The shock behind the quarrel
IPv4 scarcity did more than make old internet addresses expensive. It changed the political economy of the institutions that record, allocate and police them. A regional internet registry can look like a quiet technical body when the pool it administers is abundant, an allocation is mostly an entry in a database, and disputes can be handled as routine membership administration. Once the residual pool becomes finite, tradable and capitalized into business plans, the same registry looks very different. It becomes the point where a public coordination function meets private balance sheets, legal claims, customer continuity and regional development policy.
AFRINIC is the sharpest case because it had to administer a scarcity regime while its own legitimacy was under repeated stress. The African Network Information Centre serves Africa and the Indian Ocean region. Its own public materials describe a member-based, nonprofit organisation registered in Mauritius, responsible for distributing and managing IPv4, IPv6, and autonomous system numbers. Its service catalogue also includes WHOIS, RDAP, reverse DNS, routing-registry services, DNSSEC, and resource certification. In ordinary times those descriptions sound administrative. In the scarcity era they describe a ledger with consequences: the registry record affects routing, financing, customer assurance, legal standing, and the ability of operators to use scarce IPv4 capacity without renumbering live networks.
The economic shock is easy to miss if the issue is reduced to an AFRINIC controversy, a Cloud Innovation dispute, a court fight in Mauritius, or an argument about whether address holders "own" number resources. Each description catches part of the truth and loses the larger pattern. The deeper question is what happens when an institution designed to keep a reliable ledger also acts as a discretionary gatekeeper over an asset-like resource that the market has already priced. Scarcity does not abolish policy. It makes policy more consequential. It does not turn every address holder into an owner in the property-law sense. It does mean that revocation, transfer delay, opaque eligibility review, or uncertain registry continuity can impose costs that look very much like capital impairment.
Lu Heng's public notes frame the issue in terms that are useful even to readers who disagree with his commercial position: protect the ledger, not the gatekeeper. That distinction matters. A registry earns legitimacy when it keeps accurate records, applies published rules predictably, resolves disputes through accountable procedures, and preserves continuity for the networks that depend on it. It loses legitimacy when the market sees it as a single institutional chokepoint able to reinterpret eligibility, usage, voting, transfer, or continuity rules after capital has already been committed. The first model reduces scarcity cost. The second adds a gatekeeper premium on top of scarcity itself.
AFRINIC's recent history shows all the moving parts. Its exhaustion page records entry into IPv4 Exhaustion Soft Landing Phase 2 on 13 January 2020. Its fee schedule ties annual membership fees and allocation or assignment fees to the resources held, but those fees are service and membership charges rather than market prices for an asset. The Internet Governance Project's 2021 analysis put numbers on the price gap, noting IPv4 market prices rising from about $8 per address in 2017 to about $30 by 2021, with a /16 block potentially worth roughly $2 million. KrebsOnSecurity reported in 2019 that allegations about altered AFRINIC records and address sales had entered public view. The Number Resource Organization reported in September 2023 that the Supreme Court of Mauritius had appointed an official receiver for AFRINIC to preserve the business, oversee elections, and move toward a restored board and CEO. The Register then chronicled the troubled election path through 2025 and the continuing litigation and intervention risk in 2026.
The public reader should not collapse these events into a single moral tale. Allegations are not judgments. Party statements are not neutral evidence. Official registry pages are factual exhibits, not proof of institutional virtue. The important connection is institutional and economic. A shrinking IPv4 pool created rationing. Rationing created transfer pressure. Transfer pressure made leasing and secondary-market structures more attractive. Market value made registry discretion more expensive. Legal instability made ledger continuity more valuable. AFRINIC's legitimacy is therefore being tested not because IPv4 is technically exotic, but because scarcity turned a technical record into a capital surface.
From stewardship to rationing
The registry model began with stewardship language. Internet number resources are globally unique identifiers. They must be recorded somewhere, delegated through some hierarchy, and updated through some accountable process. AFRINIC's policy manual describes a bottom-up policy development process in which proposals are submitted, debated by the community, and adopted through the registry's procedures. It states the principles of openness, transparency, and fairness. It treats public address space as a resource to be managed in the regional internet community's interest, not as ordinary inventory to be auctioned to the highest bidder.
Scarcity did not invalidate that model. It changed what stewardship required. A registry with abundant IPv4 space can rely heavily on needs assessment and member documentation because a mistaken allocation may be administratively regrettable without being systemically explosive. A registry with a final pool is different. Every allocation has an opportunity cost. Every denial may push an operator toward the transfer market. Every delay can force customers into leasing, renumbering, carrier-grade NAT, or postponement of deployment. Every discretionary review affects not just the requesting member but the implied value of other blocks and the risk premium that buyers, lessors, lenders, and customers attach to address continuity.
