The market APNIC can see only partly
IPv4 leasing is often treated as a secondary market footnote: a temporary substitute for purchase, a bridge until a buyer raises capital, or a practical arrangement between a holder with idle space and an operator with customers to serve. That description is not false. It is too small. In the APNIC region, leasing is also a parallel allocation system created by the gap between registry recognition and operating demand.
The public registry answers one set of questions. Which legal entity is recorded as the recognized holder of a number resource? Which contacts are attached to that record? Which public registry services point outward from it? A leasing market answers another set. Who can make the addresses usable this month? Who can announce the prefix, place customers on it, pay for it, defend its reputation, answer abuse mail, maintain reverse DNS, request route authorization, and keep the service online when a customer contract depends on it? When those answers diverge, the market does not wait for a neat institutional theory. It builds around the divergence.
That is the fact APNIC should take seriously. Leasing exists because the Asia-Pacific region contains too much variation for a single allocation-era story to describe. Some networks have address space but no immediate deployment demand. Some have demand but not the balance sheet to buy. Some can navigate registry process quickly. Others face National Internet Registry paths, local documentation burdens, language barriers, unfamiliar corporate forms, sanctions screening, cross-border payment review, tax questions, banking delays, or uncertainty over whether a formal transfer can close in the time their customers require. Some operators can use IPv6 aggressively. Many still cannot substitute IPv6 for IPv4 customer reachability, mail delivery, hosting panels, fraud controls, payment integrations, VPN gateways, mobile access, residential broadband, enterprise allowlists, or legacy software.
Leasing appears when operational need has a shorter clock than registry finality. It lets a data-centre provider add customers before it can finance a portfolio purchase. It lets a small ISP preserve continuity while customer growth outruns its historic allocation. It lets a cloud or hosting company absorb a demand spike without committing capital to addresses it may not need forever. It lets a holder earn yield from dormant inventory without selling strategic optionality. It lets a broker package capacity for users that cannot navigate the address market alone. It lets a multinational place usable IPv4 capacity near an Asian customer base while formal registration, corporate approval, tax review or inter-regional transfer timing would delay the launch.
None of this makes leasing illegitimate. It can preserve service continuity, release idle capacity, reduce waste and help smaller networks that would otherwise be priced out of outright purchase. But it changes the location of responsibility. The registry record may still point to the lessor. The economic benefit may be split among lessor, broker and lessee. Operational control may sit with the lessee or its downstream customers. Abuse complaints may reach a contact who does not control the user causing the problem. RPKI may depend on a certificate held by the lessor. Reverse DNS may remain inside the lessor's management. Route objects may be created by an intermediary. Reputation damage may attach to the prefix after the lessee has left. Payment risk may sit in escrow, prepayment, deposit or monthly settlement terms. Compliance screening may be performed by the broker, lessor, bank, lessee, customer, or nobody with a full view.
That is why leasing can become shadow allocation. It is not shadow allocation merely because it is private. The internet has always depended on private contracts. It becomes shadow allocation when the real control surface of the addresses moves into private arrangements while the public registry continues to describe only the formal holder. The ledger remains correct in a narrow sense, yet economically incomplete in the sense that matters to operators, counterparties, abuse desks, routing-security systems and customers.
The answer is not to pretend leasing does not exist. It does. Nor is the answer to turn APNIC into a price regulator, credit allocator or moral court deciding who deserves scarce addresses. APNIC is not equipped for that role, and the neutrality of the registry would suffer if it tried. The useful institutional line is narrower and more practical: the registry layer should remain legible enough that address use, authorization, responsibility, provenance and contactability do not disappear when a private lease carries the operating reality.
Scarcity did not create one APNIC market
The Asia-Pacific region is not a single commercial setting. It contains mature carrier markets, high-growth access markets, island networks, national telecom incumbents, hyperscale platforms, local hosting providers, research networks, security companies, content networks, universities, banks, public agencies, small ISPs and internet exchanges. The same /24 can have very different value depending on where it is needed, which customers it supports, how quickly it can be routed, how clean its reputation is, and how much registry or legal process surrounds it.
APNIC's public material identifies a service region spanning 56 economies and recognizes seven National Internet Registries: APJII in Indonesia, CNNIC in China, IRINN in India, JPNIC in Japan, KISA in Korea, TWNIC in Taiwan and VNNIC in Vietnam. That structure is sensible for a region of this size. It gives large economies local-language and locally based registry service while keeping them inside a regional policy community. It also has market consequences. A direct APNIC-held resource, an NIR-mediated resource, a historical block, a resource acquired through a merger, a domestic transfer, a cross-border transfer and an inter-RIR transfer are not identical economic instruments, even if the global routing table sees only prefixes.
Leasing grows in that difference. A buyer of address space must care about recognition finality. A lessee often cares first about usability. Can the prefix be announced from the intended ASN? Can the lessor create and maintain the right Route Origin Authorizations? Can reverse DNS be delegated? Can route objects be made to satisfy upstream filters? Can the block pass reputation checks? Can the lessee put customers on it without waiting for the full legal and registry process required by purchase? Can the contract start this quarter rather than after the next budget cycle? Can the operator avoid carrying a long-lived scarce asset on the balance sheet while still keeping customers online?
This is not only a transfer-friction story, though transfer friction is one cause. Leasing also follows from uneven working capital. An address purchase concentrates cost upfront. A lease spreads cost over time. That difference matters to small ISPs, regional hosting companies and young data-centre businesses whose customer revenue arrives monthly while purchase prices require immediate capital. A network may have a profitable use for a few thousand addresses and still be unable to justify or finance a purchase. The customers may be real, the need urgent and the project economically sound. The balance sheet may simply be too thin.
