It argued that the dispute cannot be understood without the rise in the market value of IPv4 and the attempt to allocate regionally controlled resources at administrative prices. The exact figures in any given market report change over time, and prices vary by block quality and transaction conditions. The structural point is more durable: when an administratively allocated resource can support much higher private returns than the fee paid to hold it, arbitrage becomes predictable rather than anomalous.

AFRINIC's fee schedule illustrates the institutional mismatch. The schedule charges sign-up, allocation and annual membership fees, with categories that rise by prefix size. Those fees finance registry operations and membership services. They are not designed to capture the full market value of each routable IPv4 address. That is normal for an RIR; registries are not auction houses. But when a registry operates a need-based allocation system in a world where addresses carry significant market value, it must expect applicants, brokers, lessors and networks to optimise around the gap.

Scarcity also changes the politics of doctrine. In an abundant world, a registry can speak comfortably of stewardship, demonstrated need and community policy. In a scarce world, the same words decide who absorbs opportunity cost. If a member can lease addresses, an address block is a stream of revenue. If a registry can reclaim addresses after reassessing use, the member faces regulatory risk. If transfers are restricted by region, value depends on geography and compliance interpretation. If transfers are liberalised, critics may argue that resources intended for network development are being exported.

None of these positions is self-evidently absurd. A registry that ignores fraud or non-use will invite abuse and public anger. A registry that tries to freeze economic reality inside an earlier allocation philosophy may create black markets, selective enforcement and endless litigation. A resource holder that treats registration as absolute property may ignore the community bargain that made allocation possible. A registry that treats the holder as a temporary tenant with weak reliance rights may make long-term network planning impossible. The economics are uncomfortable because each side has a defensible interest.

The harder question is whether AFRINIC's institutional design can handle those interests. A ledger-first system would acknowledge scarcity, permit market-facing adaptation where policy allows it, and police fraud or misrepresentation through narrow rules. A gatekeeper system would retain broad discretion to decide whether a holder's current use still matches the original justification, even years later and after commercial reliance has formed. The first model risks under-enforcement. The second risks ex post control over commerce. AFRINIC's crisis is the story of that trade-off becoming visible.

AFRINIC's institutional stress test

AFRINIC entered this period with reputational damage already on the record. KrebsOnSecurity reported in December 2019 that allegations from researcher Ron Guilmette and journalists in South Africa concerned address blocks that had allegedly been commandeered from African organisations and sold through companies tied to a former AFRINIC policy coordinator. The report said the executive resigned after the allegations became public, and AFRINIC's chief executive at the time said the organisation was investigating. Those were allegations and reported facts, not a final public judgment in that article.

For the ledger versus gatekeeper lens, the lesson is not merely that corruption can occur. It is that a registry's records are themselves an asset. If historical WHOIS data, dormant company records or internal authority can be manipulated, the harm is not confined to one mistaken file. Trust in the registry's chain of title weakens. Legitimate holders may worry that records can be changed without adequate control. Outsiders may question whether scarce resources were allocated or transferred through fair procedures. A ledger that cannot protect its own integrity invites demands for stronger gatekeeping.

That demand is understandable. After an internal records scandal, a registry leadership team may feel compelled to show that it can audit, reclaim and discipline. The public expects a steward to protect common resources, and network operators expect a registry to prevent hijacking or fraudulent claims. Yet post-scandal enforcement carries its own danger. An institution trying to repair past weakness can overcorrect by expanding discretion, especially where policy language is broad and litigation risk is underestimated. In such moments, legitimacy depends on process as much as outcome.

The Cloud Innovation dispute put that tension into concrete form. IGP reported that AFRINIC questioned Cloud Innovation over alleged discrepancies between registered usage and actual deployment, alleged inconsistency with needs originally expressed, and an interpretation of membership rules tied to services in the AFRINIC region. Cloud Innovation disputed AFRINIC's position. IGP also noted that the company's agreement with AFRINIC contained language linking use of number resources to the need justified in the application. The legal significance of that language belongs to courts and contracts; the economic significance is broader.

If a registry can require renewed justification whenever a holder changes commercial use, the registry has substantial continuing power over the holder's business model. In a dynamic network market, addresses move among customers, products, regions and operating architectures. If every material change becomes an occasion for registry permission, the RIR ceases to be a ledger and begins to resemble a regulator of business plans. If the registry lacks that power entirely, a need-based allocation regime can be gamed at the moment of application and emptied of meaning afterward.

