APNIC and the economics of transfer market architecture

The market behind the registry update

The most important moment in an IPv4 transfer is not the first conversation with a broker, the signed purchase agreement, the buyer's wire, the seller's board approval or the route announcement that follows. In the Asia-Pacific region the decisive moment is usually quieter. APNIC, or an APNIC-linked National Internet Registry, recognizes the change and updates the registry state.

That act is settlement. It is not settlement in the narrow sense of a central securities depository moving shares between accounts. Internet number resources are not ordinary property, land, spectrum licences or warehouse receipts. They are unique identifiers whose usefulness depends on common recognition, public records and operational coordination. Yet IPv4 scarcity has pushed those identifiers into a commercial environment where they are bought, leased, financed, escrowed, audited, disputed and folded into customer infrastructure. In that environment the registry update performs a market function much closer to settlement finality than to routine clerical maintenance.

A simple account of the transfer system says that IPv4 is scarce, unused resources may move from one holder to another, the recipient must satisfy applicable policy and APNIC records the result. That account is not wrong. It is too thin. The more important question is how APNIC's rules, National Internet Registry relationships, documentation requirements, inter-RIR compatibility, historical-resource treatment, broker practice, escrow timing, corporate-authority checks, RPKI state, reverse-DNS continuity and compliance pressures shape the market in which transfers occur.

The Asia-Pacific IPv4 transfer market is not one market. It is several overlapping markets joined by a recognition layer. A Japanese network working through JPNIC, an Indian network interacting with IRINN, a mainland Chinese holder dealing with CNNIC, a Vietnamese operator, an Australian cloud platform, a Singapore data-centre business, a Pacific access provider, a global buyer receiving addresses from another RIR region and a legacy holder with incomplete historical paperwork do not face the same transaction surface. They may all end at a registry record. They do not arrive there through identical legal, linguistic, institutional or operational paths.

That variety makes APNIC's role more important, not less. APNIC serves 56 economies across Asia and Oceania and identifies seven National Internet Registries in its region: APJII in Indonesia, CNNIC in China, IRINN in India, JPNIC in Japan, KISA in Korea, TWNIC in Taiwan and VNNIC in Vietnam. APNIC presents NIRs as separate non-profit entities incorporated where they operate, with their own memberships and local-language services, while adhering to regional policy. It also states that each economy may have only one NIR and that a moratorium on new NIR applications was made permanent in 2024. These are not decorative governance details. They are market-structure facts. They mean that the route from seller to buyer may pass through local institutions, regional rules and sometimes another regional registry before final recognition is achieved.

In a region this diverse, APNIC is best understood as a neutral settlement ledger. Its central task is not to make IPv4 cheap or expensive, not to choose preferred buyers, not to rescue an allocation-era theory of need after exhaustion and not to become a discretionary allocator by another name. Its task is to make legitimate transfers recognizable, final, secure and operationally continuous while preventing fraud, duplicate claims and abuse of residual rationing rules. If APNIC does that predictably, it lowers transaction costs. If it does it unpredictably, it becomes a checkpoint that raises them.

The difference is paid for in the market. A clean APNIC-recognized transfer carries title-like confidence even if the legal vocabulary avoids ownership. A transfer exposed to uncertain documentation, inconsistent NIR handling, unresolved corporate authority, unclear inter-RIR compatibility, routing-security disruption or open-ended need review carries a discount. The discount may not appear on APNIC's accounts. It appears in the price a seller accepts, the escrow conditions a buyer demands, the warranties a broker writes, the legal fees a small operator cannot absorb, the time a deal spends in limbo and the decision by some networks to lease capacity rather than purchase it.

IPv4 transfer policy is therefore not only policy. It is market microstructure. It decides how quickly information becomes reliable, how much risk is pushed onto the buyer, how sellers prove provenance, how brokers earn their spread, how escrow survives registry delay, how cross-border friction is priced, how small operators obtain liquidity and whether the official ledger reflects commercial reality or forces that reality into private workarounds.

The useful line is simple. APNIC should verify identity, authority, provenance and continuity. It should protect the ledger against false claims. It should maintain Whois, RDAP, reverse-DNS, routing-registry and RPKI coherence through the transfer. It should coordinate cleanly with NIRs and other RIRs. But it should not become a price regulator, moral allocator or capital controller after exhaustion. The market needs APNIC to settle recognized reality, not to decide whether the reality deserves to exist.

Recognition finality is the product

In any mature asset market, finality is a source of value. The seller knows when it has delivered. The buyer knows when it can rely on the acquisition. Lenders, auditors, customers and counterparties know which record to inspect. Disputes may still happen, but the baseline state is legible. Without finality, every transaction remains partly provisional. A provisional market is an expensive market.

APNIC's own transfer conditions make finality visible in administrative language. A transfer occurs when Internet number resources are moved from one legal entity to another. APNIC processes transfer requests under policy and updates the APNIC Whois Database to reflect the result. When a transfer is complete, the source entity no longer has rights to the transferred IP addresses or AS numbers and the resources are registered to the recipient. That is close to the language of settlement even if APNIC does not describe itself as a clearing house.

The market reads the record that way because the APNIC entry is the shared reference point. A buyer may have a signed contract before APNIC acts. A seller may have received funds into escrow. A network may have tested route announcements. Yet until registry recognition is achieved, several questions remain open. Will the recipient be accepted? Will the source be treated as the currently registered holder? Are the resources subject to the five-year restriction attached to addresses delegated from the 103/8 final pool? Does the recipient's 24-month use plan satisfy policy? Are there unpaid account issues? Do legal documents prove the relevant merger, acquisition or authority? Are route records, RPKI material and reverse DNS expected to change or disappear? Is an NIR involved? Is a counterpart RIR's policy compatible?

Each question is a settlement-risk category. The more predictable the answer, the more valuable the resource. The less predictable the answer, the more the market prices APNIC-region addresses not as pure capacity but as capacity plus registry execution risk.