Uncertain demand creates the same pattern. A hosting platform may need capacity for a new product without knowing whether the product will survive. A broadband operator may need addresses while carrier-grade NAT is being tuned. A regional cloud provider may need IPv4 for an enterprise customer whose contract runs three years, not forever. A security company may need clean ranges for a managed service. A gaming, payments or SaaS platform may need IPv4 because counterparties, fraud systems, monitoring tools or older software still fail on IPv6-only assumptions. Buying a long-lived asset for a short or uncertain revenue stream can be rationally unattractive. Leasing is an option-like instrument. It buys time, flexibility and usability.
Geography and institutions add another layer. A network in a smaller economy may be far from the main broker networks. Its corporate documents may not be in English. Its legal form may be unfamiliar to counterparties. Its bank may impose cross-border payment review. A seller may be in an NIR economy while the buyer is not. A buyer may face sanctions, export-control, telecom-license or customer-screening questions that delay payment even when registry policy is not the main obstacle. These frictions do not abolish demand. They redirect demand toward structures that can be executed with less formal finality.
Leasing is therefore an adaptation to heterogeneity. It is not an exception at the edge of a clean transfer market. It is the market's way of converting scarce, lumpy and institutionally encumbered capacity into smaller, time-bound, financeable units of use. The more diverse the region, the more valuable that conversion becomes.
Leasing as working-capital arbitrage
The phrase "IPv4 leasing" can sound like a simple rental. Economically it is closer to working-capital arbitrage. One side holds a scarce asset whose immediate operating value may be lower than its market value. The other side has operating demand but wants to preserve cash. The lease converts the holder's inventory into income and the operator's capital expenditure into operating expenditure. That conversion is the core product.
For the lessor, idle address space is inventory. It may have been obtained years ago, received through corporate history, held as reserve, recovered from a discontinued product, or accumulated by a larger network whose present needs differ from its original allocation. Selling the block releases capital but ends future upside. Leasing preserves optionality. The holder earns yield while keeping the formal registry position. If IPv4 prices rise, the holder still owns the asset-like interest. If a strategic use appears later, the holder may reclaim the block after the lease term. If investors or lenders value recurring revenue from scarce inventory, leasing can make an otherwise dormant resource look like a yield-bearing asset without liquidation.
For the lessee, leasing turns a scarce capital good into a monthly input. That may be the difference between serving customers and declining business. A small hosting company may not have the cash to buy a /22, but it may have customers who will cover the monthly lease cost. A regional ISP may be able to finance backhaul, routers, power and installation but not a large address purchase. A cloud provider may prefer to match address cost to revenue from a specific customer segment. A start-up may need IPv4 during growth but avoid buying until its long-term product mix is clearer.
This is not inherently a loophole. It is inventory finance. Warehousing, aircraft leasing, equipment leasing, spectrum leasing, dark-fibre leases and data-centre capacity contracts all perform similar functions in other industries. The owner of a scarce or expensive asset supplies use to an operator that cannot or should not bear full ownership cost. The market does not treat that as suspicious when the asset is a truck, an office building, a server rack or a fibre pair. It should not become suspicious merely because the asset is a number resource, provided the public responsibilities attached to use remain visible and enforceable.
The important difference is that IPv4 depends on a public coordination layer. A leased truck can be identified by its license plate and insured by its operator. A leased office has a landlord, a tenant, utility accounts and local records. A leased prefix sits inside a registry system designed mainly around holder recognition, not around temporary delegation of operational control. That is where the institutional problem begins.
When leasing is clean, the lessor and lessee align the operational pieces. The lessor confirms that it is the recognized holder. The contract defines permitted use, term, payment, termination, abuse duties, subdelegation, routing arrangements, RPKI, reverse DNS, contact handling, reputation warranties, lawful-use obligations and return conditions. The parties agree who creates ROAs, who maintains route objects, who answers abuse mail, who may announce from which ASNs, who receives notices from network operators or public authorities, who handles blacklisting, what happens after a corporate control change, and how customer migration occurs when the lease ends.
When leasing is weak, only part of that is true. The lessee routes the space but cannot update public data. Abuse reports go to stale contacts. The lessor retains RPKI control and can break reachability by mistake, delay or dispute. Reverse DNS points to names the lessee cannot change. A broker holds the commercial relationship but not the registry authority. A downstream customer receives addresses without knowing whom the registry recognizes. The lessor collects rent but does not monitor reputation. The lessee bears service risk but cannot prove authority to third parties. A payment dispute becomes a routing event.
Working-capital arbitrage then becomes institutional arbitrage. The market is no longer only shifting payment timing. It is shifting responsibility away from the public record. That does not mean leasing should be banned. It means leasing needs a legible responsibility layer.
APNIC does not need to know lease prices, approve monthly rates, judge whether a holder earns too much, or choose between lease and sale as a preferred outcome. But if the registry record remains silent while the operating reality moves elsewhere, the public layer loses some of the information that makes it useful. The task is not to control the capital bargain. It is to keep the operational truth discoverable enough for the internet to function safely.
The small-operator liquidity trap
The strongest case for taking leasing seriously is not the convenience of large market actors. It is the small-operator liquidity trap. The old allocation world assumed that a network could document need and receive space according to policy. Exhaustion changed that. After exhaustion, need alone does not produce supply. Supply must be found, priced, contracted and made usable. That transition favours operators with cash, staff, lawyers, brokers and time.
Small networks often have little of any of those. Their need is immediate and practical. Customers require public addresses for hosting, remote access, payment systems, cameras, mail servers, enterprise VPNs, static-IP packages, gaming services, compliance tools or old applications that cannot sit comfortably behind shared NAT. The operator may be small because the market is small: a regional town, an island, a rural area, a specialized hosting niche, a local enterprise base, an education network or a developing broadband footprint. The customer demand can be socially and commercially valuable even when the operator's balance sheet is modest.