The stress test was therefore not whether AFRINIC should have any enforcement authority. It plainly must. The stress test was whether enforcement was bounded, transparent, proportionate and predictable enough to preserve confidence in the ledger. IGP argued in 2021 that AFRINIC's action against Cloud Innovation was an overreaction to earlier problems and that the regional-use premise was contested. AFRINIC and its supporters have taken a different view. The central point for readers is that the dispute exposed a gap between administrative allocation language and the economics of commercial reliance.

Enforcement and the cost of discretion

Discretion is expensive even when it is lawful. It imposes compliance costs on members, legal costs on the registry, uncertainty costs on customers and legitimacy costs on the community. The cost rises when a registry's decision can affect a holder's ability to route, lease, transfer or support addresses already embedded in contracts. AFRINIC's experience shows how enforcement that begins as resource stewardship can quickly move into courts, bank accounts, voting rules, institutional rescue and global oversight.

The legal history has many branches, but the economic pattern is clear. AFRINIC challenged a member's use of IPv4 resources. The member resisted, litigation followed, and the institution became increasingly constrained. IGP described a 2021 bank-account freeze in Mauritius as a crisis for registry operations. Later, The Register repeatedly described AFRINIC as unable to appoint a board or chief executive and unable to perform all functions during the years of dispute. Different parties attribute responsibility differently, and some allegations remain contested. But the institutional cost of discretion is no longer theoretical.

The Register's March 2026 reporting captured the competing narratives. AFRINIC accused Cloud Innovation, Larus and associated advocacy campaigns of driving litigation and procedural roadblocks. In response, Lu Heng told The Register that the structural issue was a registry model concentrating high-consequence power over economically critical number resources without matching legal and financial liability. He argued that an administrative body had become a gatekeeper over commercial activity. Those are his claims, not findings. Their significance is that they name the economic anxiety created by discretionary enforcement.

AFRINIC's rebuttal, as reported by The Register, is rooted in standard RIR doctrine: IP addresses are not owned as traditional property. That doctrine matters. If numbers were treated as ordinary private property, the community's ability to conserve, register and recover misused resources could be weakened. Yet "not property" does not mean "no reliance." A business may not own a licence, a route approval or a regulated entitlement in a full proprietary sense, but withdrawal can still destroy value and trigger due-process concerns. The law often recognises reliance without converting everything into property.

That middle ground is where a ledger-first model is strongest. It would avoid simplistic ownership language while still treating stable registration as an operational reliance interest. It would distinguish fraud at application from later change in use. It would distinguish inaccurate records from unacceptable commercialisation. It would distinguish a holder's conduct from the interests of downstream customers. It would reserve total revocation for the clearest cases and use narrower tools wherever possible. In institutional terms, it would discipline the registry's own power before disciplining the member's business.

The alternative is a spiral. Broad registry discretion provokes litigation. Litigation weakens the registry. Registry weakness prompts emergency oversight and proposals for stronger central control. Stronger central control confirms the fear that the registry has become a gatekeeper. Members then fight harder over board seats, bylaws, proxy rules and jurisdictional leverage. At each stage, the ledger becomes less visible and the political value of controlling the institution rises. AFRINIC's recent years are a warning that discretionary power can create the very capture risks it claims to prevent.

Receivership and elections

Receivership was meant to restore continuity. The NRO's September 2023 statement said the Bankruptcy Division of the Supreme Court of Mauritius had appointed a receiver, restrained AFRINIC from relocation, takeover, merger, restructuring or management control, and tasked the receiver with overseeing elections, forming a proper board and appointing a chief executive. The NRO welcomed this as a positive development that would help members keep receiving registry services and allow AFRINIC to return to functional governance. That is the official institutional position, and it is an important factual exhibit.

But receivership is a bridge, not a destination. It can preserve assets, keep staff working and stop a governance vacuum from becoming an operational collapse. It cannot by itself resolve the deeper question of what rights resource members have, how voting authority is verified, how policy should treat IPv4 commercialisation, or how courts should understand a local company performing a global coordination function. The receiver inherits those problems. If the bridge becomes the arena for another fight, receivership can become part of the risk surface rather than merely the repair mechanism.