This is why the word "title" must be handled carefully but not avoided analytically. Registries often prefer terms such as delegation, allocation, assignment, resource licence, registration or right of use. Those terms matter. They reflect the fact that number resources exist inside a coordination system rather than as ordinary chattels. The market, however, still needs title confidence. It needs to know that the seller is the legitimate holder for transfer purposes, that the buyer will become the recognized recipient, that the registry will not later treat the transaction as invalid because of an undisclosed condition, and that operational services tied to recognition will move coherently. Whether lawyers call that ownership, a licence, a registration interest or an exclusive right of use, the economic requirement is the same: the recognized claim must be transferable and reliable.

APNIC's policy contains both a confidence-building function and a friction-generating function. On the confidence side, the source must be the currently registered holder and not involved in a dispute over the resource status. The minimum transfer size is a /24. Transfers must be accurately reflected in the Whois Database. APNIC maintains a public log of transfers. It defines source and recipient conditions. It recognizes historical-resource transfers under a separate path. Those features reduce uncertainty.

On the friction side, IPv4 transfer recipients must demonstrate plans for use within 24 months. Existing IPv4 holders must also show past usage rate and evidence of compliance with APNIC policy for previous delegations. Inter-RIR transfers depend on counterpart RIR compatibility. Addresses delegated from the 103/8 free pool cannot be transferred for at least five years after original delegation. APNIC also states that transfers outside policy are not recognized and that organizations holding such transfers must return the resources to the appropriate Internet Registry. These rules may have legitimate purposes. They are also market design.

The policy question is not whether settlement should have conditions. Every trustworthy settlement system has conditions. The question is whether the conditions protect finality or undermine it. Fraud checks protect finality. Clear proof of corporate authority protects finality. Public transfer logs protect finality. RPKI and reverse-DNS continuity protect finality. Broad, discretionary, slow or inconsistently explained review weakens finality by making the transaction depend on institutional judgment that parties cannot price in advance.

Finality also has a time dimension. A transfer completed after three days is not the same economic product as a transfer completed after three months, even if the final record is identical. The difference is carrying cost, financing cost, opportunity cost and counterparty uncertainty. APNIC therefore should not measure the transfer function only by whether it eventually records legitimate transactions. It should measure how much uncertainty the settlement layer inserts before recognition.

The useful market metric is not simply the number of transfers. It is the cost of reaching finality. That cost includes request preparation, documentary evidence, account status, NIR coordination, legal translation, use-plan submission, staff questions, waiting time, escrow extensions, RPKI transition work, reverse-DNS changes, public-log timing and post-transfer corrections. A low-cost settlement architecture makes these items legible and bounded. A high-cost architecture leaves them case-specific and opaque.

The best way to understand APNIC's market role is therefore not to ask whether it approves or disapproves of trading. The trade already exists before the registry action: in the negotiation, the escrow instruction, the price, the warranties and the buyer's planning. APNIC decides whether that private bargain becomes recognized public state. Recognition finality is the product it supplies to the market.

Provenance becomes price

In an exhausted IPv4 market, provenance is not a footnote. It is price. A buyer does not merely ask how many addresses are in the block. It asks where the block came from, how it has been used, whether the source is the recognized holder, whether the registration history is coherent, whether a prior allocation was subject to restrictions, whether there are disputes, whether abuse reputation is manageable, whether associated route and reverse-DNS records can move cleanly, whether the block was historical, transferred, allocated, assigned, leased or dormant, and whether the seller can prove authority to act.

APNIC's policy itself creates categories of provenance. It distinguishes current delegated resources, historical resources, unused or excess IPv4, AS numbers and resources transferred through merger, acquisition or reorganization. It also distinguishes ordinary APNIC-region transfers from inter-RIR transfers. The 103/8 free-pool rule creates another category: addresses delegated from that pool cannot be transferred for at least five years after original delegation and should be returned if the original reason no longer applies during that period. That rule is a rationing-protection device. It is also a liquidity attribute.

Market participants will price those attributes even if the registry does not. A block free of lock-up, with clear registered-holder status and simple account history, will be easier to sell. A block near the boundary of a waiting period, tied to a merger, claimed through historical documentation or subject to ambiguous local records will require more diligence. A block whose history includes spam, hijack allegations, shell-company movement or inconsistent routing will carry reputational and operational risk. A block that APNIC can process only after substantial evidence will be less liquid than one that requires routine confirmation.

Historical resources show the point well. APNIC's transfer material says historical resources may be transferred and that APNIC will register transferred resources in the Whois Database. Its policy says APNIC does not require technical review or approval of the resource's current use to approve a historical transfer and does not review the agreements between the parties or control the type of agreement. That is a ledger-like posture. It focuses on recognition rather than commercial terms. For a settlement layer, the instinct is sound.

Yet historical transfers also show why provenance cannot be ignored. If resources are not currently held under an APNIC account, the recipient must verify legitimate holdership. That verification may be straightforward if the paper trail is clean. It may be difficult if the holder has changed names, merged, dissolved, reorganized, moved jurisdictions or lost old records. In Asia-Pacific economies with different company registries, languages and corporate-law conventions, the same APNIC requirement can impose very different practical burdens. The registry sees a legitimate evidence requirement. The market sees a provenance discount.

Allocated resources create a different issue. When a current APNIC account holder transfers unused or excess IPv4, the source must be the registered holder and must not be involved in a dispute. The recipient must show a detailed plan for use within 24 months. Existing holders must add past usage rate and compliance evidence. These conditions link provenance to eligibility. The cleaner the seller's history, the easier the source side. The clearer the buyer's operational plan, the easier the recipient side. Any uncertainty widens the bid-ask spread.

Corporate transactions add another layer. APNIC's merger, acquisition or reorganization material says that when business structure changes, the organization taking over needs an APNIC account and the resources must be transferred. The source account must initiate the request in MyAPNIC, provide relevant legal documents such as a sale or transfer agreement, and provide documents issued by governing authorities from the economies where the entities reside. The account must be open; if a renewal invoice has been issued, it must be paid before the transfer request can be submitted. These details matter because they make corporate authority part of settlement.