Outright purchase is difficult in that setting. A block is lumpy. Prices move. Brokers expect proof of seriousness. Sellers may prefer larger deals. Legal review costs do not scale down neatly. Escrow fees, due diligence and document preparation impose fixed costs. Registry process adds uncertainty. If the buyer must show need, corporate authority, account standing and policy compliance, the burden may be manageable but still expensive relative to the transaction. A large carrier treats the process as routine. A small ISP experiences it as a financing obstacle.
Leasing solves the cash-flow mismatch. It lets the operator convert expected customer revenue into address access. The lease can be sized closer to demand. It can start sooner. It can be cancelled, reduced or expanded more easily than a purchase. It avoids tying scarce cash into a long-lived asset when the operator also needs routers, support staff, ducts, power, colocation, marketing and customer installation. In this sense, leasing can be pro-competition. It prevents address scarcity from becoming an entry barrier that only the best-capitalized networks can cross.
The trap is that the same structure that helps small operators also makes them dependent. A small lessee may have little bargaining power over RPKI, reverse DNS, abuse handling, renewal, migration time or price escalation. It may accept terms that let the lessor withdraw capacity quickly. It may build customer packages around prefixes it does not control. It may have no practical ability to move customers if the lease ends, the price jumps, the lessor sells the block, the broker disappears, or the registry account becomes contested. It may be the party closest to the customer and farthest from the official record.
This dependence is not theoretical. IPv4 addresses become embedded. Customers create DNS records, firewall rules, reputation history, payment-verification patterns, VPN tunnels, allowlists, security exceptions and support expectations around them. A hosting provider that loses a prefix does not merely replace numbers. It migrates customers. It explains outages. It repairs email reputation. It updates reverse DNS. It fights blacklists. It may lose customers whose own systems cannot be renumbered quickly. A broadband ISP that changes address pools may face customer complaints, CGNAT pressure, lawful-intercept configuration changes, geolocation errors and support load.
Leasing therefore transfers some of the capital burden out of the small operator while leaving much of the continuity burden inside it. That is not inherently unfair if the operator understands and prices the risk. It becomes unfair when the public registry and the contract environment make the risk hard to see. If the lessee cannot verify provenance, cannot confirm who controls RPKI, cannot know whether the lessor is in good standing, cannot rely on contact continuity and cannot discover whether the block has a serious reputation history, it is not entering a market with transparent risk. It is entering an information maze.
APNIC's role should be to reduce that maze where the public coordination layer is implicated. The small operator does not need APNIC to subsidize prices. It needs a registry environment in which the minimum facts that affect continuity are visible: recognized holder, abuse and routing contacts, authorization path, relevant public restrictions, dispute status where disclosure is appropriate, and clear mechanisms for maintaining public data during temporary use. That makes leasing safer without turning the registry into a lender, broker or landlord.
CGNAT and the incomplete IPv6 substitution
Leasing cannot be understood without the unresolved transition problem. IPv6 deployment is real and important, but it has not eliminated operational dependence on IPv4. The APNIC region contains advanced IPv6 networks and many environments where IPv4 remains the customer-facing compatibility layer. The result is not a clean migration. It is a long dual-stack and translation economy.
Carrier-grade NAT is the most visible symptom. CGNAT lets access providers place many customers behind fewer public IPv4 addresses. It is economically rational when addresses are scarce. It also creates costs. It complicates logging, abuse attribution, troubleshooting, port availability, gaming, remote access, lawful requests, enterprise support and some application behaviour. It can reduce customer experience and increase support load. It can make a network look cheaper in the address market while more expensive in operations.
A small ISP facing customer growth may choose between buying addresses, leasing addresses, expanding CGNAT, pushing IPv6 harder, or accepting service degradation. None of those choices is ideological. They are P&L decisions. If IPv6 alone cannot satisfy customers and buying IPv4 requires capital the operator does not have, leasing becomes the practical middle path. It preserves continuity while avoiding a major capital purchase.
Hosting and cloud providers face a different but related problem. Many services still assume IPv4 reachability because customers, enterprise partners, payment providers, anti-fraud systems, DNS tooling, monitoring platforms and old software stacks assume it. An IPv6-only server may be technically sound and commercially unusable for the wrong customer base. The operator may therefore lease IPv4 not because it rejects IPv6, but because it sells into an internet that has not completed the substitution.
This distinction matters. If a registry or policy community treats leasing as evidence that operators are refusing progress, it misreads the market. Leasing may indicate that IPv6 substitution is incomplete, that customer continuity is more urgent than architectural preference, and that operators are rationally avoiding forced renumbering or service loss. It may also indicate that addresses are priced high enough to justify yield, but not high enough for every user to buy. That is a normal scarcity market.
The danger is that delayed substitution can make temporary leasing permanent in practice. A lessee begins with a one-year bridge. Customers are placed on the prefix. Reverse DNS, mail reputation, geolocation and security rules accumulate. The lease renews. The operator's business plan assumes continued access. The lessor's yield becomes recurring income. The prefix becomes part of the lessee's operating identity even though the registry still describes someone else.
At that point the institutional issue is no longer whether the lease was a practical bridge. It is whether the public record reflects a durable operational dependence. A registry that sees only the holder may miss where continuity risk really sits. If a lessor-controlled RPKI change invalidates a route, the customer outage occurs at the lessee. If an abuse desk cannot reach the real operator, reputation cost spreads. If the lessor's account is suspended or a corporate dispute freezes changes, the lessee's customers become collateral.