The 2025 election sequence showed that risk. The Register reported in April 2025 that AFRINIC, without a board for years, was preparing elections under a receiver, with senior British lawyers appointed to a nomination committee because of concerns about possible interference. In June, The Register reported that ICANN had sought changes to election oversight and clarification of why Cloud Innovation appeared in corporate records in a way that raised questions. The Supreme Court ordered a communique saying that the listing was erroneous, while declining to reconstitute the nomination committee.

The same election then deteriorated. The Register reported that voting was suspended shortly before the scheduled close after questions arose about powers of attorney. It cited claims from South Africa's Internet Service Providers' Association and others that some representatives found votes or voting authority claimed on their behalf without proper authorisation. Those were allegations reported during a disputed process, and the bodies involved did not answer all media questions. The receiver annulled the election, citing concerns about voter documentation and the need to protect transparency and fairness.

Later in 2025, The Register reported that AFRINIC had elected eight directors, giving it a chance to convene a board for the first time since 2022. Yet the same report noted that the institution was not out of danger: critics questioned election arrangements, court challenges were expected, a government investigation remained in the background, and a criminal investigation into the earlier election was under way. The point is not that no board could be legitimate. It is that repeated uncertainty over who may vote and how votes are authorised turns governance into a market for procedural leverage.

For a ledger institution, that is a serious problem. The member register, voting credentials and corporate membership categories are not side issues. They determine who controls the body that controls the resource ledger. If the voting system is doubted, every subsequent policy decision inherits a legitimacy discount. If a member can obtain influence through ambiguous credentials, capture fears rise. If outsiders or official actors try to correct those fears without clean authority, overreach fears rise. AFRINIC's electoral troubles therefore belong in the same economic story as IPv4 scarcity: control of governance has value because control of the ledger has value.

ICANN, NRO and emergency guardianship

The AFRINIC crisis forced the rest of the internet-number system to confront a question it had long preferred to leave abstract: what happens if an RIR cannot function? The Register reported in 2025 and 2026 that the RIR community and ICANN were revisiting ICP-2, the policy framework for recognising RIRs, so that it would cover the full lifecycle of a registry and include assistance or derecognition mechanisms. In February 2026, The Register reported that work on the revised policy was nearing a form expected to be ready for approval later that year.

Emergency guardianship is necessary in some form. A regional registry cannot be allowed to fail in a way that creates conflicting number claims, breaks registry services or destabilises routing operations. The NRO statement on receivership emphasised continuity, member services and AFRINIC's commitments under ICP-2 and memoranda with other RIRs and ICANN. The Register's reporting also described ICANN letters and interventions after election irregularity allegations and in litigation concerning efforts to wind up AFRINIC. The official system was not idle.

Yet emergency guardianship has its own gatekeeper risk. If ICANN or other RIRs can replace, discipline or derecognise a registry, the global coordination layer gains leverage over regional self-governance. That may be unavoidable in an extreme case. But the legitimacy of such power depends on narrow triggers, public procedure and a clear separation between preserving the ledger and choosing policy winners. A mechanism designed to stop registry collapse should not become a means of imposing an official narrative about IPv4 markets, transfers or member politics.

The Register's May 2026 report gives an example of the distinction. ICANN sought to intervene in a case involving an application to wind up AFRINIC. An ICANN spokesperson said the purpose was to help the court understand AFRINIC's unique role and to clarify that numbering resources allocated through AFRINIC are not assets of AFRINIC available for distribution in a winding-up. That position is important to registry continuity. It protects the idea that number resources are coordinated public identifiers, not corporate property to be divided like office furniture.

But the same principle does not answer every economic question. Saying that resources are not AFRINIC's assets does not decide how much reliance a holder has after lawful allocation. Saying that a registry should not be liquidated does not decide whether a particular enforcement policy is proportionate. Saying that global coordination matters does not decide whether regional-use restrictions are wise or how transfer rules should treat already allocated resources. Emergency guardians should clarify the ledger for courts; they should be careful about converting that role into merits judgment on every dispute.

That caution is not anti-ICANN or anti-NRO. It is a condition of their effectiveness. If the central layer is seen as protecting the continuity of records and services, it can stabilise a crisis. If it is seen as protecting incumbent discretion from legal accountability, it may deepen resistance. A ledger-first emergency doctrine would prioritise data integrity, service continuity, transparent member status, and temporary operational support. It would leave contested economic policy to legitimate regional and global policy processes, with courts deciding concrete legal disputes on evidence rather than institutional prestige.