In an ideal case, corporate authority is obvious. In the real market, it can be messy. A holding company may sell a subsidiary. A network may be acquired through an asset sale rather than a share sale. A founder may still control old maintainer access. A dissolved company may hold registry records that a successor claims. A court may appoint an administrator. A state-owned telecom may need ministerial approval. A company in one jurisdiction may merge into another with different document names. A private agreement may be commercially complete while the registry still needs public or official evidence. Every ambiguity becomes a settlement-risk premium.

APNIC should not ignore that risk. A registry that accepts forged authority destroys trust. But it should not let authority verification become an unlimited discretionary process. The market needs a finite evidence map: which documents prove authority, which documents prove succession, which documents prove name change, what translations are acceptable, when notarization or apostille is needed, how NIR documents are recognized, what happens when a court or insolvency office is involved, and what interim state preserves operational continuity while proof is completed.

Provenance is not only about preventing fraud. It is about making good resources cheaper to trade and suspect resources visibly suspect. A well-designed ledger does both. It raises the value of clean history by making it easy to prove, and it lowers the value of unclear history by forcing uncertainty into explicit evidence rather than rumor. That is how settlement architecture becomes price architecture.

One region, several transaction paths

The phrase "APNIC transfer market" hides as much as it reveals. The region includes economies with deep capital markets and economies where small network operators still face thin administrative capacity. It includes large incumbent telecoms, hyperscale platforms, national operators, data-centre groups, hosting providers, access ISPs, universities, public-sector networks, internet exchanges, regional carriers, security firms and small businesses that need a few hundred addresses to keep customers reachable. Some are direct APNIC account holders. Some work through NIRs. Some interact with brokers abroad. Some are recipients from other RIR regions. Some hold historical resources that predate the current registry style.

This variety creates segmentation. Segmentation is not the same as fragmentation, but it has similar economic effects. A block whose provenance is straightforward, whose holder is a direct APNIC account holder, whose documents are in an expected format, whose use history is clean and whose buyer already holds APNIC resources will trade differently from a block that sits behind a corporate restructuring, an NIR relationship, historical evidence, translated documents and a buyer entering APNIC for the first time. Both may be legitimate. They are not equally liquid.

NIRs are especially important because they localize service while preserving regional policy. That is a sensible institutional compromise. It reduces language and cultural distance in large economies. It lets hostmasters work with members in local conditions. It also creates a layer of practical market difference. Transfer scenarios recognized by APNIC include transfers between two APNIC accounts, between APNIC and an NIR, between two NIRs, between an NIR and another RIR, and between APNIC and another RIR. A market participant is therefore not merely asking whether an IPv4 block can be transferred. It is asking which institutional path the transfer must follow.

The settlement architecture must make those paths economically comparable. It does not have to make them identical. Local service will differ. Legal documents will differ. Language will differ. Fee schedules may differ. But the settlement meaning should be clear across paths: who is recognized, what must be proved, what happens to associated records, what timelines are normal, what can delay the request, how disputes are isolated and when the recipient can rely on finality.

If that comparability is weak, segmentation becomes price discrimination by friction. Resources from a smoother path earn a liquidity premium. Resources from a more uncertain path trade at a discount or require stronger warranties. Brokers with experience in the difficult path gain power. Small sellers may lose value because they cannot package provenance to the market. Small buyers may pay more because they cannot explain their 24-month plan, obtain early comfort or manage registry exchanges without help.

This may not be APNIC's intention. It is the predictable result of a market where official recognition is scarce and process knowledge is unevenly distributed. Microstructure always rewards those who understand the settlement venue. In equities that means market makers, clearing brokers, custodians and high-frequency firms. In IPv4 it means transfer brokers, specialist counsel, escrow providers, registry consultants, NIR-aware intermediaries and large operators with repeat experience. The smaller and less frequent the participant, the more it pays to navigate the venue.

The region's geography intensifies the effect. A cross-border buyer may need to understand local corporate authority, account standing, government-issued documents, translation expectations, sanctions screening by banks or escrow providers, tax treatment, export or telecom-licence concerns and the operational history of the prefix. A domestic NIR transfer may involve local-language support but also local institutional practice that an international buyer does not understand. An inter-RIR transfer may require matching the APNIC side with an ARIN, RIPE NCC, LACNIC or AFRINIC policy environment. The resource is numerical and global. The transaction is legal and institutional.

APNIC cannot remove all these differences. It can reduce the uncertainty they create. A registry that publishes precise checklists, standard evidence categories, common defects, normal timelines by path and aggregate delay statistics turns segmentation into manageable complexity. A registry that leaves participants to learn by experience turns segmentation into information rents for intermediaries.

Brokers perform real work: supply search, demand screening, policy interpretation, document coordination, escrow timing and provenance warnings. The problem begins when registry architecture makes broker dependence unavoidable even for ordinary transfers. A good settlement system lets brokers add value. A poor one lets them sell access to procedural knowledge that should have been public.

Path-dependence also matters. Once an economy has an NIR, resources and member relationships develop around that institution. The permanent moratorium on new NIR applications freezes that institutional map rather than making NIR status available to other economies that might prefer local service. The result is not only a governance choice. It is a transfer-market choice. Some participants approach APNIC through long-established local infrastructure; others do not. The market will price the difference if APNIC does not make cross-path settlement meaning transparent.

Eligibility rules and the old allocator reflex

APNIC's recipient eligibility rules are usually described as policy compliance. In market terms they are microstructure. They decide who can buy, how quickly they can buy, how much documentation they must submit and whether capital alone is sufficient to obtain recognized control. That has direct effects on price, liquidity and participant composition.

