Summary

  • A sale-and-leaseback has at least three separate layers: a commercial transfer, an RIR registration event and a lease of use back to the former holder. A valid registry update proves an administrative change; it does not alone determine accounting derecognition, tax ownership or secured-credit classification.
  • The market evidence is real but limited. LARUS publicly offers a model in which a holder sells IPv4 addresses and leases back needed capacity, while Cogent's public filings show a different route to liquidity through notes secured by addresses, customer leases and receivables. No public dataset supplies the number, volume or performance of completed IPv4 sale-and-leasebacks.
  • International accounting starts with control. IFRS 16 sends the sale leg to the IFRS 15 transfer test; if the transfer is not a sale, the seller keeps the asset and recognises a financial liability. Above-market sale proceeds can also be treated as additional financing.
  • Lease term and repurchase rights are not decorative. A lease covering nearly all economically useful life, an obligatory renewal, a fixed-price return, a nominal purchase option or a seller right that prevents the buyer from directing disposition can make the transaction behave like secured funding.
  • A genuine structure leaves the buyer with meaningful residual exposure and decision rights while giving the seller a defined, time-bounded right to use identified prefixes. A sham merely circulates cash and paperwork while the original operator keeps the benefits, burdens and inevitable path back to the asset.
  • Number Resource Society should publish a neutral evidence and classification framework, not a moral approval list. Its contribution is to make control, term, options, cash flows and registry events comparable while preserving private terms and jurisdiction-specific legal judgment.

One transaction, four different questions

The phrase sale-and-leaseback makes a single event sound settled. It is actually a bundle of questions answered by different institutions for different purposes.

The first question is contractual: what did the seller promise to transfer, and what did the buyer agree to provide? The answer lies in the purchase agreement, the lease, related options, guarantees, service arrangements and the circumstances in which they were negotiated.

The second is administrative: which organisation does the relevant Regional Internet Registry recognise as the holder after closing? ARIN, RIPE NCC, APNIC and their counterparts process number-resource transfers under regional policies and agreements. Their records are essential to reliable coordination. A purported buyer that never becomes recognised may not have received the practical position the transaction assumed.

The third is financial reporting: did the buyer obtain control of the asset for purposes of the applicable accounting framework, and did the seller retain only a right of use? IFRS 16 directs the parties to the IFRS 15 sale test. United States GAAP under Topic 842 and Topic 606 also focuses on transfer of control, with specific treatment for finance leasebacks and repurchase options.

The fourth is legal or tax characterisation: does the transaction create a true sale and lease, or does it function as secured financing? The governing law may look at rights, risks, term, termination and repurchase. United States tax doctrine examines objective economic realities and the benefits and burdens retained. Secured-transactions law can recharacterise some leases, although rules written for goods cannot simply be pasted onto an intangible registration relationship.

These questions can produce answers that are related but not identical. A registry can correctly update its holder record even if an auditor concludes that the seller did not surrender enough control for accounting derecognition. A tax authority can call an arrangement financing without directing the RIR to reverse a registration. A lender can take a security interest in contractual and economic rights even where the registry refuses property language.

The analytical error is to let one answer swallow the others. Registry staff are not the global accounting board. Auditors do not originate BGP announcements. A tax label does not update RDAP. The purchase agreement cannot force a registry to ignore its policy. A useful test keeps the layers connected but separate.

The market has offered the structure, but the denominator is missing

Sale-and-leaseback is no longer merely a thought experiment in IPv4 marketing. LARUS publicly describes an offering in which an organisation sells IPv4 addresses to LARUS and leases back the capacity it needs. The page presents immediate monetisation, continued operational use and movement of direct registry exposure to a specialist first-party operator as the commercial rationale.

That is first-party evidence that a market entity offers the structure. It is not an audited transaction register. The page does not publish a complete executed agreement, the number or value of completed transactions, the range of lease terms, repurchase provisions, defaults, registry outcomes or customer continuity results. Its claims about resilience and risk transfer are commercial claims that need transaction-level testing.

Another public financing route provides a useful comparator. Cogent Communications disclosed two series of notes issued by a bankruptcy-remote IPv4 subsidiary in 2024 and 2025. The collateral included IPv4 addresses, customer leases, receivables and related assets. The structure used reserves, coverage and utilisation triggers, defective-transfer indemnification and substitution rights. It raises liquidity from IPv4 value and cash flows without being described as a sale-and-leaseback between the operating user and an outside buyer-lessor.