The economics of the gatekeeper

The gatekeeper temptation comes from a real problem: the registry has to say no. It must reject unjustified requests, prevent duplication, update inaccurate records, recover resources issued under fraud, and enforce community policy. In IPv4 scarcity, saying no feels like public-interest stewardship. Without some gatekeeping, a need-based registry can become a vending machine for private arbitrage. The issue is not whether gatekeeping exists. The issue is whether the registry's gatekeeping is constrained by rules that make it compatible with ledger trust.

Economists would describe the risk as hold-up. A firm obtains resources under one institutional understanding, builds customers and contracts around them, and later faces a registry's revised or newly aggressive interpretation of permissible use. The registry may have a plausible policy rationale. The firm may have behaved opportunistically. But the timing matters. Once reliance exists, the registry's threat to withdraw or disable resources gives it bargaining power far beyond the original application review. If that power is broad, members will price registry risk into every plan.

The opposite risk is moral hazard by holders. If a member can state a need, receive scarce resources at administrative cost, and later monetise them without meaningful constraint, the allocation system can be stripped of its public rationale. Other networks that need addresses may be crowded out. Dormant or misappropriated blocks may circulate. Abuse may increase if records are poor. The KrebsOnSecurity report about alleged internal manipulation and sales of blocks from defunct or acquired organisations is a reminder that a weak registry can be exploited from inside and outside.

AFRINIC therefore needs neither a pure property regime nor a pure permission regime. It needs a credible commitment regime. Members should know which facts are material at application, which later changes require notice, which changes require approval, which breaches justify suspension, which justify revocation, and how downstream continuity is protected. The registry should know that it can act against fraud, false records and clear policy breaches without turning every commercial evolution into a full re-litigation of the original need. Courts should see a rulebook, not improvisation.

The gatekeeper model also invites political competition. If the registry can decide market access, control of the registry becomes economically valuable. Board elections then become contests not only over stewardship but over distributional power. Proxy rules, nomination committees, membership classifications and bylaw interpretations become instruments for securing influence over scarce-resource policy. The Register's reporting on AFRINIC's election controversies, proxy allegations, board disputes and bylaw tensions should be read in this light. Governance process is not separate from resource economics; it is the route through which resource economics seeks authority.

A ledger-first model reduces the prize. It does not make elections unimportant, but it lowers the value of capturing office by binding officeholders to clearer, narrower discretion. It makes the registry's main promise boring: accurate records, consistent policy execution, fair procedure, technical reliability and transparent change control. In ordinary political language, boring can sound inadequate. In infrastructure governance, boring is often the highest compliment. The more exciting the registry becomes, the more likely it is that the ledger is being asked to do gatekeeper work.

A ledger-first reconstruction

AFRINIC's recovery should be judged less by triumphant announcements than by whether it reduces discretion while strengthening enforcement. The Register reported in February 2026 that AFRINIC said it was close to approving a budget and action plan, that staff morale had improved, and that a strategy for 2027 to 2030 was being developed. Those are positive operational signs if they are followed by durable governance repair. They do not settle the institutional economics. A board, budget and strategy can restart the machine; they do not by themselves decide what kind of machine it is.

The first requirement is a clean resource ledger. AFRINIC should be able to show that registration records, member records, contact data, reverse-DNS dependencies, RPKI-related services and transfer records are accurate or under documented correction. Historical irregularities should be handled through published remediation categories: fraud allegation, dormant-holder uncertainty, documentation gap, disputed entitlement, court-constrained status and ordinary update. The purpose would not be to shame members or expose sensitive operational details. It would be to make the registry's own state legible enough that policy is not built on fog.

The second requirement is a resource-review rulebook. Reviews should have defined triggers, scope, evidence standards, timelines, member response rights, confidentiality protections and appeal channels. Random or politically motivated review should be excluded. So should unlimited fishing expeditions into downstream customer use unless policy clearly authorises such inquiry and explains why it is necessary. If the registry needs information to verify compliance, it should ask for the minimum necessary information, protect commercial confidentiality and explain how the information maps to a specific rule.

The third requirement is proportional remedy. Not every breach should lead to revocation. Some record defects require correction. Some payment failures require ordinary billing enforcement. Some usage changes require notice or reclassification. Some misrepresentations require suspension of further allocations. Fraud may require recovery. Downstream users need transition protections where possible. The registry should make clear when it is protecting uniqueness and data integrity, when it is enforcing scarcity policy, and when it is dealing with contractual breach. Mixing these categories increases litigation risk and weakens public understanding.

