The question hiding inside the address book
AFRINIC's conflicts are often described as fights about governance, elections, litigation, IPv4 exhaustion, regional development, or one large address holder. Each description captures part of the public record. None quite reaches the economic core. Beneath the procedural noise sits an older and more consequential question: what does it mean to hold title to an internet number resource when the registry says the holder does not own the number, the market prices recognised control as an asset, customers rely on that control as infrastructure, and some records were created before the modern regional registry system fully existed?
Title is used here in an institutional sense, not as a claim that IPv4 blocks are land, spectrum or freehold property. The practical question is simpler and harder. Who has the recognised claim to use, route, administer, transfer, secure, defend or rely on a block, and under what conditions can that claim be disturbed? In an age of abundant addresses, that question could hide behind registry jargon. In an age of scarcity, leasing, transfers, litigation and balance-sheet reliance, it decides value.
AFRINIC is an unusually sharp test because its formal role looks administrative while its decisions now have economic consequences. Its public materials describe a nonprofit, member-based organisation registered in Mauritius, serving Africa and parts of the Indian Ocean, and distributing IPv4, IPv6 and autonomous system numbers under community-developed policies. Its policy manual treats registration, uniqueness, aggregation and conservation as working goals of the IPv4 system. Its exhaustion page records entry into Soft Landing Phase 1 on 31 March 2017 and Phase 2 on 13 January 2020. Its transfer policy allows IPv4 resources to move from existing members or from legacy resource holders in the AFRINIC service region, but under conditions that include recipient need review and, for transferred legacy resources, loss of legacy status.
Those official facts describe the machinery. They do not settle the economics of old allocations. Legacy or early allocation records carry history: informal assignment practices from the research-network era, regional transitions, records inherited by later registries, companies that merged or disappeared, and uses that changed as the internet became commercial. A line in a registry database may therefore be both a technical record and a contested economic instrument. It can support a running network, a leasing contract, a transfer negotiation, a customer dependency, a route-security assertion and a lawsuit.
Lu Heng's public commentary has pressed the underlying contradiction in deliberately stark terms. An IP address is a unique numerical label, not a physical object; the internet needs a consistent record, not metaphysical ownership language. Yet operators build businesses around addresses, markets price them, and registry institutions retain power to alter records while usually accepting little direct economic downside. The institutional-economics point is clear: the registry record, operational reliance and economic title have drifted apart.
AFRINIC's legacy-allocation problem is where that drift becomes visible enough to fight over. A registry can correctly say that it records and coordinates number resources rather than selling land. A holder can correctly say that its recognised control has become the basis of contracts, customers and revenue. A buyer can correctly price transferability. A court can correctly refuse to treat number resources as the registry's own assets while still needing to decide whether resource holders have reliance interests against registry action. The danger begins when one layer is used to erase the others.
Three kinds of title
The narrowest layer is the registry record. It says that a particular block is associated with a holder, contact, status and related objects in a database. It helps prevent conflicting claims. It supports troubleshooting and operational coordination. AFRINIC's policy manual gives registration serious force: assignments, allocations and sub-allocations of public address space are to be registered, and registration data must be correct. That is a strong statement about record authority. It is not, by itself, a complete statement about title.
The second layer is operational reliance. A network does not use an IP address because the entry is elegant in a database. It uses the address because routers, customers, abuse desks, security systems, hosting platforms, banks, cloud vendors, reputation services and counterparties treat the address as part of a working service. Reverse DNS, WHOIS or RDAP information, route objects and RPKI materials do not themselves make packets flow, but they help a claimant show that its operation is recognised. When a customer buys hosting, connectivity or cloud infrastructure, it rarely reads a regional registry policy manual. It relies on the operator's ability to keep the network reachable.
The third layer is economic title. It appears when the operational position can be sold, leased, financed, pledged, valued, discounted or litigated. The Internet Governance Project's 2021 analysis of the AFRINIC crisis reported that IPv4 market prices had moved from roughly $8 per address in 2017 to roughly $30 by 2021, using a /16 block as an example of a resource worth about $2m. Prices vary by size, reputation, routing history and transaction risk. The institutional point does not depend on any one quote. Once a market pays for recognised use of a scarce identifier, the registry record becomes part of an asset package even if registry language resists the property label.
The confusion comes from treating these layers as identical. A registry can say, plausibly, that it keeps a record and does not issue deeds. A holder can say, plausibly, that it has built a business around a recognised resource. A market can price, plausibly, the holder's expected control over that resource. A court can protect the coordination function by refusing to let number resources be distributed as a registry corporation's assets, while also recognising that record changes can impose serious harm on those who depend on them. These propositions are not mutually exclusive. They are different parts of the same institutional arrangement.