The central rule is the 24-month use plan. Recipients that do not already hold IPv4 resources must demonstrate a detailed plan for the use of the transferred resource within 24 months. Recipients that already hold IPv4 resources must demonstrate that plan, show past usage rate and provide evidence of compliance with APNIC policy concerning past delegations. APNIC's transfer conditions say recipient APNIC accounts will be asked to provide a detailed plan for use of transferred resources. That language is clear enough as policy. Economically, it means APNIC retains an allocation-era need test inside a post-exhaustion transfer market.

There is a defensible argument for some needs-based guardrail. IPv4 is scarce, and APNIC's policy goal is to move addresses from those who no longer need them to those who do. Without any recipient scrutiny, speculative or fraudulent demand could absorb scarce supply, and final-pool rules might be evaded. But there is also a serious cost. A use-plan requirement turns a market purchase into an administrative forecast. The buyer is not merely proving identity, authority and non-dispute status. It is asking the settlement layer to accept a future business plan.

This changes the market's shape. Large buyers can document use. They have network plans, customer forecasts, utilization history, compliance teams and lawyers. They may already hold IPv4 and can show past usage rates. Small operators may have genuine demand but weaker documentation. New entrants may need addresses precisely because they are building capacity, not because they can prove historic use. A cloud or data-centre business may face fluctuating demand. A leasing platform may hold resources for downstream customers whose exact use varies. A security provider may need addresses for rotating infrastructure. A use plan may be operationally real while administratively hard to express.

The incidence of the rule is therefore unequal. It does not simply prevent bad transfers. It raises the cost of transfers for participants whose business model, size or geography is less legible to APNIC. A rule may apply to all recipients equally, but its cost falls more heavily on the less administratively sophisticated. The result can be a liquidity discount for sellers whose likely buyers are small or novel, and a liquidity premium for buyers that already look like APNIC's expected applicant.

The five-year 103/8 restriction has a different microstructure effect. It creates a temporary non-transferable class for addresses delegated from the final pool. That may protect the integrity of residual allocation policy by preventing immediate conversion of rationed space into cash. Yet it also creates inventory segmentation. A block inside the restriction is less liquid than a block outside it. The market must know the original delegation date and whether the restriction applies in mergers, acquisitions or reorganizations. APNIC's condition says it does. That matters for corporate buyers because acquiring a company with 103/8 space does not eliminate the restriction.

Waiting periods and eligibility rules can therefore produce hidden discounts. A seller may insist that a block is equivalent to another /20. A buyer will ask whether APNIC can complete recognition now, whether a use plan will be accepted, whether associated history creates questions and whether a final-pool restriction applies. If the answer is uncertain, the buyer's price falls. If the seller needs liquidity, the discount becomes real.

The risk is that APNIC becomes a capital controller without admitting it. A capital control does not require the registry to set prices. It exists whenever movement of a scarce capital-like input depends on discretionary official permission beyond narrow settlement integrity. If a buyer can pay, a seller can deliver, authority is clean, fraud is absent, operational continuity can be maintained and the registry still withholds recognition because it dislikes the buyer's forecast or business model, the registry is no longer only protecting the ledger. It is allocating after exhaustion.

The distinction is not anti-policy. It is pro-clarity. Some rules are ledger safeguards. They verify source, recipient, authority, non-dispute status, provenance and operational transition. Some rules are market controls. They restrict who may receive resources, when resources may move and what future use is acceptable. Both may be justified, but they should not be confused. APNIC's legitimacy would be stronger if every transfer rule were labelled by function: anti-fraud, provenance, residual-pool protection, inter-RIR compatibility, operational continuity or recipient market control.

That labelling would not settle every argument. It would make the argument honest. The market can tolerate cost better when it knows what the cost is buying. It can also distinguish a rule that preserves the integrity of the registry from one that continues the old allocator reflex after the free pool has lost its central economic role.

Escrow, brokers and the price of uncertainty

IPv4 brokers are sometimes discussed as if they sit outside the registry system. In practice they are one of its market responses. They exist because finding supply is difficult, but also because settlement is not frictionless. The broker does not only match buyer and seller. The broker packages provenance, interprets policy, stages paperwork, coordinates escrow, warns about waiting periods, manages expectations around APNIC or NIR processing and sometimes absorbs reputational risk by refusing suspicious blocks.

Escrow performs a similar function. In an ordinary sale of a simple good, money and title can move close together. In an IPv4 transfer, payment, registry recognition and operational usability may not be simultaneous. The seller wants assurance that the buyer can pay. The buyer wants assurance that APNIC will recognize the transfer. The broker wants assurance that its commission survives delay. The escrow provider wants conditions that are observable and enforceable. The resource may not be fully delivered until APNIC updates the record, associated operational records are handled and the buyer can use the block without unacceptable uncertainty.

Registry finality therefore enters private contract design. Parties may define release conditions around APNIC approval, Whois update, public transfer log entry, confirmation emails, route-object changes, reverse-DNS readiness, RPKI arrangements or absence of registry objection. If APNIC processing is predictable, escrow can be simple. If processing is unpredictable, escrow becomes longer, more conditional and more expensive.

This is where the incidence of registry cost becomes visible. A seller with clean, APNIC-direct resources can demand faster escrow release and fewer warranties. A seller with NIR complexity, historical documents or inter-RIR movement may need to accept delayed release. A buyer entering APNIC for the first time may face membership-fee or account-creation timing. A seller transferring out of APNIC may need to understand that associated APNIC Whois records, such as sub-assignments, route records and domain records, can be deleted from the APNIC database when the outbound inter-RIR transfer completes. If the buyer expected continuity of those records, private agreement must address the transition.

Brokers price this complexity. Sometimes they do so explicitly through fees. More often they do it through spread, deal selection and advice. A broker may steer a buyer away from a cheaper block because the provenance looks hard. It may advise a seller to fix account or documentation issues before listing. It may ask for stronger representations about lack of disputes, sanctions exposure, abuse history, corporate authority and compliance. It may prefer counterparties that have completed APNIC transfers before. Each practice is rational. Each is also evidence that the official settlement layer creates private market structure around itself.