APNIC should not force IPv6 by making IPv4 leasing opaque or precarious. That would punish the networks carrying the current internet. The better approach is to acknowledge that IPv4 remains economically necessary in many settings and to make the responsibilities around temporary use clear. IPv6 progress and IPv4 leasing transparency are not opposites. A cleaner leasing layer can reduce the damage of the prolonged transition while operators continue investing in dual-stack, translation and eventual substitution where it is commercially possible.
NIR paths and cross-border asymmetry
National Internet Registries are one of APNIC's distinctive regional features. They help localize registry service in large and complex economies. They also create differences in how market participants experience address access. A leasing market will find and price those differences.
An operator in an NIR economy may have local support, local language and familiar procedures. That can reduce friction. But a cross-border lessee or lessor may not know how the local layer interacts with the regional record, which documents are needed, who can change public contact data, how quickly routing-security changes can be made, or how a lease should be reflected without implying a formal transfer. A broker may understand the path; the end user may not. That knowledge gap becomes an information rent.
Direct APNIC account holders face a different set of issues. They may have a clearer regional interface but still need to manage cross-border contracts, tax treatment, payment rails, corporate authority, escrow, sanctions screening and legal enforceability. A lease between two firms in the same economy is different from a lease among a lessor in one jurisdiction, a broker in another, a lessee in a third and downstream customers across several more. The addresses are global. The contract is local. The registry layer is regional. The route is operational. The bank is regulated elsewhere. No single actor naturally sees the whole map.
That creates two forms of asymmetry. The first is process fluency. Repeat participants know what evidence satisfies a hostmaster, which registry fields matter, how to structure RPKI and reverse DNS obligations, which escrow terms are common, how abuse clauses are enforced, and when a proposed use will raise questions. First-time participants do not. They may pay a broker because the broker knows the path, not because the broker found unique supply.
The second is bargaining asymmetry. The party that controls the registry account usually controls the public levers. If the lessor is sophisticated and the lessee is not, the lease can leave the lessee dependent on lessor action for every operational change. If the lessee is a large platform and the lessor is a small holder, the lessee may demand warranties, indemnities and service levels the lessor cannot fully price. If the broker stands between them, it may collect a spread by reducing friction, or it may benefit from opacity that keeps both sides dependent.
Intermediaries are not the enemy. In a fragmented region they can match supply and demand, standardize contracts, screen counterparties, coordinate escrow, arrange route authorization, track reputation and translate registry requirements. But the market should distinguish useful execution from avoidable opacity. If a broker earns a fee by carrying diligence and risk, that is productive. If it earns a fee because ordinary users cannot understand public responsibilities, the registry layer is too obscure.
APNIC cannot and should not standardize all private contracts across 56 economies. It can standardize what the public layer needs to preserve: contactability, authorization, holder accountability, operator reachability and continuity. The NIR dimension makes that especially important. Regional confidence requires a minimum common meaning across local paths. If a lease involving an NIR-held resource is operationally opaque to parties outside that local path, liquidity suffers. If local implementation can preserve contactability and routing-security continuity while giving local service, leasing becomes safer. The market does not require every path to be identical. It requires the responsibility surface to be intelligible.
The broker's rent and the price of hidden information
Every scarce market produces intermediaries. IPv4 is no different. Brokers, leasing platforms, consultants and specialist counsel exist because finding supply, checking provenance, arranging terms and navigating registry process are not trivial. In the APNIC region, where legal systems, languages, NIR paths and operator sizes vary widely, intermediaries are often useful.
The economic question is what exactly they are paid for. A broker can create value by finding a legitimate lessor, screening a lessee, checking abuse reputation, confirming registry status, coordinating ROAs, arranging reverse-DNS delegation, drafting abuse clauses, setting payment milestones, monitoring route origination and responding when a prefix is blacklisted. Those are real services. They reduce risk and make scarce inventory usable.
A broker can also profit from information that should not be scarce. If only insiders know which public fields matter, how to reach the right contact, how to identify a contested prefix, how to structure a lease without breaking route security, or how to interpret basic APNIC and NIR expectations, the broker earns rent from public-layer opacity. The fee may be legal and commercially accepted, but the institutional design is poor. The registry layer should not unintentionally create a market for decoding the registry layer.
Leasing magnifies this because the transaction does not end at recognition finality. In a sale, the market's attention often concentrates on closing. In a lease, the relationship continues. Monthly payments, renewals, abuse events, customer changes, route changes, reputation problems, geolocation complaints, RPKI updates and termination procedures all require continuing coordination. An intermediary that stays involved may be valuable. An intermediary that disappears after matching leaves the parties with a living dependency they may not know how to operate.
Informational rents appear most sharply in reputation. IPv4 blocks have histories. They may have been used for clean access networks, mail abuse, bulletproof hosting, hijack episodes, proxy services, botnet infrastructure, VPN products, cloud workloads, or nothing visible for years. Reputation does not always belong neatly to the holder or lessee. It attaches to the prefix, origin ASN, reverse-DNS patterns, abuse response history and external lists. A lessee may pay for addresses that look available but carry hidden deliverability or filtering costs. A lessor may place a clean block with a lessee that damages it and returns it impaired.
The broker who can read those histories earns a legitimate fee. But a healthier market would make some basic signals easier for ordinary participants to inspect. Not every blacklist or private reputation score can be public or accurate. Yet registry data can help by preserving contact continuity, route authorization clarity, public transfer or status signals where applicable, and stable abuse channels. The less the registry layer says, the more the market depends on private intelligence that small operators cannot afford.