AFRINIC's transfer policy shows the layers colliding. The source of a transfer must be the current rights holder recognised by AFRINIC and must not be involved in a dispute as to the status of the resources. That is registry-record language. The recipient must show need and become an AFRINIC member under current policies. That is policy-control language. If the resources are legacy resources, they cease to be regarded as legacy once transferred. That is title-conversion language. A single transfer therefore changes not only a database entry but also the regulatory character of the block and the economic expectations attached to it.
This is why title cannot be reduced to a slogan about ownership. The real question is whether the chain of recognition is clear enough that parties can transact without guessing how much value the registry might later remove. If a holder can never know whether a changed customer mix, a leasing arrangement, a merger, a geography question, a billing dispute or a policy reinterpretation will reopen the original allocation rationale, the market will discount the block. If a registry can never correct fraud, stale records or false claims because all reliance is treated as absolute ownership, the record will decay. A credible system needs both reliance and correction.
The economic frame is therefore not pro-holder in the crude sense. It is pro-institutional clarity. Markets do not need a registry to bless every commercial plan. They need to know which facts matter, which records are final, which defects can be cured, which breaches can lead to severe remedies, and which decisions are independently reviewable. Registries do not need to become commodity exchanges. They need enough discipline that their recordkeeping function does not become a hidden option over other people's operating capital.
Why legacy space is different
Legacy address space is not merely old address space. It is a reminder that the internet's institutional architecture arrived after many operational facts were already in place. AFRINIC was the last of the five Regional Internet Registries to be created, and its own exhaustion materials say it has managed a pool of number resources since 2005. By then, the internet was no longer a research club. But much of the global IPv4 endowment still reflected earlier allocation practices and regional asymmetries.
Old records therefore contain two histories. One is technical: who was entered as holder, which block was associated with which organisation, whether contacts were maintained, whether successors were recorded, whether reverse DNS or routing data were updated, and whether the block was visible in use. The other is economic: whether the holder treated the resource as a costless administrative allocation, a business input, a tradable asset, a leasing base, a dormant reserve or a strategic option. These histories rarely line up neatly. A database may show an old entity whose legal successor is hard to prove. A block may be unrouted but valuable. A business may have changed from local access to global hosting. A policy rule written for new allocations may be applied in an environment shaped by older grants.
AFRINIC's transfer policy recognises legacy status just enough to reveal the problem. It allows IPv4 resources to be transferred from a legacy resource holder in the AFRINIC service region. The policy therefore cannot pretend that legacy resources are irrelevant. But it also says transferred IPv4 legacy resources will no longer be regarded as legacy resources. The act of moving the block through the current registry process converts its status. A holder contemplating transfer must price not only buyer demand but also the loss of a historically different position. That is not an administrative footnote. It is an economic term.
The legacy overhang becomes explosive during institutional stress. If a registry is trusted, parties may tolerate historical imperfection because routine corrections are expected to be fair. If a registry is distrusted, every old record becomes a potential battlefield. A dormant or weakly documented block looks like an opportunity to a bad actor, a liability to the registry, a risk to buyers and a possible windfall to a successor. If the registry is both recordkeeper and judge of old facts, the value of controlling registry interpretation rises.
The water-company analogy in Lu Heng's public notes is polemical but useful when stripped down. A monopoly utility does not own the house merely because the household depends on its pipes. In registry terms, a recordkeeper does not become owner of the resource or the network merely because users depend on the recordkeeping system. But the same logic cuts both ways. A holder does not become immune from proof merely because it has relied on an old entry. The record must be strong enough to support reliance and auditable enough to correct defects.
Legacy title is hard precisely because it asks for continuity without amnesia. It cannot romanticise all early holders as innocent pioneers. Nor can it let the current registry rewrite history as if years of recognised use created no legitimate expectation. It needs a proof standard: what documents establish successor status, what operational history matters, what treatment applies to dormant blocks, how competing claims are frozen, and when a verified old record becomes final enough for transfer or financing.
AFRINIC inherited that problem in a region with a small global share of IPv4, late institutional formation and high development stakes. That makes the politics sharper, but it does not abolish the economics. A block's value still depends on recognised control, transferability, reputation, routing continuity and legal confidence. If legacy status is unclear, every one of those variables becomes unstable. The task is to define what proof is required, what reliance deserves protection and what conversion happens when old records enter current market channels.