This dependence is healthy only up to a point. Brokers should be paid for search, diligence and execution. They should not be paid merely because APNIC's public process is too hard for ordinary members to understand. A well-run settlement layer reduces avoidable broker rents while preserving useful broker services. It lets a small operator complete a straightforward transfer without needing an expensive intermediary. It lets complex transactions use brokers for complexity, not for access to practical knowledge that should be public.

There is also an abuse dimension. Brokers and escrow providers are natural control points for fraud prevention. They can identify fake sellers, suspicious authority documents, hijacked resources, shell buyers, money-laundering concerns and sanctions exposure. APNIC should welcome that market discipline. But it should not outsource policy clarity to private intermediaries. If brokers are the only people who know how a transfer will likely be treated, the market has become too dependent on informal settlement intelligence.

The strongest architecture uses brokers as optional specialists and APNIC as a predictable final-recognition layer. The weakest uses brokers as translators of opaque registry discretion. The difference matters most for small operators, which cannot pay large spreads or legal bills simply to acquire the minimum capacity needed for business continuity.

Escrow design also reveals the boundary between commercial freedom and registry responsibility. APNIC does not need to review the price, the spread, the financing cost or the allocation of risk between buyer and seller. It does need to make clear what event counts as registry completion, what associated records will change, what defects can stop recognition, and how a party can cure them. The private contract can then price the remaining risk. Without that clarity, escrow becomes a crude insurance policy against institutional uncertainty.

Inter-RIR compatibility is bridge architecture

Inter-RIR transfers expose the settlement nature of APNIC most clearly. A domestic transfer is one registry updating its own record. An inter-RIR transfer requires two registry systems to recognize the same change without leaving a gap, conflict or double claim. That is bridge architecture between ledgers.

APNIC's policy says it will process and record inter-RIR IPv4 transfers only when the counterpart RIR has an inter-RIR transfer policy that permits transfers between APNIC and that region. The space must meet a minimum /24 size. The address space should be under the management of the RIR where the source holds an account. The authentic holder should match the source without disputes. The source and recipient conditions are defined by the RIR where each party holds an account. If the recipient is in APNIC, APNIC's recipient conditions apply. If the recipient is elsewhere, the other RIR's recipient rules apply.

This is elegant in principle and costly in practice. Each RIR has its own history, contracts, policies, eligibility rules, legacy-resource treatment, review style, sanctions exposure, dispute posture and terminology. Inter-RIR compatibility does not mean identical rules. It means that the two systems can produce a mutually recognized settlement. For the market, compatibility is binary at the legal threshold but continuous in transaction cost. Two RIRs may formally permit transfers while still imposing enough documentation, timing and recipient-review friction to make transactions difficult.

APNIC's region adds another layer because NIRs can sit in the path. The official transfer scenarios include NIR inter-RIR transfers. That means the bridge may connect an NIR-linked environment to another RIR. The buyer or seller may be dealing not only with APNIC and a counterpart RIR but also with local NIR processes and local documents. A global buyer may look at the same prefix and ask whether it is APNIC-direct, NIR-administered, historical, previously transferred, final-pool restricted or tied to local corporate formalities. Those questions affect price.

Inter-RIR movement also raises geopolitical and compliance concerns. Banks, escrow providers, brokers and counterparties may screen parties for sanctions, beneficial ownership, money-laundering risk, export restrictions or reputational exposure. APNIC may also have obligations under applicable law. Those obligations should be handled as legal constraints, not as discretionary geopolitical preference. The registry should verify identity, authority and lawful ability to complete the recognition event. It should not use vague risk language to decide which jurisdictions, business models or market participants deserve scarce addresses.

The distinction is subtle but important. Fraud control protects settlement. Sanctions compliance respects external law. Capital control decides whether private parties may move a scarce resource because the registry dislikes the destination, price or commercial rationale. The first two are legitimate constraints. The third is a different power and should require a much higher justification.

Inter-RIR transfers also reveal why APNIC should publish compatibility and delay data. Market participants need to know not only whether a route is formally possible, but how it behaves in practice. Which counterpart RIR paths are common? Which are rare? What are typical processing times? What kinds of documentation failures occur? How often do requests stall because recipient rules differ? How are NIR cases handled? How often are outbound-associated records deleted and what operational warnings are given? Private transaction details can remain confidential. Aggregate settlement performance should not.

Without that information, the market constructs its own risk tables through broker memory. That is inefficient and unequal. Large repeat buyers learn. Small operators guess. A neutral registry should reduce that asymmetry by publishing the evidence needed to price settlement risk.

The word "compatibility" should therefore be read economically, not only formally. A bridge is not good because it exists on a map. It is good because traffic can use it at predictable cost. APNIC's inter-RIR design should be judged by that standard. It should not be enough that transfer policy can, in principle, connect to another RIR. The relevant question is whether legitimate transfers can cross the bridge without avoidable uncertainty, operational breakage or institutional delay that only repeat specialists can price.

Operational continuity is part of delivery

An IPv4 transfer is not complete in any useful economic sense if the buyer receives a registry update but loses operational continuity. The APNIC record matters because it connects to public registration, routing information, reverse DNS and RPKI. These are not extras. They are the operational surfaces through which recognized control becomes usable.

APNIC's Whois material describes the Whois Database as publicly searchable and as storing information about IP address ranges, routing policies, reverse-DNS delegations and network contact information. APNIC's routing-registry material describes the Internet Routing Registry as a distributed routing-information database used by operators for debugging, router configuration, route filtering and planning. Its reverse-DNS material explains that reverse DNS translates IP addresses to hostnames, that IANA delegates reverse zones corresponding to APNIC allocations, and that APNIC delegates authority for reverse zones to name servers provided through domain records. Reverse zones generated from APNIC Whois are refreshed on APNIC name servers on a regular cycle.