The comparison matters because it reveals choice. An operator seeking cash can sell and cease using addresses, borrow directly, securitise lease cash flows, contribute assets to a controlled financing vehicle, lease unused space to customers, or sell to an investor and lease back capacity. Those forms place control, residual value, registry exposure, balance-sheet treatment and operational continuity differently.

No public source reviewed for this article provides a census of IPv4 sale-and-leasebacks. RIR transfer logs do not disclose whether a registered transfer includes a leaseback. Private leases are often absent from public registration data. Financial statements may aggregate number resources with other intangibles or describe financing without publishing prefix-level terms. Market providers do not publish every unsuccessful negotiation.

Therefore no global rate appears here. There is no claim that sale-and-leaseback represents a particular percentage of transfers, that most such arrangements are financing, or that a stated share is abusive. The evidence supports a classification test and the existence of an offered structure, not a population estimate.

This absence is itself a governance fact. Moral claims about prevalence run ahead of the data. The correct response is not compulsory disclosure of every private contract. It is a neutral framework that lets parties, auditors, lenders and courts identify the features that matter when classification is required.

Start by identifying exactly what was sold

A useful analysis begins before labels. It identifies the subject of the transfer.

An IPv4 transaction may cover the registered position for specified prefixes, contractual rights under a registry agreement, related database records, RPKI administration, reverse-DNS authority, route objects, letters of authorisation, customer leases, receivables, reputation history and operational services. These elements do not always move together.

The purchase agreement should list the prefixes and the current recognised holder. It should state which rights and obligations are intended to transfer, what remains with the seller, and which effects depend on registry action. It should identify existing leases, customer uses and route authorisations. It should explain whether the buyer will become the direct holder, use a subsidiary, appoint the seller as servicer, or rely on another registration arrangement.

The leaseback then defines the right returned to the seller. Is it the exclusive use of the exact prefixes? A quantity of addresses that the buyer may substitute? A managed network service? A right to originate routes from the seller's ASN? A right to serve existing customers? The more fungible the promised capacity, the less the seller necessarily controls the particular transferred block. The more exact and irreplaceable the prefixes, the more the seller may have retained the practical core.

Services must be separated from use rights. The buyer may maintain registry records while the seller operates the network. The seller may provide billing, abuse handling and routing services to the buyer. Those arrangements can be commercially sensible, but if the buyer cannot direct the important decisions or replace the servicer, the claimed transfer deserves closer examination.

Proceeds matter too. A true sale price is consideration for rights transferred. Rent is consideration for use returned. A deposit, prepaid rent, guarantee or reserve changes the buyer's exposure. If nearly all sale proceeds are locked to fund lease payments and repurchase, cash may have circulated without creating meaningful new risk for the buyer.

The registry record is one component of this map. It can show that the recognised holder changed. It does not reveal the lease term, call option, put option, residual economics, side letters or who directs the addresses day to day. Classification fails when a one-line registration is asked to carry a hundred-page contract.

Control is the centre of the accounting test

IFRS 16 does not accept the title "sale-and-leaseback" as an accounting conclusion. Paragraph 99 requires the entity to apply IFRS 15's requirements for determining when a performance obligation is satisfied in order to decide whether the transfer is a sale.

IFRS 15 describes control as the ability to direct the use of an asset and obtain substantially all remaining benefits, including the ability to prevent others from directing use and obtaining benefits. Indicators include a present right to payment, legal title, physical possession, significant risks and rewards, and customer acceptance. No single indicator automatically decides every case.

Physical possession is a poor fit for IPv4, but the control logic is valuable. Who can decide whether to sell the block, lease it to another user after the leaseback ends, split it, substitute it, change service providers or bear a loss if market value falls? Who obtains lease revenue? Who can prevent unauthorised use? Who benefits from appreciation? Who carries the risk that registry recognition, reputation or demand weakens?

If the transfer qualifies as a sale, IFRS 16 requires the seller-lessee to recognise a right-of-use asset for the retained use and only the gain or loss relating to rights transferred. The buyer-lessor accounts for the purchase under applicable standards and the lease under lessor rules. If consideration or rent is off market, below-market terms are treated as prepaid lease payments and above-market terms as additional financing.

If the transfer does not qualify as a sale, the accounting is more direct. The seller-lessee continues to recognise the transferred asset and recognises a financial liability equal to the proceeds. The buyer-lessor does not recognise the transferred asset and recognises a financial asset equal to the proceeds. The paperwork may say sale, but the accounts describe funding.