The official grammar of non-ownership
The RIR system has long resisted the idea that internet number resources are ordinary property. There are good reasons for that resistance. IPv4 addresses and ASNs are globally unique identifiers. Duplicate assignment would break coordination. Routing is not guaranteed merely because a registry allocates a number. Conservation, aggregation and registration impose constraints that do not look like ordinary ownership. If every holder could treat a number as an unrestricted commodity, the registry's ability to prevent fraud, stale records and conflicting claims would weaken.
AFRINIC's policy manual contains this grammar in several places. It calls IPv4 addresses public or global identifiers, says uniqueness is an absolute goal, and requires registration to support troubleshooting. For ASNs, it says directly that assignment or registration does not confer ownership and describes users as custodians rather than owners. For IPv6, it says address space should not be considered freehold property and is licensed for use rather than owned. The IPv4 transfer section is less metaphysical and more operational, but it too treats transfers as policy-governed changes in recognised rights, not as simple conveyances of private property.
Contemporary reporting captures the same tension. The Register's March 2026 reporting described AFRINIC's rejection of asset-style framing by noting that IP addresses are not owned as traditional property, while also observing that addresses are often bought, sold and leased. The contradiction sits in that juxtaposition: not property in formal language, traded in practice. A position that refuses one word cannot make the market effects disappear. It can only decide how to regulate them.
Non-ownership language is most useful when it prevents category mistakes. Numbering resources are not AFRINIC's corporate assets. ICANN's reported intervention in Mauritian winding-up proceedings in 2026 sought to explain that numbering resources administered through AFRINIC should not be treated as assets available for distribution if the registry were wound up. That position protects the global coordination function. If a registry corporation's creditors or shareholders could divide number resources like office furniture, internet-number coordination would collapse into corporate insolvency law.
The same language becomes dangerous when used to deny reliance. A taxi licence, spectrum authorisation, banking charter, airport slot or registry record may not be ordinary property, but the holder's reliance can still be economically serious. Law and policy often protect notice, proportionality, appeal and legitimate expectations without turning every entitlement into freehold ownership. The public interest in correction does not require arbitrary revocation. The absence of property in the old sense does not mean the absence of title in the practical sense.
AFRINIC's challenge is to keep the useful part of non-ownership language and discard the lazy part. The useful part says the registry must preserve uniqueness, prevent duplication, correct fraud, keep data accurate and avoid treating number resources as its own balance-sheet assets. The lazy part says holders should accept indefinite uncertainty because they were never owners. That is not a stable bargain. It asks operators to invest, route, serve customers and pay fees under a record whose value can be reopened after reliance has matured.
The better grammar would distinguish property, record, entitlement and reliance. AFRINIC need not say that an IPv4 block is land. It should say what a recognised holder can rely on, what conditions attach to that reliance, what process is required before disturbance, what remedies are proportionate, and what happens to legacy status on transfer, merger, dispute or court order. Clarity would not end controversy, but it would move the fight from metaphysics to rules.
Scarcity turned records into economic claims
IPv4 scarcity did not create old records, but it made them valuable. AFRINIC's exhaustion page gives the chronology. After the IANA free pool reached a critically low level in 2011, each RIR received a final /8. By September 2015, four of the five RIRs had exhausted their free pools. AFRINIC entered Phase 1 in March 2017 and Phase 2 in January 2020. During that interval, the gap between administrative allocation fees and market value became too large to ignore.
The Internet Governance Project argued in 2021 that AFRINIC held a very small portion of global IPv4 while, for a period, being the only region with a large pool of previously unallocated IPv4 available for nominal administrative fees. That created an arbitrage problem. AFRINIC's fee schedule charges membership and allocation fees by category; those charges support registry services and are not market-clearing prices for a scarce asset. A /16 can attract annual fees far below the millions of dollars implied by transfer-market prices. A registry can be nonprofit and still control access to value it does not itself price.
Scarcity also changes the meaning of delay. In an abundant system, a rejected request or a slow correction may be an inconvenience. In a scarce system, rejection pushes a network toward the transfer market, leasing, renumbering, carrier-grade NAT or postponed deployment. A transfer hold can destroy a transaction. A status dispute can trigger customer migration or legal warranties. A need review can become a valuation event. The words in the policy manual may remain the same; the stakes around them change.