These factual details have a market consequence. When a block moves, the buyer needs not just recognition but operational handover. Contacts must be current. Abuse handling must point to the right party. Route records may need to be recreated or moved. Reverse DNS must not break unexpectedly. RPKI and route-origin authorizations must align with the new routing plan. Customers should not experience mail reputation failures, security-filter problems, route invalids or DNS confusion because the official settlement layer updated one part of the state while another lagged.

RPKI raises the stakes. Resource certification ties cryptographic assertions to recognized number resources and lets relying parties validate route-origin authorization. If a transfer changes recognized control without a clean RPKI transition, the result may be invalid routes, stale authorizations or a gap in security posture. Conversely, if RPKI authority can be disrupted for reasons unrelated to resource validity, registry power becomes operational leverage. That is dangerous. RPKI should protect routing integrity, not become a pressure instrument in commercial or administrative disputes.

Reverse DNS is less glamorous but often just as practical. Mail systems, security logs, reputation tools, customer applications and troubleshooting processes may depend on it. APNIC's own reverse-DNS material makes clear that delegations are generated from registered domain records and then propagated. During a transfer, the timing and deletion or recreation of such records can matter. A buyer using a newly acquired block for mail or customer services may care more about reverse-DNS continuity than about abstract policy language.

Outbound inter-RIR transfers illustrate the risk. APNIC's transfer conditions state that when resources are transferred from an APNIC account holder to a member of another RIR, associated records in the APNIC Whois Database, including sub-assignments, route records and domain records, if any, will be deleted. That may be technically necessary because the resource leaves APNIC management. It is also operationally consequential. A serious settlement architecture should make the consequences plain before closing, provide transition guidance and coordinate timing so that the recipient can recreate necessary state in the destination system.

The principle should be delivery-versus-title. Registry title-like recognition is necessary but not sufficient. Delivery includes the records and security state that make the addresses usable. APNIC does not need to operate the buyer's network. It should not guarantee global routability; no registry can. But it should treat registry-adjacent operational surfaces as part of the transfer package. A transfer process that surprises parties with deletion, certificate invalidity or reverse-DNS gaps creates avoidable market cost.

This is another place where aggregate metrics would help. APNIC could report transfer-related support categories without exposing private deals: reverse-DNS transition issues, RPKI or ROA support tickets, route-record corrections, abuse-contact updates, NIR coordination cases, inter-RIR outbound record deletion notices and post-transfer corrections. Such reporting would discipline the settlement layer. It would show whether operational delivery is improving or whether hidden friction is being pushed onto operators after recognition.

The market wants finality, but it also wants continuity. A registry that gives finality without continuity may satisfy its database. It does not satisfy the economy using the database.

Leasing and the liquidity signal

IPv4 leasing is often treated as a policy problem before it is understood as an economic response. It exists because transfer markets are costly, because buyers may need capacity temporarily, because small operators cannot afford outright purchase, because sellers may prefer yield over sale, because users may want address continuity without registry-facing exposure, and because a world of slow or permissioned transfers creates demand for flexible use.

Leasing is not automatically virtuous. It can obscure responsibility if the registry record shows one holder while another party operates the addresses. It can create abuse risk if contact data is wrong. It can complicate RPKI if authorization is unclear. It can expose downstream users if the lessor loses control, fails to renew, violates policy or becomes subject to dispute. It can create a broker chain in which no one clearly owns operational accountability. These are real risks.

But treating leasing primarily as evasion misunderstands market causality. Leasing grows when permanent transfer is expensive, slow, uncertain, capital-intensive or administratively hard. The more APNIC's transfer architecture behaves like a predictable settlement layer, the more leasing competes on ordinary commercial grounds. The more the architecture behaves like a discretionary checkpoint, the more leasing becomes a way to route around recognition friction.

For a small operator, leasing may be rational even if ownership-like control would be preferable in theory. Buying a block requires capital, due diligence, escrow, APNIC account status, usage-plan documentation, possible transfer fees, post-transfer management and continuing registry exposure. Leasing converts some of that into operating expense and places registry-facing responsibility upstream with a specialist holder. That may be more expensive per address over time, but cheaper in administrative risk. In institutional-economics language, leasing substitutes contract with a lessor for direct participation in the registry settlement venue.

APNIC should not try to abolish this economic response by moralizing it. It should make the ledger accurate enough that leasing does not require opacity. The registry record should support clear contacts, routing authorization, abuse pathways and operational delegation where policy permits. If the registered holder remains responsible, that responsibility should be legible. If the operational user handles abuse, that should be reachable. If RPKI is delegated or managed through the holder, the authorization chain should be understandable. If a dispute occurs, the registry should preserve the last verified state where possible while isolating the dispute.

The wrong response is to treat leasing as inherently suspicious and force it into informal arrangements. That makes the public record less useful. Commercial reality will not disappear because registry language dislikes it. It will become harder to see. The right response is to distinguish between ledger questions and commercial questions. Who is the recognized holder? Who is authorized to route? Who handles abuse? Who controls reverse DNS? What is the RPKI state? Is there a conflict claim? Those are registry-relevant. What price the parties agreed, whether the user should have bought instead, or whether the lessor's business model is philosophically attractive are not registry settlement questions.

Leasing also disciplines transfer architecture. If many operators prefer leasing because transfer recognition is unpredictable, that is a signal about APNIC's settlement cost. If leasing grows because operators want flexibility, that is normal market adaptation. If leasing grows because small buyers cannot pass use-plan review or cannot afford process risk, that is a policy-incidence problem. APNIC should not respond by suppressing the signal. It should ask why direct transfer is not meeting demand.

The broader point is liquidity. A market can be formally open and practically narrow. If only large, repeat, administratively fluent buyers can purchase cleanly, others will seek substitutes. They may lease, use more NAT, delay deployments, buy through intermediaries or accept operational arrangements that are less transparent than direct holding. The registry may see fewer transfer requests and conclude that the system is stable. The market may see suppressed demand.