AFRINIC's Soft Landing policy is the point at which this change became visible in public rules. The registry's exhaustion page explains that since 2005 AFRINIC had managed a pool of internet number resources and delegated them to organisations that could justify need. It also says IPv4 resources are scarce and that the community supported a Soft Landing policy in 2011 to guide exhaustion, conserve the pool, and support transition to IPv6. In 2017 AFRINIC entered Phase 1. On 13 January 2020 it entered Phase 2. Under the published Phase 2 framework, requests are processed through tickets, complete applications proceed to evaluation, members must satisfy contractual checks, and the minimum and maximum IPv4 allocation or assignment sizes are /24 and /22.
Those details matter because they turn policy into rationing. A /22 maximum is not just an engineering parameter. It says late-stage IPv4 demand will be met in small increments, if at all. First-come, first-served processing and completeness rules become allocation machinery. A 90% efficient-use expectation becomes a gate through which additional demand must pass. The final-pool framework was not a price mechanism; it was a rationing mechanism built from documentation, eligibility and registry review. It may have been a reasonable conservation strategy. It still placed the registry in the middle of a market that had begun to reveal a price for what the rationing system distributed at administrative cost.
The contrast with other RIR regions sharpened the pressure. AFRINIC's exhaustion page notes that by 24 September 2015, APNIC, ARIN, LACNIC, and the RIPE NCC had already exhausted their free IPv4 pools and were allocating from the final /8 received from IANA. AFRINIC entered that late-scarcity period with a different pool position. The Internet Governance Project emphasised that AFRINIC had only ever held a small share of global IPv4, but that it was, for a time, the remaining region with a large pool available through administrative allocation. That asymmetry created a price gradient: in exhausted regions operators increasingly had to obtain IPv4 through transfers, while in AFRINIC the rules still offered a route to addresses through membership, need, and fees far below secondary-market value.
A price gradient invites arguments about fairness. One argument says Africa's remaining pool should be preserved for African networks and regional development. Another says a globally routed number resource cannot be kept inside a regional economic wall without creating arbitrage, corruption risk, and coercive registry power. Both arguments have internal logic. The policy failure begins when the registry behaves as if the economic gradient does not exist. If the institution insists that the only relevant fact is formal eligibility, while the market sees millions of dollars in potential value, the registry will be surprised by the intensity of disputes. If holders insist that market value alone should decide usage and transfer, they will be surprised by the endurance of public-resource rhetoric and community policy constraints.
The institutional task is therefore not to pretend that scarcity can be administered as if it were abundance. It is to build rules that acknowledge scarcity without giving the registry unlimited discretion. Rationing can be legitimate if it is clear, prospective, proportionate, and reviewable. It becomes dangerous when it turns the registry into a judge of changing business models, customer geographies, or economic motives without equally strong procedural constraints. AFRINIC's Soft Landing rules were designed to manage the pool. The legitimacy test is whether the institution can manage the pool without converting scarcity into a permanent license to second-guess the networks and businesses that depend on recorded number resources.
Administrative fees and market value
The central economic fact is the gap between AFRINIC's administrative fee structure and the market value of IPv4 addresses. AFRINIC's fee schedule says the organisation charges members to support its operations, with annual membership fees based on categories derived from billable resources held. For LIRs, the table places a /16-to-less-than-/14 category in the medium band, with annual fees far below the market value that IGP assigned to a /16 in 2021. The same schedule lists allocation fees for approved resources and separate end-site, IPv6, ASN, academic, critical-infrastructure, and transfer-related rules. The point is not that AFRINIC was secretly selling property too cheaply. It is that the institution was charging service fees in a world where the resource recorded through that service had become capitalized.
This difference between fee and value is common across registry systems, but AFRINIC's timing made it acute. A member paying annual fees is not buying ordinary title to an address block. The registry and much of the RIR community have long resisted property language. The Register's 2026 reporting recorded AFRINIC's view that IP addresses are not owned as traditional property, even while acknowledging that addresses are bought, sold, and leased. That tension is not a semantic inconvenience. It is the heart of the scarcity problem. A resource can be non-property in registry doctrine and still carry market value for those who depend on it. A contract can describe use rights, custody, policy compliance, and revocation while the surrounding market prices those rights according to continuity, routability, and transferability.
Capital does not wait for doctrinal agreement. An operator using a block in production has invested in network architecture, customer contracts, abuse handling, reputation, firewall rules, geolocation, routing policy, reverse DNS, and operational processes. A buyer in the transfer market prices expected control over a block. A lessee pays for the ability to use capacity without funding a purchase. A lender or investor may discount a business whose number-resource continuity is uncertain. The fact that the registry does not call the resource property does not make these investments disappear. It shifts the argument to enforceability, revocability, and the credibility of the registry ledger.