Market reliance is not limited to outright transfers. Leasing matters because it converts recognised use into recurring revenue without necessarily moving the registry record. A hosting company, cloud operator or access network may lease capacity for customers whose services depend on stable address reputation and continuity. The registry may view some leasing arrangements with suspicion, especially if they obscure control chains, regional use or abuse accountability. Yet the market's existence is not hypothetical. Reporting, participant materials and public commentary all describe a world in which IPv4 is leased, bought, sold or otherwise monetised.
That monetisation changes what title defects cost. A clerical error in an abundant, low-value resource system is annoying. A defect in a scarce, priced and leased block can impair a transaction, interrupt revenue, undermine financing or invite litigation. A buyer will discount a block if the source's status is unclear. A lessor will discount future revenue if the registry might reclassify the arrangement. A customer will discount a network if routing continuity or abuse reputation is unstable. Clean title is therefore not a lawyer's luxury. It is market infrastructure.
Lu Heng's public note on the present registry model puts the argument bluntly: once IPv4 became scarce, transferable, priced, financeable and operationally indispensable, the old low-value registry shell no longer matched the object it governed. In less loaded language, the registry record became a price-bearing institution. It did not merely describe who had numbers. It affected liquidity, collateral-like value, revenue stability and counterparty confidence. That does not prove that every market actor behaves well. It proves that enforcement and title rules must be designed for a market that already exists.
AFRINIC's more restrictive regional-transfer posture, as described in public reporting and participant commentary, sharpens the conflict between scarcity-era transferability and regional control. Supporters see regional limits as a way to prevent scarce African-issued resources from being drained into global inventory. Critics see them as lock-in that reduces value and concentrates registry power. The economics of title sit between those positions. If a resource cannot move, its value is not the same. If it can move without verification, the record is not safe. The real question is what conditions make mobility legitimate without turning every old allocation into a permanent hostage of current policy.
Scarcity also exposes a development paradox. A regional policy may seek to preserve capacity for African networks, which is a serious objective. But a low-liquidity and legally uncertain regime can make resources inside the region less financeable, less transferable and more politically risky. Operators who need capacity may then face higher costs, not lower ones. The region is not helped if protection turns into uncertainty. Development policy works best when it lowers transaction costs for genuine operators. It works badly when it makes the registry's discretion the most valuable scarcity input.
AFRINIC as a title-economics stress test
The Cloud Innovation dispute is often narrated as a confrontation between AFRINIC and a controversial member. That is factually true and analytically incomplete. The deeper issue is whether a registry can reopen the meaning of an allocation after a holder has built a business around it, and whether the holder's reliance limits the remedy even if a policy breach is alleged. That is title economics in its most practical form.
IGP's 2021 account reported that AFRINIC challenged Cloud Innovation over the use of a large IPv4 holding, including concerns about the countries of use, the original justification and the requirement that services originate in the AFRINIC service region. Cloud Innovation disputed AFRINIC's interpretation and argued, in substance, that business use evolves and that constant re-justification would make the registry a planner of network operations. Those are party positions and analysis, not final findings. They still show the issue that scarcity makes unavoidable: when does a registry review remain a record-integrity exercise, and when does it become control over a business model?
The distinction matters because remedies are not neutral. Asking for updated contact records is one thing. Holding a transfer is another. Threatening deregistration of live resources is more severe still. A holder that has customers, contracts and routing history built around a block will treat withdrawal as a continuity event. A registry that believes the original allocation bargain has been breached will treat inaction as a threat to policy integrity. Once both sides see the same registry entry as existential, litigation becomes predictable.
The case also highlights how regionality becomes a title term. A transfer or use rule tied to the AFRINIC service region may be understandable as conservation policy. But networks and customers do not always map neatly onto registry geography. A company can be incorporated in one place, serve users in another, announce routes from a third and contract with infrastructure operators elsewhere. A hosting platform can support African customers and international customers on the same architecture. A leased block can have an operator, a customer and an abuse contact in different jurisdictions. The more a registry tries to police these facts through broad discretion, the more it inspects business models rather than record integrity.
AFRINIC's institutional stress increased the price of that ambiguity. The NRO's September 2023 statement welcomed the appointment of an Official Receiver by the Bankruptcy Division of the Supreme Court of Mauritius, describing a mandate to preserve the status quo, oversee elections, facilitate a proper board and appoint a chief executive. The Register later reported election suspension, annulment, a subsequent board election in 2025, continuing critics, further legal exposure and ICANN intervention in 2026. Those facts do not decide the merits of any holder's claim. They show the governance environment in which title risk is priced.