For APNIC, leasing should be read as evidence, not simply as a target. Some leasing is ordinary commercial preference. Some leasing is a symptom of high fixed cost. The policy task is to tell the difference and to make the direct transfer path cheap enough that leasing is chosen for business reasons rather than forced by settlement friction.

Small operators pay the highest friction rate

Transaction costs are not neutral. A legal bill, a three-week delay, a notarization requirement, an unclear NIR path or a confusing use-plan request does not mean the same thing to every participant. For a large carrier or cloud platform, it is annoying. For a small ISP, hosting company, university network, island operator or early-stage data-centre business, it can decide whether the transaction is possible.

This is why transfer-market architecture should be judged by incidence, not just rule text. APNIC may apply the same formal conditions to all recipients. The practical burden falls unevenly. Large organizations can hire counsel, maintain compliance teams, collect utilization evidence, attend policy meetings, understand APNIC language, extend escrow and wait. Small operators often cannot. They may have one engineer handling routing, billing, customer complaints, abuse mail and registry paperwork. They may not know whether a use plan is persuasive. They may not have easy access to official corporate documents in English. They may depend on an NIR hostmaster or broker to explain the process.

The old needs-based language sometimes presents this inequality as fairness. Everyone must show need; therefore everyone is treated equally. But equal review is not equal cost. Price is visible. Discretion is not. A small operator can budget for a known transfer price. It struggles to budget for uncertainty, delay and the possibility that an administrative reviewer will not understand its business model.

This is not an argument for abandoning verification. A forged transfer harms small operators too. A hijacked block can damage innocent networks. Abuse-heavy leasing can pollute address reputation for everyone. The point is proportionality. A routine transfer of a clean /24 to a small operator should not require the same practical effort as a complex cross-border acquisition of a large portfolio with historical records and multiple jurisdictions. Evidence burden should scale with risk, not with the registry's institutional comfort.

Small-operator liquidity matters on both sides. A small holder that no longer needs a block may want to sell it to fund network upgrades, debt, IPv6 deployment, fibre, equipment or business survival. If transfer friction is high, the holder receives less. The discount may be captured by a broker, a better-capitalized buyer or simply lost in a failed sale. A small buyer that needs a block may lease instead because purchase is too complex, or may buy through a broker at a higher all-in cost. In both cases the settlement layer changes economic outcomes.

APNIC can reduce this burden without becoming lax. It can provide a small-transfer pathway with clear evidence checklists, plain-language use-plan examples, NIR-specific guides, translated support where practical, early comfort on obvious eligibility questions, published service timelines, staged cure processes and predictable treatment of operational records. It can report median and percentile processing times by transfer type and size. It can identify common reasons for delay. It can make it easier to know before signing whether a block is subject to lock-up or dispute. It can separate transfer-record integrity from broader account issues unless the latter directly matters.

The goal should be to minimize fixed cost per transaction. Fixed costs are the enemy of small-market liquidity. They make large transactions efficient and small transactions uneconomic. A transfer architecture that works only for large blocks and sophisticated parties will gradually centralize liquidity even if the rules never say so.

This matters for the Asia-Pacific because many socially valuable networks are not the largest. Rural access providers, island connectivity projects, regional ISPs, local hosts, educational networks and smaller enterprises may depend on modest address capacity. If APNIC wants a healthier regional internet economy, lowering settlement cost for legitimate small transactions is more useful than preserving allocation-era rhetoric after the free pool has gone.

There is also a democratic-economy point. Policy meetings and mailing lists tend to overrepresent people with time, travel budgets, institutional fluency and English-language confidence. Transfer rules affect many participants who are absent from those rooms. A rule can be developed openly and still impose unseen costs on absent parties. For high-consequence transfer rules, market-impact analysis should ask who bears the cost, what problem is being solved, whether the rule is a ledger safeguard or a market control, and how implementation will be measured.

Fraud control must not become capital control

Every serious transfer market attracts abuse. IPv4 is no exception. Scarcity and price create incentives for forged authority, hijacked resources, shell sellers, manipulated contact records, undisclosed disputes, sham corporate transactions, bad-faith leasing, abuse-heavy buyers and money movement designed to evade scrutiny. APNIC would fail its members and the internet if it treated every transfer request as innocent paperwork.

Fraud control is therefore a core settlement function. It is not optional. The registry must verify that the source is the legitimate holder. It must ensure that the resource is not subject to an unresolved dispute that makes transfer unsafe. It must check that the recipient is a real organization with the account relationship required by policy. It must demand sufficient legal documents for mergers, acquisitions and reorganizations. It must protect historical records. It must prevent duplicate recognition. It must maintain a public transfer log. It must coordinate with counterpart RIRs when resources cross regions. It must avoid enabling known unlawful conduct.

But fraud control has a boundary. It should not become capital control. The settlement layer should ask whether the transfer is valid, authorized and operationally coherent. It should not decide whether the price is too high, whether the buyer is too commercial, whether a seller should have kept the addresses, whether leasing is morally inferior to direct network operation, whether a large buyer already owns enough, or whether a smaller buyer's business plan is institutionally preferred. Those are market judgments unless a specific policy rule clearly and narrowly applies.

This boundary is especially important after exhaustion. During the free-pool era, needs assessment was the ordinary allocation mechanism. The registry was deciding how much scarce new space an applicant should receive from the common pool. After exhaustion, transfers mostly move resources already held by someone else. The seller gives up recognition. The buyer receives recognition. The registry still must protect accuracy and policy, but the normative justification for broad allocator discretion is weaker. It is no longer distributing free inventory. It is recognizing a legitimate change in control of an existing resource.

APNIC's current framework still contains allocation-era logic. The 24-month plan, past usage rate and compliance evidence extend need-based review into transfers. The five-year 103/8 rule protects final-pool integrity. These may be defensible. The danger is mission creep. A review designed to stop abuse of scarce public inventory can become a general veto over market movement. A compliance check designed to prevent fraud can become a broad audit of business model. A dispute check designed to prevent duplicate claims can become a pause button for tactical objections.