Payment structure is another source of rent. Deposits, monthly prepayment, escrow, termination windows, service-level promises, indemnities and penalties reflect real risks: non-payment, abuse, sanctions exposure, reputation damage, sudden withdrawal, route invalidation and customer loss. Escrow can reduce credit risk, but it does not solve control. If the lessor alone can change RPKI or reverse DNS, escrowed money does not prevent an outage. If the lessee alone controls downstream customers, a deposit does not prevent abuse from damaging the prefix.
The better contract therefore needs money terms and control terms: who may announce, who requests ROAs, who maintains reverse DNS, who handles abuse, who may sublease, which customer categories are prohibited, what happens after sanctions or blacklist events, how much notice precedes route changes, and how return and reputation damage are handled. APNIC should not write those contracts for the market. But APNIC can make clear which public coordination facts must remain accurate no matter what private terms are chosen. The broker's job should be commercial execution and risk placement, not making public responsibility disappear.
The operational bundle that makes a lease usable
A leased IPv4 block is not delivered merely by saying "you may use it." Usability is an operational bundle. The lessee needs global routing to accept the announcement, security systems not to reject it, reverse DNS to support services, contacts to receive complaints, and records that do not contradict the real operator. If these pieces remain under lessor control without clear duties, the lease is fragile.
RPKI is the most visible control point. Route Origin Authorizations can make a route look valid or invalid to networks that use origin validation. If the lessor holds the certificate and must create ROAs for the lessee's ASN, the lessee depends on the lessor's responsiveness and competence. A typo, delay, account problem, termination dispute or misunderstanding can make the route invalid or leave it less trusted than expected. If a broker requests the change, another layer is added. If the lessee changes upstreams or origin ASNs, the dependency repeats.
Reverse DNS is less dramatic but commercially important. Mail systems, security tools, enterprise customers and diagnostics often depend on usable reverse-DNS delegation. A lessee that cannot update reverse DNS may struggle to provide hosting, mail, VPN or managed services. A lessor that delegates reverse DNS but does not define return procedures may face cleanup problems when the lease ends. Stale reverse DNS can create reputational confusion long after a customer has left.
Internet Routing Registry data and route filters add another layer. Some networks still rely on IRR route objects to build filters. A lessee may need route objects aligned with its origin ASN. The authority to create or maintain them can depend on registry records, maintainer credentials, authentication paths and upstream expectations. If the lease ignores route objects, the lessee may obtain theoretical use of addresses that upstreams will not route cleanly.
Contact data is the simplest and most neglected issue. Abuse contacts, technical contacts and administrative contacts are the public path by which problems reach a responsible party. In a lease, responsibility may be split. The lessor is the recognized holder. The lessee controls customers. A downstream customer may be the source of abuse. A broker may receive the complaint first. If the public record exposes only the lessor and the lessor forwards slowly, abuse response suffers. If the public record exposes only the lessee without preserving the holder's accountability, the registry may lose the link to the recognized resource holder. The answer is not necessarily one contact. It may be layered contactability.
This operational bundle is where shadow allocation becomes risky. The registry can be accurate about holder identity and still fail to describe who can fix a live problem. A network operator deciding whether to filter a prefix, a victim reporting abuse, a public authority seeking a responsible contact, or a customer diagnosing reverse DNS does not care only who signed the original allocation papers. It needs the party that can act.
The market often tries to solve this privately through contract clauses. Private clauses do not help outsiders who need contactability. They also do not help when the parties disagree. If the lessee says the lessor failed to update RPKI, the lessor says the lessee failed to pay, and customers are offline, the public internet needs a stable operational state while the commercial dispute is resolved. The registry should not adjudicate the payment dispute, but it should be designed to avoid making customers hostage to a private argument.
One useful principle is continuity of last verified operating state. If a lease is active and public route authorization exists, routine commercial disagreements should not casually erase route validity before notice and process. If a lease ends, return procedures should be predictable enough that the lessee can migrate and the lessor can protect the block. If abuse is severe, contacts and escalation should lead to a party capable of action. None of this requires APNIC to police contract morality. It requires the registry layer to understand that temporary use creates real operational dependencies.
Abuse, reputation and compliance risk
IPv4 reputation is an asset and a liability. It affects email deliverability, fraud scoring, content-platform access, payment processing, enterprise allowlisting, security filtering, geolocation trust and the willingness of upstreams or customers to accept a service. In leasing, reputation risk often moves faster than formal accountability.
A lessor that leases clean space to a bad lessee can receive back a damaged prefix. The market value of the block may fall. Future lessees may demand discounts. Upstreams may ask questions. Abuse desks may remember the holder's name. The lessor may face complaints for conduct it did not perform. If the lessor has many leases, it becomes a portfolio manager of reputation risk, not merely an address holder.
A lessee faces the opposite risk. It may receive a block with existing blacklist entries, geolocation errors, poor mail reputation, association with prior abuse, or stale route records. The lease price may look attractive while the operational cost is high. Cleaning reputation takes time. Some reputation systems are opaque. Some databases lag. Some counterparties do not distinguish between prior and current users. The lessee may pay for capacity that cannot serve the intended customers immediately.
Brokers and leasing platforms sit between these risks. Good intermediaries screen counterparties, monitor abuse, maintain reputation histories, define permitted-use categories and intervene before a prefix is seriously damaged. Weak intermediaries treat reputation as someone else's problem. The difficulty is that public registry data alone does not reveal enough, while private reputation intelligence may be inaccessible to the smaller operators most exposed to harm.
Sanctions and compliance screening intensify the issue. The APNIC region includes cross-border counterparties whose banks, vendors, customers or regulators may face different legal obligations. A lease can involve a holder in one economy, a lessee in another, downstream users elsewhere, payment through a third, and traffic visible globally. The lessor may be concerned about sanctioned entities, prohibited industries, export-control exposure, fraud, gambling, adult services, cryptocurrency abuse, proxy networks or other high-risk uses. The lessee may be concerned that a lessor, broker or payment path becomes restricted during the term.