That is why IPv4 leasing has become central to the dispute. Leasing is not merely a way to monetise addresses. It is a way to separate use from direct registry holding, upfront capital expenditure from operating expenditure, and customer continuity from the legal exposure of a direct RIR account. LARUS's public materials market "first-party IPv4 leasing" on precisely this continuity logic: fewer intermediary layers, direct lessor accountability, and upstream absorption of registry-side risk. NRS presents the issue more politically, arguing that registry discretion has become economic power and that holders should control their IP assets. These are participant claims, not neutral adjudications. They are nevertheless evidence of how scarcity is being understood by market actors: the address is valuable, but the reliability of the registry relationship is part of the product.
This creates a paradox for a registry that wants to reduce speculation. The more discretionary the registry becomes, the more valuable continuity structures become. If direct holding exposes an operator to unpredictable policy review, possible revocation, costly litigation, or unclear transfer treatment, the operator may prefer a leasing provider that claims to absorb that upstream risk. If registry rules make transfers slow or regionally constrained, holders may lease rather than sell. If registry legitimacy is weak, customers may pay a premium for a counterparty that can promise legal stamina. A gatekeeper trying to suppress market behavior can therefore intensify it by making certainty scarce as well.
Scarcity cost is layered. The obvious layer is the direct market price of an address. Around it sit the administrative cost of obtaining, documenting and maintaining the registry relationship; the legal cost of defending a use right; the operational cost of renumbering or replacing a block; the reputational cost of losing clean routing history or customer trust; and the option value of holding spare IPv4 capacity when future supply is uncertain. When the registry is perceived as a reliable ledger, those layers remain more separable. When it is perceived as a discretionary gatekeeper, they combine into a larger risk premium.
This premium is paid by more than speculators. A national telecom operator, a hosting company, a university network, an internet exchange, a cloud platform, or a public-service provider may not trade IPv4 aggressively. Yet each depends on predictable registration and policy execution. If AFRINIC cannot allocate residual space, process transfers, update records, maintain reverse DNS, support RPKI, or resolve membership questions without legal or procedural turmoil, ordinary operators pay through delay and uncertainty. The economics of scarcity therefore cannot be confined to headline valuations or courtroom claims. It appears in procurement choices, network planning, the timing of IPv6 transition, and the willingness of investors to fund IPv4-dependent services in the region.
Transfers, leasing, and the regional border
The transfer problem sits where regional policy meets a global routing system. AFRINIC's policy manual contains a section on IPv4 resource transfers within the AFRINIC region, added in 2017. Its fee schedule separately says transfers must comply with whatever transfer policy or guideline is in force, and it distinguishes transfers between existing resource members from transfers to a new organisation. The transfer language is administrative, but the economic question is larger: when an address is globally routable, how much regional control can a registry impose on its use or movement before the control itself becomes a source of scarcity cost?
IGP's 2021 analysis argued that the attempt to put a regional border around AFRINIC-issued addresses was the structural cause of the Cloud Innovation dispute. The article described Cloud Innovation as having received rights to millions of IPv4 numbers from AFRINIC and leasing them to customers, many outside Africa. It also described AFRINIC's 2020 and 2021 correspondence asserting concerns about discrepancies between registered usage and actual countries of use, the original needs justification, and the requirement that members originate services in the AFRINIC service region. Cloud Innovation disputed AFRINIC's interpretation, arguing that business use evolves and that constant re-justification would turn the registry into a central planner for network operations.
This dispute should be read with care. The fact that Cloud Innovation had a commercial interest does not answer the policy question. The fact that AFRINIC had policy concerns does not prove that unlimited revocation discretion was legitimate or prudent. The fact that IPv4 leasing exists does not establish that every leased block is abusive. The fact that addresses are globally routed does not make regional policy irrelevant. The public record supports a narrower conclusion: once market value rose, both sides had incentives to treat registry interpretation as existential. The holder saw resource withdrawal as a threat to customers and revenue. The registry saw out-of-region usage and leasing as a challenge to the premise of regional allocation.
The legitimacy problem is compounded when rules are applied after the fact. Operators change customers, deploy in multiple countries, use cloud and transit arrangements outside incorporation jurisdictions, and shift allocations as business models evolve. A policy system can require documentation and accuracy. It can penalise fraud. It can distinguish assignment, sub-allocation, leasing, and transfer. But if ordinary network evolution triggers a broad new needs assessment, the address holder is never finished acquiring the right to use the block. The capital value of the block is then discounted by the possibility of future reinterpretation.
Regional-use restrictions are attractive because they promise to keep scarce resources available for the region that received them. They also create enforcement problems. A customer may be incorporated in one country, serve users in another, announce routes from a third, and contract with infrastructure providers elsewhere. A hosting provider may use African-registered resources for international customers while still operating in the region. A content platform's address usage may not correspond neatly to the place where value is created. The more the registry tries to police these distinctions through discretionary approval, the more it must inspect business models rather than maintain the ledger.