Applying this to IPv4 requires care about the asset recognised in the first place. Entities may have different accounting histories for acquired and legacy address space. The article does not assume every holder carries a separately recognised IPv4 intangible at a stated value. The sale-and-leaseback test still matters to the extent the applicable framework identifies an asset and a transfer.

The lesson is institutional rather than mechanical: accounting asks who obtained control and what rights remained. An RIR update is strong evidence of a changed recognised relationship, but it cannot answer the complete control inquiry.

Repurchase rights can turn exit into an illusion

A repurchase clause is often the fastest way to understand the transaction.

If the seller must buy the addresses back, or can do so for nominal consideration, the buyer may never have obtained a meaningful residual position. If the buyer must return them at a fixed price that repays its advance plus an agreed yield, the cash flows resemble principal and interest. If the seller has a call option so valuable that exercise is economically inevitable, the buyer's apparent upside may be only a financing return.

IFRS 15's repurchase guidance examines forwards, call options and put options when deciding whether control transferred. A seller's obligation or right to repurchase can prevent the customer from obtaining control because the customer's ability to direct use and obtain substantially all benefits is constrained. The detailed result depends on the type of right, price and incentives.

United States GAAP provides a useful comparator. FASB materials explaining Topic 842 state that a seller-lessee repurchase option generally precludes sale accounting unless the exercise price is fair value at exercise and substantially the same alternative assets are readily available in the marketplace. A finance leaseback or sales-type lease also prevents the buyer-lessor from being treated as having obtained control for the sale test.

IPv4 makes the alternative-asset question difficult. Two /20 blocks contain the same number of addresses but may differ by registry region, transfer status, route reputation, aggregation, geolocation history, customer embedding, RPKI state and operational acceptance. A contract that says the seller may repurchase the exact original prefixes is not the same as a right to acquire equivalent capacity at then-current fair value.

Options are not automatically abusive. A fair-value right of first offer may protect commercial continuity without making return inevitable. A narrowly defined purchase right after buyer default can be a remedy. A seller put at market value may address an investor exit. The analysis asks who controls the option, how price is set, whether exercise is compelled, and what genuine alternatives exist.

Side letters count. A lease may appear terminable while a separate service agreement, guarantee or customer commitment makes termination economically impossible. A repurchase option can sit in an affiliate rather than the seller. Classification follows the integrated bargain, not the document title.

A credible transaction therefore discloses every call, put, right of first refusal, mandatory purchase, renewal and substitution mechanism to the parties responsible for legal and accounting analysis. It also records the valuation method at exercise. Hidden repurchase economics are the shortest route from a defensible sale to an undisclosed loan.

Term measures how much economic life the seller kept

The lease term is not merely a billing interval. It allocates economic life.

A one-year leaseback of part of a portfolio, with the buyer free to redeploy the addresses afterward, looks different from a twenty-year exclusive lease of the exact block with automatic renewals. The first leaves a visible residual opportunity with the buyer. The second may give the seller almost everything that mattered while the buyer receives a scheduled return.

IPv4 has no physical useful life in the ordinary sense. Technical standards do not prescribe a depreciation date after which the addresses wear out. Economic life depends on continued demand, IPv6 substitution, policy, routing acceptance, regulation, customer systems and the cost of renumbering. That uncertainty makes term analysis more important, not less.

The parties should estimate the period over which the block is expected to produce material benefits under stated scenarios. They should compare that horizon with the non-cancellable lease, renewal options the seller is reasonably certain to exercise, purchase obligations and termination penalties. A lease that captures nearly all expected benefits can resemble financing even if the calendar term sounds finite.

Termination rights must be real. If the seller can terminate but only by paying the investor the remaining rent, expected residual value and a premium, the obligation is economically close to non-cancellable. If the buyer can terminate and reallocate the addresses upon seller default, it carries a genuine residual and remarketing task. If registry policy or customer dependence makes repossession practically impossible, the written right may be weaker than it appears.

Substitution changes the analysis. A buyer allowed to replace the exact prefixes with equivalent addresses may control a pool rather than one block. The seller receives capacity; the investor manages assets. But equivalence must preserve routing and service needs. A purely paper substitution right that no customer could tolerate does not establish practical control.