The market does not wait for constitutional serenity. A buyer, lessor, lender or customer must decide whether a block's recognised status will survive board disputes, receiver decisions, policy reinterpretations, court orders and transfer holds. Even if packets continue to route, the administrative path may be priced as risky. That is why "mere governance" is not a side issue. When the institution controlling recognition is unstable, governance becomes part of title.
The dispute should not be reduced to a morality play. AFRINIC may have legitimate reasons to examine documentation, regional use, abuse accountability and whether a source is in dispute. Cloud Innovation and related market actors may have legitimate reasons to object to retrospective reinterpretation, severe remedies and regional lock-in. Interested participants on both sides use language that favours their commercial or institutional position. The article's point is narrower: a system built for administrative coordination now carries asset-like consequences, and the rules have not caught up.
That mismatch creates incentives for escalation. If the holder thinks the registry can destroy value without bearing comparable liability, it will litigate aggressively. If the registry thinks a member is converting regional allocations into private inventory, it will defend its policy role aggressively. If courts are asked to decide under time pressure, they may be forced to translate internet-number categories into ordinary corporate, contractual and property concepts. The cost of unclear title architecture is paid through litigation, delay, public distrust and valuation discounts.
Fraud, stale records and the cost of proof
Any serious title analysis must also face the risk of fraud. AFRINIC's record problem is not only that legitimate holders may suffer from uncertainty. It is also that old, stale or poorly controlled records can be abused. KrebsOnSecurity reported in 2019 on allegations that blocks associated with defunct or acquired organisations had found their way into the hands of marketing firms and that companies tied to a former AFRINIC policy coordinator appeared to have sold such blocks. The report cited an estimated market value above $50m for the documented addresses and quoted AFRINIC's then chief executive as saying the organisation was investigating. The article was not a final court judgment. It was a warning about how valuable historical ambiguity had become.
Dormant records invite both predation and overcorrection. If a company disappears, merges, changes name, loses documents or stops routing a block, the registry may lack a clean chain from original allocation to current claimant. A bad actor may exploit the gap. A legitimate successor may struggle to prove its position. A buyer may not know whom to trust. A registry under pressure may decide that aggressive review is safer than passivity. Each response has costs. The absence of clean provenance makes those costs unpredictable.
Legacy title therefore needs an audit vocabulary. Not every old record is fraudulent. Not every unrouted block is abandoned. Not every transfer from a legacy holder is suspicious. Not every suspicious pattern justifies immediate revocation. A credible registry would classify problems: verified holder, successor review, dormant contact, historical documentation gap, competing claim, fraud investigation, court hold, payment hold, transfer hold and record-correction required. Each category should carry a process and a consequence. The goal is not to expose sensitive data or punish history. It is to reduce the fog in which both fraud and arbitrary power thrive.
AFRINIC's own manual already points toward such discipline. Registration data must be correct. Transfers require the source to be recognised and not in dispute. Reverse DNS depends on appropriate registered assignments or sub-allocations. These rules imply a record with evidentiary standards. They become dangerous only if the standard is applied unpredictably, retroactively or selectively.
The market benefits from auditability. A buyer or lessee wants to know whether a block's history is clean enough to support the deal. A lender wants to know whether the registry can disturb the record on grounds that cannot be priced. A customer wants to know whether service continuity depends on a hidden dispute. A regulator or court wants to know whether the registry is correcting a defect or asserting a policy preference. Clean provenance lowers transaction costs because it lets each party price actual risk rather than rumour.
Lu Heng's public note arguing that internet number resources are not political property says the network that built, paid for, operated and bore the downside of a resource is the substantive asset holder, while the registry records and coordinates. That formulation is strongest when the holder's chain is clean. It is weaker when provenance is opaque. The answer is not to abandon the insight that operational reliance matters. It is to demand proof good enough that reliance and correction can coexist.
This is the institutional-economics case for a serious legacy review process. The review should be strict about identity, authority and successor claims. It should be careful about fraud. It should not treat age as guilt. It should not treat value as evidence of misuse. It should separate record defects from policy disagreement. It should give reasons, cure periods and review rights. Without such discipline, old allocation records become lottery tickets for opportunists and weapons for gatekeepers.
Lock-in and the discount on economic title
AFRINIC members do not join a philosophical seminar. They join because they need number resources, registry services, reverse DNS, routing-security support, WHOIS or RDAP records, transfer processing, invoices and a recognised relationship with the regional registry. AFRINIC's fee schedule makes that relationship concrete: members pay annual fees based on resource categories, allocation or assignment fees when resources are approved, late-payment penalties and other charges. The schedule also says transfer parties must be in good standing before transfers are considered. The member pays for a service, but the service conditions the economic status of the resource.