Leasing exposes the weakness of a rigid regional-border theory. If a holder cannot transfer a block freely but can lease use of it, the economic benefit of the block can still move. If the registry tries to prohibit or punish leasing without clear prospective rules, the dispute moves into contracts, courts, and public campaigns. If the registry accepts leasing but requires disclosure, abuse handling, accurate contact data, and continuity controls, it can reduce operational harm while acknowledging market reality. None of these choices is easy. The worst approach is to deny the economic function of leasing while exercising broad power over it case by case.
AFRINIC's 2026 disputes show the point. The Register reported AFRINIC accusing Cloud Innovation, LARUS, and associated campaigns of trying to paralyse the registry through litigation and procedural roadblocks. The same report recorded Lu Heng's response that the structural issue is high-consequence power over economically critical number resources without commensurate legal and financial liability. Later reporting described a Larus press release about a first-party IPv4 leasing platform, AFRINIC's response that a Mauritius court order had not approved leasing or commercialisation of AFRINIC-allocated resources, and an interim order directed at statements implying judicial approval of leasing or monetisation. The narrow public lesson is that leasing, legal status, and registry authority have become inseparable in the market's perception.
That inseparability is precisely why a registry must be predictable. A holder should not be able to launder unsupported claims through court language or marketing copy. A registry should not be able to turn every economic use it dislikes into a discretionary threat. The public interest lies in a ledger that records who has which resources, under what public rules, with what transfer path, what dispute process, and what operational obligations. The more those questions are settled by communiques and injunctions, the more scarcity cost rises.
Record integrity and the ledger function
Scarcity makes the registry database more valuable, and value makes record integrity more important. KrebsOnSecurity's 2019 reporting showed why. Based on work by researcher Ron Guilmette and South African reporting, Krebs described allegations that an AFRINIC policy coordinator, Ernest Byaruhanga, had links to companies involved in selling African IPv4 blocks and that official records had been changed around blocks associated with defunct or acquired organisations. Krebs reported that Byaruhanga had resigned and that AFRINIC's then chief executive said the organisation was aware of the allegations and investigating. The cited material did not provide a final public adjudication. It did, however, put the integrity of registry records into public dispute before the Cloud Innovation litigation became the main story.
This matters because the ledger is the registry's core product. Policy debates can be fierce, but they depend on a stable record of which resources exist, who holds them, what status they have, and how changes are authorised. A registry can survive disagreement over transfer policy. It is much harder to survive doubt over whether records can be altered, restored, challenged, or explained transparently. Under scarcity, a mistaken or unauthorised record change is not just a clerical error. It can shift millions of dollars in implied value, affect routing reputation, and change who can claim standing in later disputes.
The distinction between protecting the ledger and protecting the gatekeeper is useful here. Protecting the ledger means preserving accurate, auditable, durable records even when the institution around them is under stress. It means records should not depend on factional control of a board, a receiver's discretion, a delayed election, or the mood of a policy dispute. Protecting the gatekeeper, by contrast, means treating institutional authority as self-validating: the registry did it, therefore the record is legitimate; the registry says the holder is non-compliant, therefore revocation is justified; the registry says continuity is preserved, therefore members need not worry. Scarcity makes that posture untenable. The ledger must be more trustworthy than the officeholders who administer it.
AFRINIC's own service list underscores the dependency. WHOIS and RDAP records, reverse DNS, routing-registry entries, RPKI, and member-service systems are not decorative. They are how network operators, counterparties, abuse desks, auditors, and customers observe legitimacy. If those systems continue to function while governance is impaired, the registry can preserve some confidence. If their update processes become slow, contested, or legally uncertain, the market starts pricing operational risk into every AFRINIC-linked block.
Record integrity also affects the election story. A member-based registry derives governance authority from who can participate in its processes. The Register's 2025 reporting described concerns over credentials, powers of attorney, voter documentation, and the June 2025 election annulment. IGP's June 2025 article discussed confusion around Cloud Innovation's classification in corporate records and the Mauritius court's treatment of that issue. These are not just election-process footnotes. In a registry, the member record helps decide who can choose the board that oversees the policy process that governs resource records. Ledger continuity, member legitimacy, and policy authority are nested inside one another.
This is why post-crisis recovery cannot be measured only by whether AFRINIC has directors. A board is necessary, but not sufficient. The registry must be able to show that resource records are stable, changes are auditable, member categories are legally coherent, transfer and leasing interpretations are prospective and published, and disputes have predictable appeal paths. In a scarce-resource environment, "trust us" is too expensive. The market will ask for proof because the cost of being wrong is not an abstract governance embarrassment. It is renumbering, litigation, customer loss, or stranded capital.
Courtrooms as scarcity institutions
The AFRINIC crisis also shows that courts become part of scarcity governance when registry rules and market value collide. In 2021 IGP reported that the Supreme Court of Mauritius had provisionally frozen up to $50 million in AFRINIC bank accounts in connection with the Cloud Innovation dispute. IGP criticised both AFRINIC's risk management and Cloud Innovation's legal escalation, while also questioning the proportionality of the bank-account freeze. Whatever one thinks of that interpretation, the institutional lesson is plain: a registry decision over number resources can trigger remedies that affect the registry's own operating capacity.