Those are not registry price questions. APNIC should not decide whether a lease price is fair because a sanctions-screening clause is expensive. It should not bless or condemn commercial sectors unless a narrow legal obligation applies. But the registry layer should make contactability and responsibility visible enough that lawful notices, abuse complaints and operational warnings reach someone who can act. If the public record is silent about the real operator, compliance risk does not vanish. It moves into private files and delayed email chains.
Escrow and payment terms also reflect reputation risk. A lessor may require deposit or prepayment because abuse can damage the block faster than monthly revenue covers. A lessee may require service credits or termination rights because route invalidation can destroy customer trust. A broker may hold funds to manage disputes. These terms are rational. They are market attempts to price risks that the registry does not carry.
The institutional danger is mispriced externality. If a lessee damages reputation and walks away, the lessor bears residual loss and future users inherit friction. If a lessor withdraws authorization abruptly, the lessee's customers bear outage costs. If abuse contacts are stale, victims and networks elsewhere bear investigation costs. If APNIC sees only the formal holder and not the temporary responsibility layer, the registry's public function underestimates the risk the market is distributing.
Reputation is therefore a balance sheet nobody records in the registry. It should not be turned into a subjective score controlled by APNIC. That would create new problems. But the conditions that allow reputation to be managed - accurate contacts, clear authority, route-security continuity, provenance and the ability to distinguish current from past operational responsibility - are registry-adjacent. Making those conditions clearer is a legitimate registry function.
When private allocation becomes shadow allocation
Private allocation is not automatically harmful. Every ISP assigns addresses to customers. Every hosting provider allocates space to servers. Every cloud platform gives customers address use under contract. The internet has always depended on downstream assignments that do not individually appear as full public registry changes. The question is when ordinary downstream use becomes shadow allocation.
The line is crossed when private arrangements replace the public registry as the place where meaningful control can be understood. If the formal holder merely owns the registry account while another party controls routing, customers, abuse response, reverse DNS, RPKI requests, commercial revenue and practical continuity, the public record is no longer describing the real control surface. It is describing a legal anchor without the operating map.
In small customer assignments this may not matter much. A broadband customer receiving one dynamic address from an ISP is clearly inside the ISP's operating responsibility. A server customer receiving one address from a hosting provider is clearly inside the provider's abuse and routing environment. A large lease of a prefix to an independent operator is different. The lessee may originate routes from its own ASN, serve its own customers, maintain its own abuse desk, control its own network and build its own reputation. The lessor may not see daily operations. The relationship starts to resemble allocation, even if the registry still calls the lessor the holder.
The APNIC region's diversity makes this line harder to see. A resource may be leased across economies, through a broker, to a network whose customers are elsewhere. NIR involvement may add local service paths. Historical resources may carry older records. Some leases may be temporary capacity for a project; others may become quasi-permanent operating blocks. Some lessees may be infrastructure providers; others may be resellers. Some may sublease further. Without public responsibility signals, outsiders cannot distinguish these cases.
Shadow allocation creates several institutional problems. It weakens accountability because the party on the public record may not be the party causing operational effects. Complaints bounce. Investigations slow. Networks make blunt filtering decisions because fine-grained responsibility is unavailable. It weakens provenance because a future buyer, lessee, upstream or customer may not know how the block was used, by whom, under whose authority, and with what operational history. It weakens continuity because a private dispute can affect routing, RPKI, reverse DNS and customer service. It weakens policy feedback because APNIC may see formal transfers and allocations while missing how address capacity actually circulates after exhaustion.
The remedy is proportionality, not making every downstream use a registry event. Large, independent, route-originating or long-duration leases should have a way to make operational responsibility visible without becoming transfers. Contact delegation, route authorization, reverse-DNS responsibility and abuse escalation can be clarified without publishing price or commercial terms.
APNIC should ask a practical question: when a block is leased, can the public internet identify who can act on operational problems, who authorized the route, how reverse DNS is maintained, and whether the recognized holder remains accountable? If the answer is yes, the lease is less shadowy. If the answer is no, the market has created an off-ledger allocation whose risks will appear later in abuse, disputes, outages and discounts.
Legibility without capital control
There is a tempting but wrong response to leasing risk: regulate the market as if APNIC were a public utilities commission. Set approved terms, inspect prices, decide acceptable margins, require moral justification for each lease, punish holders who monetize scarcity, or force lessees toward purchase or IPv6. That path would confuse the registry's function with economic planning. It would also push leasing further into opacity.
APNIC's comparative advantage is not price judgment. It is the maintenance of a reliable, trusted, public coordination layer. In leasing, that means illuminating responsibility without controlling the bargain. This is the line between legibility and capital control. Legibility asks whether the public record is accurate enough for the internet to identify authority, reach responsible contacts and preserve routing-security continuity. Capital control asks whether the institution should decide who may monetize scarcity, at what price and under which commercial terms. The first is a registry function. The second is a different political economy.
A useful framework separates four categories. The first is holder accountability. The recognized holder remains responsible for the resource relationship with APNIC and for ensuring that public data is not misleading. Leasing should not let a holder collect income while disclaiming all consequences of use. If a holder permits another operator to use space, the holder should ensure reachable contacts, route authorization, reverse-DNS arrangements and abuse escalation are maintained.
The second is operational contactability. The party actually operating the leased space should be reachable for abuse, technical and routing matters. That does not require publishing every customer or contract. It does require serious complaints and route-security issues not to vanish into the wrong mailbox. APNIC can encourage or require contact structures that show both holder and operator where appropriate.