That creates expectations on both sides. A member expects accurate records, fair processing, confidentiality where appropriate, technical continuity, avoidance of duplicate claims and consistent application of rules. The registry expects the member to pay, keep records current, use resources according to policy and cooperate with review. Both expectations are legitimate. The problem appears when exit options are weak. If a registry is the only recognised operator for a resource's region and a specific block cannot move its administration away from institutional failure, the member bears a lock-in risk that ordinary service language understates.
The monopoly point can be stated without rhetorical excess. The absence of realistic exit increases the importance of procedural safeguards. If a customer can switch vendors, harsh terms may be disciplined by competition. If a resource holder cannot move the resource-level record without registry permission, the registry's review, billing, transfer and enforcement powers carry a monopoly premium. The market prices that premium as a discount on the block.
Lock-in affects legacy resources in a particular way. A legacy holder may have a historically different relationship to the registry than a modern member. Yet AFRINIC's transfer policy says transferred legacy resources lose legacy status. A holder that wants liquidity must therefore compare the buyer's price with the value lost through conversion. If inter-RIR or regional transfer restrictions limit buyers, the discount grows. If governance is contested, the discount grows again. A trapped asset is still valuable, but it is less valuable than one with predictable portability.
Supporters of regional restrictions argue that the discount is a legitimate cost of preserving resources for regional development. If resources administered through AFRINIC could flow freely to global buyers, late-developing African networks might lose access to scarce capacity. That concern should not be dismissed. IPv4 scarcity is real, and the original regional bargain matters. But the remedy must be judged by outcomes. A regional lock-in rule can preserve resources for development only if regional networks can obtain, use and finance addresses under clear rules. If it mainly creates uncertainty, litigation and low liquidity, it may protect a slogan while harming operators.
Member expectations therefore require a richer bargain. AFRINIC can require fees, records and policy compliance. Members can require finality, notice, appeal and clear portability conditions. The registry can preserve a regional development objective. Members can demand that the objective be translated into predictable rules rather than case-by-case moral judgment. A legacy holder can be required to prove a chain of entitlement. Once proved, it should not be left under indefinite suspicion merely because the record is old and valuable.
The same logic applies to leasing. A registry may reasonably worry that opaque leasing can hide control chains, weaken abuse response and bypass transfer policy. But a blanket suspicion of leasing can also drive activity into less transparent forms. If leasing is economically rational under scarcity, the registry should define the record, contact, abuse, sub-allocation and responsibility requirements that make it accountable. A visible leasing regime would be less dangerous than a shadow one.
The broader point is that liquidity disciplines governance. A resource that can move under clear conditions gives the holder bargaining power and gives the market a price signal. A resource that cannot move, or can move only after opaque approval, gives the registry more discretion and gives buyers more reasons to discount. Scarcity-era transferability is therefore not a side privilege. It is one of the mechanisms through which economic title becomes real enough to support investment.
Courts and the non-asset paradox
Courts entered the AFRINIC story because institutional rules were no longer enough to contain the conflict. The NRO's 2023 receiver statement treated receivership as a path toward functional governance and continued registry services. That is a factual exhibit about continuity. It does not answer title economics.
Receivership preserves an institution; it does not define every resource holder's reliance interest. A receiver can keep staff working, stage elections and protect corporate assets. But administrative continuity alone cannot decide whether an old allocation remains portable, whether changed use breaches need, whether a legacy transfer should erase status, or how far non-ownership language should limit remedies. Those questions require policy text, contracts, evidence and legal reasoning. If they remain unsettled, receivership becomes a holding pattern over a market that still needs title clarity.
The 2025 and 2026 reports from The Register show how quickly continuity and legitimacy interact. A June 2025 election was suspended and annulled after allegations about powers of attorney and voter documentation. A later 2025 election produced a board, but critics questioned arrangements and further legal challenges remained possible. In February 2026 AFRINIC signalled recovery, budget work and a strategy process. In March and May 2026, public reporting again described litigation, bylaw disputes, communications over leasing claims and ICANN intervention in winding-up proceedings. This is not ordinary administrative background. It is the environment in which title questions are being priced.