The market needs the narrow version, not the expansive one. APNIC should define fraud indicators, authority defects, conflict statuses, documentation gaps and policy blockers as precisely as possible. It should publish reasons for denial or delay in categories that members can understand. It should provide cure paths for fixable defects. It should maintain escalation or review mechanisms for high-consequence decisions. It should avoid changing expectations mid-transaction unless the change is legally required or clearly signalled.

Sanctions and compliance require the same discipline. In a cross-border region, counterparties may face sanctions screening by banks, escrow providers, brokers or corporate compliance teams. APNIC may also have obligations under applicable law. Those obligations should be handled as legal constraints, not as discretionary geopolitical preference. If a transaction cannot proceed because a party is legally restricted, the registry should say the category of restriction as far as confidentiality permits. If the issue is enhanced due diligence, the required evidence should be clear. If no legal restriction applies, sanctions anxiety should not become a hidden policy tool.

The settlement architecture should therefore be risk-based but not ideology-based. It should stop forged transfers. It should expose false provenance. It should block policy evasion where the policy is narrow and known. It should preserve records during disputes. It should support law-compliant screening. It should not regulate price discovery, suppress liquidity or decide that scarce IPv4 should move only to institutions whose plans resemble the old allocation model.

The test is simple. If APNIC's intervention protects the integrity of the record, it is a settlement safeguard. If it substitutes APNIC's commercial preference for the parties' bargain, it is capital control. The first builds confidence. The second creates discounts and workarounds.

A cheaper ledger for a scarcer market

A good registry makes important things cheaper. It makes it cheaper to know who holds a resource, cheaper to prove authority, cheaper to transfer a clean block, cheaper to identify fraud, cheaper to coordinate reverse DNS, cheaper to maintain routing-security assertions and cheaper to resolve disputes without harming innocent networks. That is the economic standard APNIC should use: whether it reduces the cost of legitimate recognition more than it increases the cost of market participation.

Several practical tests follow. Recognition finality should be explicit enough to support escrow and financing. Provenance should be standardized enough that clean source status, historical legitimacy, corporate succession, absence of dispute, 103/8 status and NIR path can be shown without bespoke negotiation. Recipient review should be narrow and measurable; if a 24-month plan, past usage evidence or compliance evidence is required, the buyer should know the standard before committing capital. NIR and inter-RIR paths should be mapped as settlement paths rather than left to specialist memory. RPKI, reverse DNS, route records, abuse contacts and public registration data should be treated as part of delivery, not as afterthoughts. Fraud control should be strong but bounded. Public transfer logs should be supplemented by aggregate friction data where confidentiality permits. Small-operator incidence should be an explicit policy test, because a rule that is cheap for a multinational and expensive for a regional ISP is not neutral merely because the wording is the same.

None of this requires APNIC to abandon stewardship. It requires APNIC to define stewardship for a post-exhaustion economy. The old question was how to distribute new scarce space without waste. The new question is how to maintain a reliable recognition layer for existing scarce space without suppressing liquidity, hiding cost or turning verification into control.

The economic future of APNIC's transfer market will be decided less by the number of IPv4 addresses remaining than by confidence in the recognition layer. Buyers will ask whether APNIC-region resources are easy to diligence, transfer, secure with RPKI, maintain with reverse DNS and move across borders where policy allows. Sellers will ask whether they can realize value without procedural uncertainty. Brokers will ask whether they are being paid for real execution or avoidable opacity. Small operators will ask whether the market is accessible without institutional fluency they do not possess.

The answer will determine liquidity. If APNIC lowers settlement cost, Asia-Pacific resources become more attractive. If it raises settlement cost, resources carry a registry-risk discount. A single-market assumption will not work for 56 economies, seven NIRs, multiple legal systems, multiple languages, direct and indirect membership paths, historical and allocated resources, inter-RIR transactions, sanctions-sensitive counterparties and networks whose administrative capacity ranges from global carrier to tiny local provider.

APNIC does not need to solve every market problem. It cannot make IPv4 abundant. It cannot guarantee global routing. It cannot eliminate bad actors. It cannot make all economies equally rich, all documents equally simple or all business plans equally legible. Its comparative advantage is narrower and more valuable: it can make the ledger trustworthy.

Trustworthy means accurate. It means final. It means bounded. It means evidence-based. It means service continuity where possible. It means public enough to reduce information asymmetry and confidential enough to protect private transactions. It means strong against fraud and modest against commerce. It means recognizing that after exhaustion, registry discretion affects capital value even when the registry insists it is only administering policy.

The right metaphor is a settlement layer, not a sovereign. A sovereign taxes, regulates, redistributes and polices according to political authority. A settlement layer verifies, records, reconciles and maintains continuity according to evidence. APNIC has moral authority when it performs the second role well. It risks legitimacy when it drifts toward the first without the law, capital, accountability or political mandate that such a role would require.

IPv4 scarcity turned registry recognition into a high-value institutional service. That fact cannot be wished away by saying addresses are public resources or that policy is community developed. The service now sits inside a market. Its design will decide who can transact cheaply, who must rely on brokers, who receives a liquidity discount, who leases instead of buying, who can prove title-like confidence and how much uncertainty the Asia-Pacific region adds to a global address economy.

APNIC can be the neutral ledger that lowers those costs. Or it can be the checkpoint that makes them larger. In a region this diverse, the difference will not remain abstract. It will be paid in escrow extensions, lost discounts, failed small-operator deals, broker dependence, operational handover errors and cross-border uncertainty.

The better path is restrained. Verify the parties. Verify the resource. Verify authority. Preserve provenance. Enforce narrow waiting rules transparently. Coordinate with NIRs and other RIRs. Maintain RPKI, reverse-DNS and registration continuity. Stop fraud. Publish friction. Then settle the transfer.

That role makes every other part of the market possible.