The fourth requirement is separation of functions. The same institution can maintain records, develop policy, adjudicate disputes and enforce contracts only if internal separations are credible. A ledger-first AFRINIC would strengthen independent appeals, publish board-conflict controls, keep staff implementation separate from political campaigns, and prevent election disputes from directly contaminating operational registry decisions. Where outside experts are needed, their mandate should be narrow and public. The receiver experience suggests that even repair mechanisms can become controversial when authority, process and communications are not clear enough.

The fifth requirement is legal translation for courts. Courts in Mauritius are being asked to decide disputes involving a local company that performs a cross-border coordination function. Judges do not need mystical language about internet uniqueness. They need a precise map of what resources are, what AFRINIC owns, what it administers, what members rely on, what can be transferred, what cannot be distributed in insolvency, and what operational harm follows from abrupt changes. ICANN's intervention on the non-asset character of numbering resources is useful, but it should be part of a broader explanatory discipline.

What to watch next

The first watchpoint is whether AFRINIC's board legitimacy stabilises. A board can exist on paper and still operate under a cloud if election procedures remain disputed. The important signals are not slogans about unity, but documentary signs: finalised election records, resolved court challenges, clear membership classifications, verified voting authority, published conflict policies and ordinary board minutes. If those appear, the institution becomes less dependent on personality. If they do not, every policy change will be filtered through the suspicion that control of the registry is still contested.

The second watchpoint is the treatment of pending and future litigation. The Register's 2026 reporting described active court proceedings over winding-up, interim orders concerning public statements, bylaw disputes and continued accusations among AFRINIC, Cloud Innovation, Larus and advocacy actors. Many claims are contested, and public reporting is incomplete. Readers should watch for whether litigation narrows toward specific legal questions or expands into a permanent arena for disabling institutional action. A functioning ledger can coexist with litigation. A registry whose every operational act becomes a court front cannot easily rebuild trust.

The third watchpoint is the revised ICP-2 process. A lifecycle policy for RIRs is overdue. The absence of a mature failure-and-repair mechanism made AFRINIC's crisis harder. But the substance matters. A good policy will define assistance, remediation, emergency continuity and derecognition with high thresholds and transparent procedure. A bad policy will give the global layer broad discretion while leaving members uncertain about regional autonomy. The difference is whether the policy protects the shared ledger or creates a larger gatekeeper above the regional gatekeeper.

The fourth watchpoint is IPv4 transfer and leasing doctrine. AFRINIC's 2026 environment cannot be governed as if IPv4 scarcity were a temporary irritant. The Register reported an AFRINIC official saying in February 2026 that unallocated IPv4 remained and that the conversation should move toward IPv6 once that pool reaches zero. IPv6 deployment is essential, but it will not erase IPv4 reliance quickly. The registry needs policy that recognises market behaviour without surrendering to abuse. Pretending that commercialisation does not exist gives discretion to whoever enforces the fiction.

The fifth watchpoint is whether AFRINIC can avoid rhetorical inflation. Official defenders sometimes speak of continuity as if it settles every contested issue. Registry critics sometimes speak of discretion as if it invalidates every act of stewardship. Neither habit helps. The strongest public-interest case for AFRINIC is not that it should be trusted because it is an RIR. It is that it can demonstrate accurate records, fair procedures, proportionate enforcement and accountable governance. The strongest market case for resource holders is not that addresses are ordinary property. It is that operational reliance deserves predictable treatment.

The ledger-versus-gatekeeper lens points to a sober conclusion. AFRINIC should remain a registry, not become an industrial-policy authority for IPv4. It should enforce rules, but the rules should be clear before enforcement threatens business continuity. It should reject the idea that number resources are corporate assets available for liquidation, while also rejecting the lazy inference that "not property" means "no reliance." It should accept help from ICANN, the NRO and courts where continuity requires it, while resisting any repair that turns emergency guardians into permanent policy masters.

AFRINIC's crisis is not only an African institutional story. It is a preview of what happens when administrative internet resources become economically significant before governance language catches up. The internet-number system was built to make uniqueness boring. IPv4 scarcity made it valuable; litigation made that value visible. The task now is to make the registry boring again without pretending the value disappeared. That means protecting the ledger, narrowing the gatekeeper, and designing procedures strong enough that no faction has to capture the institution in order to trust the record.