ICANN's reported winding-up position creates a paradox. Numbering resources administered through AFRINIC are not assets of AFRINIC available for distribution in liquidation. That is correct for registry continuity. But if the resources are not AFRINIC's assets, AFRINIC's power over the record requires especially clear limits. A registry cannot say, in one context, that the numbers are not its property and, in another, act as if institutional survival justifies open-ended control over the economic lives built on those numbers. The non-asset position protects the record from corporate liquidation. It should also discipline the registry's claim to economic control.
Courts are likely to remain important because they are being asked to translate internet coordination into ordinary legal categories. Judges need to know the difference between an address, a registry record, a Registration Service Agreement, a member vote, a transfer, a legacy status, a route object, a reverse-DNS delegation and an RPKI assertion. They also need to know which facts are contested and which are merely institutional slogans. A court that treats addresses as ordinary corporate property would harm coordination. A court that treats registry discretion as unreviewable because addresses are "not property" would harm reliance. The difficult middle path is the only useful one.
Title clarity can reduce the judicial burden. If AFRINIC publishes precise categories for resource status, transfer holds, legacy conversion, review triggers and remedies, courts can review decisions against a record. If the registry relies on broad claims of stewardship, regional destiny or institutional necessity, courts will be forced to invent categories during emergencies. That is bad for holders and bad for the registry. The more valuable old allocation records become, the less the system can afford ambiguity at the moment of dispute.
The legal lesson is not that courts should run the address system. They should not. The legal lesson is that the address system must become legible enough for courts to protect continuity without mistaking coordination for ownership or discretion for sovereignty. If the registry's own categories are imprecise, judges will supply categories from outside. Those categories may be poorly fitted to internet operations. A better title architecture would make the legal work narrower.
What a clearer title architecture needs
A better title architecture would begin by separating proof of record from approval of business model. The registry needs proof that the claimant is the recognised holder or legitimate successor, that the resource is not subject to a competing claim, that contacts are accurate, that abuse and operational accountability are reachable, and that any transfer conditions are satisfied. Those are record questions. The registry may also have policy questions about regional use, need and conservation. Those should be written with enough precision that they do not become a roving inquiry into whether staff or directors approve of a holder's commercial strategy.
Legacy resources need a conversion map. If a legacy holder signs a modern agreement, transfers a block, merges into another entity, moves the resource through an inter-RIR process or places it into a leasing structure, the consequences should be published in advance. Which rights are retained? Which old privileges end? Which current policies attach? Is the change reversible? What notices are required? What evidence proves successor status? What happens if a dispute arises after a transfer but before record update? These questions should not be answered only after millions of dollars are at stake.
Finality is the second requirement. A completed registration or transfer should not be forever vulnerable to reinterpretation. Reopeners are necessary for fraud, material misrepresentation, court orders, serious payment default, defined policy breach or competing claim discovered later. But finality should be the default after proper review. A market cannot function if every block carries an invisible option for the registry to retry the original allocation case whenever politics change.
Proportional remedy is the third requirement. Record defects should first invite correction. Payment problems should follow billing processes. Abuse-contact failures should require contact repair. Unclear customer geography should be handled under a defined rule, not as an invitation to total revocation. Fraud may justify recovery. Deliberate false claims may justify severe action. The remedy should match the breach and protect innocent downstream reliance wherever possible. Destructive remedies may sometimes be necessary, but they should be rare because they convert a policy dispute into a continuity event.
Independent review is the fourth requirement. The body that maintains the record should not be the final judge of every high-value title dispute in which its own interpretation, revenue, reputation or institutional power is implicated. AFRINIC's policy environment contains appeal concepts, but legacy title and transfer disputes need a commercially serious review path. Delay destroys value. So does secrecy. A specialised review mechanism, with reasoned outcomes and anonymised precedent where confidentiality requires it, would make disputes less likely to escalate immediately to courts.
Data publication is the fifth requirement. Aggregate statistics can show how many transfers are requested, approved, rejected, delayed, held for dispute, converted from legacy status, placed under court hold or returned for documentation. The registry need not reveal sensitive customers to publish a market-health dashboard. Without such data, each side fills the gap with anecdotes. With data, members can see whether policy protects the record or chokes it.
The sixth requirement is a disciplined leasing vocabulary. Leasing is neither automatically abuse nor automatically harmless. It can support efficient temporary use and customer continuity. It can also obscure responsibility and weaken abuse response. AFRINIC should state what registry records must show, who is responsible for abuse and contact accuracy, how sub-allocations or assignments should be recorded, how regional-use policy applies, and which arrangements create transfer-like obligations. The point is not to celebrate leasing. It is to make an existing market legible enough to supervise.