The NRO's September 2023 statement marked the next stage. It reported that the Bankruptcy Division of the Supreme Court of Mauritius had appointed an official receiver for AFRINIC under the Companies Act. The receiver's role, as summarised by the NRO, was to maintain the status quo of AFRINIC's assets, preserve the value of the business, oversee an election process under AFRINIC's constitution, facilitate formation of a proper board, and appoint a CEO. The NRO framed the appointment as a positive development for service continuity and a path back to functional governance. That was an official coordination-body view, useful as a factual account of the receiver's mandate and as evidence that continuity had become a concern across the RIR system.
Receivership is often described as a sign of failure. It may be better understood as the moment ordinary corporate law became the temporary container for a critical technical function. AFRINIC is locally incorporated, regionally consequential, and globally coordinated. Its legal personality is Mauritian; its operational impact extends across the African internet and into global routing. Courts cannot ignore the company. The internet community cannot ignore the registry function. The receiver sat between those realities.
The court process did not remove the scarcity problem. It moved it into a different institutional setting. If the receiver preserves the business but cannot produce an accepted election, uncertainty continues. If courts recognise claims that affect member status, voting rights, or communications about leasing, those orders influence the market's assessment of registry risk. If ICANN seeks to intervene in winding-up proceedings, as The Register reported in May 2026, the court must hear arguments not only about corporate disputes but about whether number resources can be treated as assets of the company and whether dissolving the registry is compatible with the continuity of the numbering system.
This is where the economics of scarcity tests old assumptions about private internet governance. The RIR model depends on community policy, member legitimacy, and contractual relationships. It also depends on domestic legal systems when contracts break down. IGP's 2023 article interpreted receivership as evidence of resilience: rule of law preserving the registry while leadership is replaced. That is one reading. A more cautious reading is that legal backstops can prevent collapse, but they cannot by themselves restore legitimacy. Courts can appoint a receiver, restrain relocation or restructuring, order communications, and decide standing. They cannot make members trust policy execution if the registry continues to behave unpredictably.
The economics of court involvement are significant. Litigation locks management attention and money. It delays allocation and transfer decisions. It increases due diligence costs for anyone touching AFRINIC-linked resources. It invites public campaigns by participants trying to shape member opinion or legal interpretation. It may affect whether operators prefer direct holding, leasing, or resources from other RIR regions. It also creates precedent risk: if one registry can be paralysed through litigation over scarce IPv4 resources, other registries and market participants will update their assumptions.
In that sense, the courtroom is not external to the scarcity market. It is one of the places where scarcity is priced. A holder with deep legal resources may be seen as more able to defend continuity. A registry with weak governance may be seen as more vulnerable to injunctions. A court order touching records, elections, communications, or winding-up can change counterparties' willingness to contract. The fewer clear ex ante rules the registry provides, the more the court becomes the ex post rule maker.
Elections and the price of legitimacy
AFRINIC's board crisis made the legitimacy cost visible. The Register reported in April 2025 that AFRINIC was preparing elections after years without a board or chief executive, with a receiver appointing senior British lawyers to oversee nominations because of concerns about potential interference. The same article noted that AFRINIC serves 54 nations across Africa and the Indian Ocean and had been unable to appoint a CEO or elect board members since 2022. A registry with no board can keep parts of the ledger running, but it cannot easily claim full policy legitimacy.
The June 2025 election was supposed to relieve that condition. It did not. The Register reported that the election was suspended minutes before the in-person voting period ended because of questions around powers of attorney, then annulled by the receiver after concerns about voter documentation. South Africa's ISP Association and others alleged irregularities around proxy voting. ICANN sent questions and warned of possible compliance review. AFRINIC remained in the limbo the election had been designed to end. The immediate issue was electoral procedure. The larger issue was whether the body that controls scarcity rules could prove that its own governance machinery was reliable.
September 2025 brought a more positive but still incomplete turn. The Register reported that AFRINIC held fresh elections and announced eight directors, creating a chance to convene a board for the first time since 2022. But the same account warned that the board faced critics, possible legal challenges, investigations, and continuing disputes. In February 2026, The Register reported AFRINIC's head of capacity building telling APRICOT that morale had improved, interim management personnel had been appointed, a budget and action plan were imminent, and AFRINIC still had 773,376 unallocated IPv4 addresses. That was evidence of recovery. It was not evidence that the scarcity-governance problem had been solved.
March and May 2026 reporting reinforced the point. AFRINIC accused Cloud Innovation, LARUS, and associated campaigns of trying to paralyse the registry through litigation and procedural roadblocks. Lu Heng responded that the issue was structural power over economically critical number resources. ICANN later intervened in proceedings linked to a winding-up application, telling The Register it wanted the court to understand AFRINIC's unique role and that numbering resources administered through AFRINIC were not assets available for distribution in a winding-up. These developments show a registry moving toward ordinary governance while still contested at the level of legal status, economic power, and continuity.