Rent structure provides another signal. Fixed payments designed to return the sale price plus a yield can look like debt service. Variable rent tied to address utilisation or market rates gives the buyer different exposure. Neither form is conclusive. IFRS 16 expressly handles variable sale-and-leaseback payments, and commercial leases can be fixed. The point is to examine the complete risk pattern.

There is no universal number of years that transforms an IPv4 lease into financing. The relevant denominator - expected economic life - is itself uncertain and transaction-specific. Any bright-line global term would be invented certainty.

Residual value reveals whether the buyer is an investor or a conduit

A genuine buyer-lessor should possess something meaningful when the leaseback ends.

That residual may be the right to renew at market rent, lease to another operator, sell the prefixes, divide or combine them, use them in its own services, or accept the loss if demand and reputation deteriorate. It may also benefit if scarcity value rises. The buyer need not operate the addresses day to day, but it should carry a real economic position beyond receiving predetermined payments.

If every plausible outcome returns the block or its full value to the seller, the buyer may be a funding conduit. Warning signs include a fixed repurchase equal to original price plus yield, a lease term covering the expected economic horizon, sale proceeds reserved to meet rent, seller guarantees of residual value, and restrictions preventing the buyer from dealing with the asset even after default.

Registry risk complicates residual value. A buyer may become the recognised holder and assume direct exposure to fees, policy compliance, transfer procedure and account continuity. That is a real burden. But it is not enough by itself if the seller indemnifies every outcome and controls every decision. Conversely, using a specialist to manage that exposure can be a genuine service even where the seller retains use.

Reputation risk is also residual. A lessee's abuse, spam or poor network hygiene can impair future value. A buyer that bears the post-lease cost has an investor's concern and should have monitoring, acceptable-use, remediation and security rights. A seller that guarantees a pristine exit has retained more risk than ordinary rent suggests.

The buyer's capital structure matters. If it borrows the purchase price on a non-recourse basis and lease payments exactly service that debt, it may still be a genuine investor, but the circularity deserves scrutiny. Who loses if the seller defaults? Who gains if market value rises? Who must find the next user? Economic substance lies in those tails.

An independent valuation should therefore include residual scenarios rather than only current sale price. It should state assumptions for lease expiry, route reputation, policy, IPv6 adoption, transfer time and remarketing cost. The article does not offer a residual percentage because no public dataset supports one.

The buyer need not be heroic. It only needs to be more than a name temporarily inserted into a registry while all meaningful benefits and burdens remain with the seller.

The registry should test registration, not moral purpose

RIR transfer policies ask administrative questions. Is the source the recognised holder? Is the resource eligible? Has the source supplied authority evidence? Does the recipient satisfy applicable requirements? Are required agreements, fees and inter-RIR coordination complete? When those conditions are met, the RIR updates its registration according to policy.

That function is important. It does not include deciding whether the parties should account for the transaction as sale, lease or financing under every jurisdiction. Nor does it require staff to judge whether raising capital through a leaseback is morally superior to a secured note.

ARIN's materials show that specified-recipient transfers have source and recipient requirements, that fees do not guarantee approval, and that completion follows approval, agreement and applicable coordination. RIPE policy says legitimate holders may transfer eligible resources and that the RIPE NCC completes a transfer by updating records. APNIC publishes its own source and recipient criteria. These are institutional conditions, not a global economic-substance code.

A registry can legitimately investigate false authority, conflicting claims, ineligible space or non-compliance with a specific policy. It should state the rule, evidence and decision. It should not deny an otherwise valid transfer because a leaseback resembles finance, nor approve it as a tax shelter because the database changed.

Moral labelling is especially dangerous because economically similar transactions can use different forms. A secured loan may leave the operator as holder while giving a lender extensive remedies. A sale-and-leaseback may move the holder record while leaving the seller operational use. A securitisation may place addresses and leases in a special-purpose entity. If registry policy privileges one label without analysing its own relevant facts, parties will optimise paperwork rather than risk.

The correct boundary also protects critics. A completed registration is not immunity from tax, accounting, creditor or insolvency law. If a court later finds fraud or a sham, the registry may respond to a lawful order through a defined correction procedure. Administrative recognition is not absolution.

Registry neutrality therefore means disciplined scope, not blindness. Verify identity and authority. Apply published eligibility rules. Record the holder accurately. Preserve a review trail. Refer legal classification to the competent forum. The RIR becomes more reliable by refusing to invent a finance ministry inside a transfer desk.