The seventh requirement is restraint in institutional storytelling. A registry's role is vital, but it is still a role. It maintains uniqueness, data integrity, registry services and fair process. It does not become the owner of every business built on the resources it records. Nor should resource holders convert reliance into immunity from proof, fraud review or policy compliance. Good title architecture is modest in both directions. It refuses confiscatory registry discretion and refuses absolutist asset claims.
None of this requires AFRINIC to become a commodity exchange. It requires the registry to recognise that its old allocation records now support a market and that title uncertainty is itself a policy outcome. A thin, accurate, auditable record with clear reliance rules would protect both regional development and market confidence better than broad discretion. The registry should be strong where it is supposed to be strong: uniqueness, provenance, data integrity, security services and fair process. It should be modest where its institutional incentives are weakest: deciding the commercial destiny of resources it does not own and cannot economically indemnify.
Uncertainty and watchpoints
Several uncertainties should guide public monitoring of AFRINIC's legacy-allocation title problem. The first is judicial treatment of the non-asset paradox. Watch whether Mauritian proceedings and any ICANN participation clarify that numbering resources are not AFRINIC assets available for winding-up while also recognising that resource holders may have reliance interests requiring due process. A judgment that protects only the registry shell will not settle title economics. A judgment that treats addresses as ordinary corporate assets will endanger coordination. The useful signal is careful separation of registry property, registry administration, holder entitlement and operational reliance.
The second watchpoint is legacy conversion. AFRINIC's policy says transferred IPv4 legacy resources will no longer be regarded as legacy. That clause deserves close attention in any future transfer, merger, restructuring or dispute. Does conversion occur automatically on record change? Does the holder receive clear notice of the economic consequences? Are there transitional protections? Does conversion affect fees, transferability, review exposure or contractual remedies? The answers will tell the market whether legacy title is a stable historical category or merely a waiting room for current policy.
The third watchpoint is the board and member register. A registry that controls economically significant records needs a trusted governance base. The 2025 election controversies, allegations over powers of attorney, later board formation and bylaw debates all matter because policy authority is economically valuable. Watch for verified membership categories, clean voting-authority rules, published board minutes, audited conflict policies and transparent treatment of resource members under Mauritian company law. If governance legitimacy remains cloudy, every title decision will carry a legitimacy discount.
The fourth watchpoint is transfer and leasing rules. AFRINIC can either publish precise rules for leasing, sub-allocation, customer geography, inter-RIR movement, good standing and resource review, or it can continue to fight those issues through communiques and litigation. The former would reduce title risk. The latter would preserve it. Particular attention should be paid to whether regional policy is applied prospectively and objectively or used retroactively to devalue older allocations.
The fifth watchpoint is provenance remediation. The 2019 Krebs reporting showed why old records cannot be trusted blindly. AFRINIC should be expected to maintain auditable categories for dormant records, successor claims, suspected fraud, historical documentation gaps, court holds and ordinary verified resources. Clean remediation would improve both registry legitimacy and market liquidity. Opaque remediation would invite accusations of selective enforcement.
The sixth watchpoint is the revised ICP-2 process. A lifecycle framework for RIR failure, assistance and possible derecognition may improve continuity, but it should not become a higher-level title gatekeeper. The global layer should protect the record and service continuity, not choose winners in regional economic disputes. The question is whether emergency authority is tied to objective failure conditions and data portability, or whether it gives central actors broad discretion over policy disagreements.
The final uncertainty is cultural. AFRINIC's title problem cannot be solved if every market claim is treated as theft and every registry concern is treated as confiscation. Old allocation records are neither sacred relics nor empty permissions. They are institutional artefacts carrying operational history, commercial reliance and public coordination risk. The registry must be able to correct false claims; holders must be able to rely on verified claims; markets must be able to price claims; courts must be able to distinguish claims from assets of the registry itself. Until those distinctions are made explicit, legacy title will remain an engine of conflict.
AFRINIC's crisis therefore points beyond AFRINIC. The internet-number system was built to keep unique identifiers coherent. It did not fully anticipate that old records would become capital-bearing instruments. The cause is scarcity, delayed IPv6 substitution, commercial adaptation and institutional inertia. The constructive response is not to pretend addresses are ordinary property, nor to pretend they are valueless permissions. It is to build a title architecture honest enough for both the record and the market: accurate data, clear provenance, protected reliance, defined conversion, proportionate remedies and reviewable discretion. Old allocation records became modern institutional conflict because the system priced them before the system explained them. The next phase depends on whether AFRINIC can explain them before more value is destroyed.