Legitimacy in this setting is not ceremonial. It has a price. A board whose election path is disputed faces a higher cost of making difficult allocation or transfer decisions. A receiver whose process is questioned has a lower ability to settle member confidence. A registry whose bank accounts, bylaws, member classifications, or communications are under litigation cannot easily persuade the market that policy risk is low. Even a correct decision becomes more expensive if counterparties assume it will be challenged.
The legitimacy price is paid in delay. It appears when operators wait for clarity before requesting resources. It appears when transfer participants demand warranties, indemnities, or lower prices. It appears when lessors market themselves as continuity shelters. It appears when ICANN and peer registries spend time designing or revising lifecycle rules for dysfunctional RIRs. It appears when public readers must distinguish between registry communiques, party claims, court orders, and independent reporting to understand whether the institution is actually recovering.
The election story also exposes a weakness in treating community as a magic word. Community policy is legitimate only if the relevant community can participate through coherent records, fair voting, transparent procedures, and accountable dispute channels. If resource members have bylaw rights that can be challenged under local corporate law, as The Register reported in connection with ISPA's bylaw analysis, the formal rhetoric of member governance may sit on an unstable legal base. Scarcity makes that instability material. Whoever controls the board helps determine how the remaining pool, transfer rules, leasing disputes, and review powers are handled.
The gatekeeper premium
A discretionary gatekeeper increases scarcity cost in ways that a neutral ledger does not. The first cost is uncertainty. If holders do not know whether a change in customer geography, leasing structure, routing pattern, or business model will trigger retrospective review, they must reserve capital and legal attention for registry risk. They may hold more addresses than needed, avoid transfers, charge customers more, or route through contractual structures designed for survivability rather than efficiency.
The second cost is litigation. Broad discretion invites broad challenges. AFRINIC's attempted enforcement against Cloud Innovation was not processed as a quiet membership matter; it became a multi-year legal conflict. IGP's 2021 account described the bank freeze and multiple court cases. The Register's 2026 reporting described continuing litigation, winding-up attempts, takedown orders, and ICANN intervention. A narrow, published, proportionate enforcement path would not eliminate litigation, but it would reduce the number of questions courts must answer. The more the registry relies on discretionary assertions, the more legal process becomes the market's substitute for predictable rules.
The third cost is due diligence. Anyone buying, leasing, financing, or relying on AFRINIC-linked IPv4 must ask not only whether the block is routed and recorded, but whether the holder's use could be challenged, whether transfer restrictions apply, whether litigation might affect continuity, whether member status is secure, and whether the registry can process updates. These questions are not free. They consume legal work, engineering review, contract negotiation, and risk discounts. Scarcity already makes IPv4 expensive. Gatekeeping makes the surrounding transaction more expensive still.
The fourth cost is political capture. When a registry controls a valuable bottleneck, board seats, nomination rules, proxy campaigns, bylaw wording, and member classifications become commercially attractive. The Register's accounts of election interference concerns, powers of attorney, Smart Africa endorsements, ICANN letters, and NRS campaigning illustrate how governance channels become part of the scarcity contest. That does not mean every participant is acting in bad faith. It means that scarcity makes governance worth fighting over.
The fifth cost is weakened conservation. This may seem counterintuitive. A registry might believe strict gatekeeping conserves the pool. But if the market expects arbitrary enforcement, rational operators may hoard addresses to avoid future dependency. They may resist returning unused space because they fear they will not be able to obtain resources later. They may use leasing arrangements that reduce transparency. They may delay IPv6 transition because IPv4 continuity risk absorbs management attention and capital. Conservation works best when rules are trusted. It works poorly when the conservation authority is seen as a threat.
The sixth cost is reputational spillover. AFRINIC is one registry, but the RIR system is a shared institutional model. AFRINIC's crisis prompted work on revised lifecycle rules for RIRs, including possible derecognition or emergency arrangements. That response may strengthen the system, but it also signals to markets that registry continuity can no longer be assumed. If one RIR's governance becomes unstable, every RIR participant asks whether similar contractual, policy, or legitimacy risks exist elsewhere. LARUS's public comparison of public RIR contract logic across regions is a participant's commercial use of that spillover. The argument resonates because the registry model everywhere combines administrative language with high-consequence control.
None of this means registries should abandon policy. The opposite is true. Scarcity requires better policy, not less. The policy must be prospective, intelligible to operators, economically realistic, and procedurally constrained. It must distinguish fraud from ordinary business evolution, inaccurate records from lawful leasing, transfer evasion from transparent market use, and regional development goals from impossible attempts to keep globally routed identifiers inside a regional container. It must also state what remedies are proportionate. Withdrawal of a live block is not the same as a request for updated documentation. Suspension of a member's voting rights is not the same as correction of a registry contact. The remedy scale must match the harm and must account for customer continuity.