The third is provenance and authorization. A lessee should be able to prove to an upstream, customer or counterparty that use of the prefix is authorized by the recognized holder. That proof should not depend solely on a private PDF nobody else can verify. RPKI, route objects, authenticated registry contacts and public delegation data can all serve as stronger signals. The more machine-readable and public the authorization, the less room exists for hijack-like confusion.
The fourth is continuity. Lease start, change and termination should be handled so that routing security, reverse DNS and contact data do not break unpredictably. If a lease ends, public records should be cleaned. If a lease renews, authorizations should not expire silently. If a dispute arises, there should be a default bias toward preserving verified operational stability while narrow issues are resolved, unless abuse or legal obligation requires urgent action.
None of those categories requires APNIC to know whether the monthly rate is high or low. None requires APNIC to approve the lessee's business model beyond narrow policy and legal constraints. None requires APNIC to decide whether leasing is morally better than purchase. They are ledger functions: accuracy, contactability, authorization and continuity.
APNIC can also improve market quality by publishing aggregate information and practical guidance without exposing confidential contracts: contact-data defects, route authorization errors, stale reverse-DNS failures, abuse-handling patterns, useful public-record checks and minimum operational clauses. The tone matters. If leasing is framed as inherently suspect, participants will hide it. If it is framed as ordinary but responsibility-bearing, they have a reason to make it legible.
Lessor-controlled dependencies and continuity
The most delicate feature of leasing is that the lessee may operate the network while the lessor controls the registry levers. This creates an agency problem. The party closest to customer harm may not hold the power needed to prevent it.
Consider RPKI. If a lessee's service depends on a ROA controlled by the lessor, the lessor's availability becomes part of the lessee's uptime. If the lessee changes upstreams and needs a new origin ASN authorized, delay can affect reachability. If the lessor's staff misunderstand a request, a valid route can become invalid. If a payment dispute escalates, the lessor can threaten route authorization. If the lessor's APNIC account faces its own issue, the lessee may suffer without having caused it.
Reverse DNS creates similar dependence. A lessee selling mail, hosting, VPN or enterprise services may need fast changes. If the lessor retains delegation, the lessee must ask. If the lessor delegates but can revoke abruptly, the lessee's services remain exposed. If the lease ends and reverse DNS is not cleaned, both parties may suffer.
Abuse handling is harder because incentives diverge. The lessee may want time to investigate a customer. The lessor may want immediate termination to protect the prefix. The broker may want to preserve revenue. Downstream victims want fast mitigation. Upstreams may threaten filtering. Without predefined responsibility, the loudest actor may drive the outcome.
This agency problem also affects investment. A lessee that cannot rely on stable access will underinvest in services tied to leased addresses. It may avoid high-value customers, refuse long contracts, or pass uncertainty into customer terms. A lessor that fears abuse will impose shorter terms, higher deposits or restrictive use clauses. Both sides are rational. The market becomes more expensive because control and operating exposure are separated.
One response is to say lessees should buy if they need control. That is commercially neat and institutionally incomplete. Many lessees lease precisely because purchase is not available, financeable or efficient. Telling them to buy does not solve the current demand. It ignores the market that already exists. Another response is to let lessors control everything and treat lessees as ordinary customers. That works for small assignments but not for independent networks using leased prefixes as part of their own service platform.
The better response is structured delegation. Contracts can define what the lessee may request and what service level the lessor must meet. Public records can show operational contacts. RPKI arrangements can be automated or pre-authorized where possible. Reverse-DNS delegation can be explicit. Abuse escalation can have timelines. Termination can include notice unless urgent abuse or legal compulsion exists. Return procedures can protect both route hygiene and customer migration.
APNIC can encourage this structure by treating route-security and contact continuity as part of responsible resource management. It does not need to take sides in commercial disputes. It can state that a holder allowing third-party operational use should maintain public data and security authorizations in a way that reflects real responsibility. If a holder wants the yield from leasing, it should also bear the duty to keep the public layer truthful.
That is the institutional bargain. Leasing can separate ownership-like economics from use. It should not separate income from accountability, control from contactability, or route authority from operational responsibility.
A more intelligible leasing market
APNIC's post-exhaustion challenge is to avoid two errors at once. The first is denial: treating leasing as peripheral, private and outside the registry's concern until something breaks. The second is overreach: using leasing risk as a reason to expand registry discretion into commercial life.
Denial is attractive because it preserves old categories: holders, allocations, assignments, transfers and policy-defined records. Private contracts are not the registry's business. That view is clean but incomplete. If leasing becomes a material way that scarce IPv4 reaches operators, it affects contact accuracy, abuse response, routing security, provenance and continuity. Those are registry concerns.
Overreach is attractive because every risk can sound like a reason for control. Abuse can justify business-model review, reputation can justify price suspicion, cross-border risk can justify discretionary approval, and scarcity can justify moral ranking of users. Soon the registry is no longer maintaining a ledger. It is deciding which commercial arrangements deserve to exist.
APNIC should choose the restrained middle. The test should be: does the intervention protect the accuracy, security and usability of the public coordination layer? If yes, it is likely legitimate. Does it instead substitute APNIC's view of price, capital allocation or moral deservingness for the parties' agreement? If yes, it is likely overreach.
This line is especially important in the Asia-Pacific because the incidence of friction is uneven. A rule that appears simple to a large operator in a mature market can be costly to a small operator in a less liquid economy. A disclosure requirement that is easy for a repeat broker may be hard for an NIR-linked local provider. A lease review designed to catch abuse may become a barrier to networks whose only problem is lack of paperwork fluency. If APNIC expands control, the burden will not fall evenly.
The registry's legitimacy comes from being cheaper and more reliable than private chaos. It should make it easier to know who is responsible, easier to secure routes, easier to maintain reverse DNS, easier to contact the operator, easier to identify forged use, and easier to preserve continuity through disputes. It should not make it harder for legitimate operators to obtain address usability because the institution is uncomfortable with the market's existence.
Leasing is also a policy signal. If operators lease because purchase is too expensive, that is a capital-market fact. If they lease because transfers are slow or uncertain, that is a registry-process fact. If they lease because IPv6 substitution is incomplete, that is an operational fact. If they lease because NIR or cross-border paths are asymmetric, that is a regional-design fact. If they lease because brokers package hidden supply, that is an information-market fact. Suppressing leasing would suppress the signal. Illuminating it lets APNIC see where the formal system fails to meet demand.
A practical agenda follows from that posture. APNIC should clarify the public responsibilities of holders who lease address space. The holder should remain accountable for accurate registry data, reachable contacts, authorized route origination, reverse-DNS arrangements where relevant, and abuse escalation. This does not mean the holder must police every packet. It means the holder cannot use private contracts to make responsibility unreachable.
APNIC should support layered operational contacts. A leased prefix may need a holder contact and an operator contact. The public record should make it possible for abuse desks, upstreams and counterparties to reach the party that can act. The design can protect privacy and avoid exposing every downstream customer, but it should not leave the internet guessing.
RPKI continuity should be treated as a leasing issue. If a holder authorizes another ASN to originate leased space, the authorization should be accurate, maintained and terminated predictably. Guidance should cover renewals, origin changes, emergency revocation, dispute handling and return. Invalid routes caused by administrative mismatch are not merely private inconvenience; they are public routing-security failures.
Reverse DNS and route-object management should also be part of responsible temporary use. The registry need not inspect every hostname or routing policy. It should make clear that leases require a maintained delegation path and cleanup obligations. Stale reverse DNS and stale route objects are small defects until they create operational confusion at scale.
APNIC should distinguish disclosure of responsibility from disclosure of commercial terms. Many market participants will resist any system they fear will expose prices, margins, customers or strategy. APNIC does not need those details for ordinary registry purposes. It needs enough to preserve contactability, authorization and accountability. Keeping that boundary clear will increase cooperation.
Practical guidance for small operators would help: how to confirm holder authority, check RPKI arrangements, verify reverse-DNS delegation, inspect reputation, define abuse duties, document permitted origin ASNs, understand termination risk, require notice periods and avoid subleasing chains that make responsibility unclear. Aggregate signals would help as well: contact failures, ROA delays, stale records, invalid-route incidents and dispute categories that threaten customer continuity. Enforcement should stay narrow. Fraudulent claims, forged authority, hijacking, unreachable contacts, persistent abuse non-response and misleading public records are registry concerns. High prices, profitable leasing, hosting use, cross-border demand, broker involvement and working-capital motives are not violations by themselves.
This agenda would make leasing less shadowy without making APNIC a leasing regulator. It would also improve transfer markets because the same facts that make a lease safe - provenance, contactability, route authorization, reputation history and operational continuity - make a future transfer safer. Formal transfers would remain the path for permanent recognized control. Leasing would remain a flexible mechanism for temporary or finance-driven use. Downstream assignments would remain ordinary customer service. The registry would not confuse those layers, but it would prevent public responsibility from vanishing between them.
The institutional line
The economics of IPv4 leasing in the APNIC region are not a story about opportunists exploiting a loophole. They are a story about scarcity meeting diversity. Registry allocation logic, transfer rules, working-capital constraints, operational deadlines, NIR paths, cross-border contracts, broker knowledge, reputation risk and incomplete IPv6 substitution do not line up neatly. Leasing is the market's way of making capacity move when the formal system is too slow, too expensive, too lumpy or too final for the need at hand.
That market can be useful. It can keep customers online. It can help small operators compete. It can release idle capacity. It can finance inventory. It can let holders earn yield without selling. It can give networks time to plan. It can reduce waste in a region where demand is uneven and fast-moving.
It can also be dangerous. It can separate economic benefit from accountability. It can leave lessees dependent on lessor-controlled registry levers. It can hide the real operator from abuse desks. It can make RPKI, reverse DNS and route objects fragile. It can turn reputation into an unrecorded liability. It can give brokers informational rents. It can create off-ledger allocations whose operational reality is invisible until a dispute, outage or abuse event exposes it.
The institutional line should be drawn with precision. APNIC should not deny the market. Denial makes the market darker. APNIC should not moralize the market. Moralizing turns a registry into a commercial authority it has no mandate to be. APNIC should not regulate price, lease term, margin, business model or capital structure except where a narrow legal or registry-integrity issue is genuinely present.
APNIC should instead protect the public layer: uniqueness, accurate records, contactability, route authorization, reverse-DNS continuity, provenance, fraud resistance and operational stability. It should make temporary use legible enough that the internet can see who is responsible without reading private contracts. It should keep the holder accountable while making the operator reachable. It should let the market price scarcity while preventing responsibility from disappearing.
That is a modest role, but not a small one. In a post-exhaustion IPv4 economy, the registry's value is no longer measured only by how it allocates new space. It is measured by how cheaply and reliably it lets existing space remain usable, traceable and safe as economic control changes. Leasing is one of the places where that measure becomes visible.
If APNIC gets the line right, leasing can be a legitimate capacity market attached to a trustworthy public ledger. If it gets the line wrong, leasing will still exist, but more of it will move through private contracts, opaque intermediaries, stale contacts and fragile route dependencies. The registry will then be formally correct and economically blind.
The better APNIC posture is simple: do not choose the price, do not choose the winner, do not pretend the market is not there. Make the responsibility surface visible. Preserve the ledger. Keep the addresses reachable. Let scarce capacity move without letting accountability disappear.