The economic test is simple: does the registry reduce the cost of scarcity by making rights of use more legible, or does it increase the cost by making continuity depend on discretionary tolerance? AFRINIC's future legitimacy depends on answering that question through conduct, not communiques.
What admitting scarcity would require
An honest scarcity policy would start by acknowledging that IPv4 has economic value even where the legal form is not ordinary ownership. The registry does not need to call addresses property. It does need to recognise that recorded use rights can support revenue, customer obligations, financing assumptions, and operational continuity. Once that is admitted, policy can focus on the real public interests: accurate records, abuse accountability, routing stability, fair access to the remaining pool, transparent transfer paths, and protection against fraud.
The next step is to narrow discretion. Needs-based allocation can remain, but the evidence required, review period, appeal path, and remedy range must be clear before the member invests. Resource reviews should not operate as open-ended audits of every future business change. If the registry believes a particular practice, such as undisclosed leasing or out-of-region transfer by another name, harms the region, it should define the practice prospectively and consult the community on enforcement. A registry that relies on surprise reinterpretation is not conserving a public resource. It is taxing continuity.
A more mature policy would also separate functions that scarcity has fused together. The ledger function should be technically conservative and highly auditable. Dispute resolution should be independent enough that parties do not view registry staff as prosecutor, judge, and record keeper. Economic-rights questions, including transfer and leasing treatment, should be handled through published policy rather than ad hoc letters. Governance participation should be reconciled with the registry's local corporate law so that resource members know which rights are operational, which are legal, and which require bylaw repair.
AFRINIC's remaining IPv4 pool adds urgency. The February 2026 report that AFRINIC had 773,376 unallocated IPv4 addresses is not a sign of abundance. It is the tail of a pool whose final allocation decisions will be watched precisely because the numbers are small. When the pool is nearly exhausted, the marginal value of legitimacy rises. A transparent /22 issuance under Phase 2 may matter less in volume than a large historical allocation, but it matters greatly as evidence of whether the institution can still apply rules without triggering another legitimacy fight.
Transfers and leasing need realism. A registry can require accurate holder data, good-standing status, abuse contacts, routing-security hygiene, and disclosure sufficient to prevent fraud. It can restrict transfers in ways adopted through legitimate policy. But it cannot make global demand disappear by refusing to name it. If leasing is treated only as suspicion, it will move into less transparent forms. If transfers are treated only as leakage, price pressure will appear as litigation, shell structures, or defensive holding. If market use is acknowledged and bounded, the registry can lower transaction cost while preserving the ledger.
The registry's own liability posture also needs scrutiny. Public participant materials from NRS and LARUS emphasise the mismatch between catastrophic operational downside for holders and limited registry-side liability. Readers should treat those materials as interested arguments, but the underlying question is legitimate. If a registry can revoke or impair use of a block that supports a live network, yet the holder's practical remedy is small relative to renumbering costs, the institution has more power than accountability. That imbalance encourages legal escalation because parties see litigation as the only way to equalise risk.
This is not a call for registries to become insurers of every business model. It is a call for proportionality. A registry responsible for public coordination cannot guarantee every customer's revenue. It can, however, avoid remedies that impose continuity-scale harm without strong evidence, clear process, and independent review. It can publish audit results for record-integrity incidents where legally possible. It can separate emergency technical steps from punitive membership decisions. It can make transfer and leasing rules operationally intelligible rather than morally theatrical.
Such reforms would not eliminate conflict. IPv4 scarcity ensures that conflict will remain. They would change the form of conflict from a contest over discretionary gatekeeping to a contest over known rules. That is the difference between a registry that reduces scarcity cost and one that multiplies it.
Why AFRINIC matters beyond one registry
AFRINIC's crisis matters because it is not simply an African governance story. It is a test of the RIR model under late-stage IPv4 scarcity. Other regions exhausted earlier, developed transfer practices under different conditions, or entered scarcity with different institutional strength. AFRINIC combined a late residual pool, regional development claims, high market value, public record-integrity allegations, a major holder dispute, receivership, contested elections, and continuing litigation. That combination made visible weaknesses that can remain hidden elsewhere.
One weakness is the ambiguity of member power. RIRs often speak in the language of community, but the legal mechanics of membership, voting, board authority, and policy ratification depend on specific corporate documents and jurisdictions. When a registry is healthy, the ambiguity may not matter. When scarce resources make governance commercially valuable, it matters intensely. A community cannot govern a scarce resource credibly if the legal identity of the community is unclear.
Another weakness is the gap between technical neutrality and economic impact. A registry may think of itself as recording unique identifiers and applying community policy. Operators experience the registry as a continuity institution. Customers experience it indirectly through uptime, reputation, and address stability. Markets experience it through price and risk. Courts experience it through contracts and corporate law. These are not contradictory descriptions. They are different layers of the same institution. Scarcity forces all layers into view at once.
AFRINIC also matters for African digital development. It is tempting to say that Africa's future depends on preserving the last IPv4 addresses for local use. It is also tempting to say the remaining pool is too small to matter and the region should simply import addresses or move to IPv6. Both claims are partial. Africa needs IPv6 growth, but IPv4 remains operationally relevant for compatibility, customer reach, legacy systems, and commercial services. The remaining AFRINIC pool cannot finance or supply the continent's long-term growth by itself, but mismanaging it can still damage trust, delay networks, and make African operators pay more for continuity.
The issue is therefore institutional quality, not address nationalism. A credible AFRINIC would help African operators by keeping records reliable, applying scarcity rules predictably, processing legitimate requests efficiently, supporting IPv6 transition, and reducing the risk premium attached to resources in its region. An AFRINIC seen as a discretionary gatekeeper would do the opposite. It would make every allocation, transfer, lease, election, and bylaw change part of a larger uncertainty discount.
The broader internet governance system has already reacted. The Register's 2025 and 2026 reporting shows ICANN, the NRO, and other RIR communities considering how to handle a dysfunctional RIR, including revised lifecycle rules and emergency arrangements. That work may be necessary. It should not become an excuse to miss the economic lesson. Derecognition rules, emergency service provisions, and peer support address institutional collapse. They do not by themselves solve the underlying scarcity problem: how to administer economically valuable number resources without pretending that registry discretion is costless.
AFRINIC's recovery, if it happens, will be measured by mundane evidence: budgets, board minutes, executive appointment, audited finances, clear bylaw repair, public policy meetings, transfer processing, RPKI and WHOIS reliability, published resource statistics, court outcomes and member confidence. The drama will attract attention, but the recovery will be bureaucratic. That is appropriate. A registry proves legitimacy by making high-value records boring again.
Uncertainty and watchpoints for public readers
Several uncertainties should guide public monitoring of AFRINIC over the next year. The first is legal continuity. Cloud Innovation's dispute with AFRINIC, the winding-up effort reported in 2025 and 2026, ICANN's intervention in Mauritius, and court orders around communications and registry status remain central to the institution's risk profile. Public readers should watch actual court orders and party filings more than campaign language. A court decision that clarifies whether AFRINIC can be wound up, how number resources are treated in such a proceeding, or what limits apply to registry-side enforcement would affect the market's assessment of every AFRINIC-linked block.
The second uncertainty is governance normalisation. The September 2025 election and the 2026 budget-and-action-plan statements were positive signals, but they are not the same as stable governance. Watch whether AFRINIC seats durable executive leadership, publishes budgets and minutes, holds regular member and policy meetings, resolves bylaw tensions, and demonstrates that the board can act without immediate procedural paralysis. Legitimacy will be earned through repeated ordinary acts.
The third uncertainty is the remaining IPv4 pool. The reported 773,376 unallocated IPv4 addresses in early 2026 should be tracked through public statistics, allocation notices, and Phase 2 policy execution. The question is not only how fast the pool declines. It is whether allocations are processed consistently, whether applicants understand the evidence standard, whether any delays are explained, and whether residual scarcity becomes another arena for accusations of favouritism or obstruction.
The fourth uncertainty is transfer and leasing treatment. Public statements from AFRINIC, LARUS, Cloud Innovation, and NRS show a live dispute over whether leasing and monetisation are legitimate continuity mechanisms or improper commercialisation of registry-administered resources. Watch for formal policy text, court findings, registry guidance, and actual transfer outcomes. A stable rule could reduce market friction even if participants dislike it. Ambiguous communiques and ad hoc litigation will increase the gatekeeper premium.
The fifth uncertainty is record integrity. The 2019 allegations reported by KrebsOnSecurity remain important because no scarcity policy can work if members doubt the ledger. Public readers should look for transparent accounting of historical record-integrity incidents where legally possible, robust change logs, clear authorisation procedures, accurate member records, reliable WHOIS and RDAP data, and evidence that reverse DNS, routing-registry, and RPKI processes remain technically dependable.
The sixth uncertainty is the relationship between local law and regional function. AFRINIC is incorporated in Mauritius but serves a continental and global coordination role. The court system is therefore not an external detail. It is part of the registry's operating environment. Watch whether Mauritian corporate-law questions around membership, directors, receivership, and winding-up are resolved in ways that make resource-member participation more coherent rather than more fragile.
The final watchpoint is whether AFRINIC reduces or raises the cost of scarcity. A healthy registry will make IPv4 scarcity easier to price by making rights of use, transfer paths, dispute rules and ledger updates predictable. A weak registry will make scarcity harder to price by adding institutional risk to an already finite resource. The public question is not whether AFRINIC can make IPv4 abundant again. It cannot. The question is whether it can stop behaving, or being perceived, as a discretionary gatekeeper and return to the more valuable role: a trusted ledger for a scarce global resource.