Economic substance distinguishes liquidity from evasion

Sale-and-leaseback can serve legitimate purposes. An operator may have valuable addresses but need capital for network expansion. It may want predictable use without maintaining direct exposure to every registry process. An investor may be willing to hold residual value and provide administration. The structure can align those preferences.

It can also be abused. Parties may try to manufacture sale proceeds while preserving all ownership economics, avoid a lender's covenant, move assets beyond creditors, disguise secured debt, obtain accounting gains, shift tax benefits without risk, or evade a transfer restriction. The same surface form can carry either substance.

United States tax authority illustrates the principle. IRS Notice 2005-13, drawing on Frank Lyon and other cases, says objective economic realities govern rather than the parties' labels. The notice targets sale-leaseback structures in which payment obligations were economically defeased and the purported buyer's downside and upside were limited. It asks whether the owner-lessor acquired significant and genuine attributes of ownership, including benefits and burdens.

That notice concerns United States federal tax and particular listed transactions involving tangible property and tax-indifferent parties. It does not classify IPv4 globally. Its value here is the method: examine risk, control, cash circularity, business purpose and residual economics.

A credible IPv4 sale-and-leaseback has a non-tax commercial explanation, independently valued consideration, a real registration transfer, a defined lease, buyer control outside the retained use, and meaningful buyer exposure. An evasive structure hides side rights, fixes every outcome, defeats creditors, misstates registry facts or uses the buyer as a nominee.

Intent alone is limited public evidence. A party can sincerely call a deal a sale while writing debt economics. A critic can sincerely dislike IPv4 markets while misclassifying a genuine transfer. Objective terms are more reliable than motive narratives.

The integrated transaction must be reviewed at closing and through performance. Later amendments can change substance. An optional repurchase can become mandatory. A short lease can be repeatedly extended. The buyer can surrender control to the seller through servicing practice. Economic classification is not protected by the original press release.

This is why the test should be evidence-led and jurisdiction-specific. It catches evasion without giving an RIR an unbounded moral veto.

The UCC analogy is useful only inside its boundary

UCC Section 1-203 distinguishes a lease of goods from a security interest by facts of each case. It provides a bright-line route to a security interest when the lessee's payment obligation is not terminable and the term covers the remaining economic life, the lessee must renew or become owner, or the lessee has a nominal renewal or purchase option. It also says that full-value payments, risk of loss, taxes, maintenance and a fair-market option do not alone create a security interest.

This is an elegant map of term and residual economics. It is not a direct legal answer for IPv4 because the section addresses goods, while number-resource interests are intangible and embedded in registry contracts. The governing law may classify those interests differently. A court may instead apply Article 9, common law, insolvency law, tax doctrine or another jurisdiction's rules.

The analogy still sharpens questions. Is the seller-lessee's obligation non-cancellable? Does the term consume the expected economic life? Is return mandatory? Is the purchase price nominal or fair value at exercise? Does the buyer retain a meaningful alternative use? These are the same economic joints that IPv4 documents can expose.

If a transaction is recharacterised as secured financing in a United States jurisdiction, attachment and perfection become central. UCC Article 9 generally requires the debtor to have rights in collateral or power to transfer them and an authenticated security agreement or another specified condition. A financing statement may be necessary but not sufficient to obtain registry recognition after enforcement.

Parties should therefore draft for both characterisations. The purchase agreement can include a back-up security grant to the extent permitted, without asserting that it conclusively solves perfection. The buyer can file where counsel advises. The documents can preserve proceeds, records, cooperation and substitution remedies if a court treats the sale as financing.

Back-up drafting is not an admission of sham. It is recognition that classification can differ by purpose and jurisdiction. Sophisticated transactions anticipate an adverse legal view instead of relying on typography.

The boundary must remain visible in public analysis. This article uses Section 1-203 as a disciplined comparator, not as proof that every long IPv4 lease is debt or that United States model law governs a RIPE or APNIC transaction.

A genuine structure leaves the buyer room to act

Consider a defensible case.

An operator holds several prefixes and uses only part of them. It sells specified blocks at an independently supported price to an unaffiliated investor. The relevant RIR recognises the buyer through the applicable transfer process. The operator leases back a defined subset for five years at a market-supported rent. The buyer may substitute equivalent capacity subject to service protections, lease unused portions to others, and sell the residual interest subject to the lease. The seller has no fixed-price call. At expiry it may bid for renewal at market terms, but the buyer can choose another user.

The buyer bears market decline, policy exposure and remarketing cost, while the seller remains responsible for its traffic and contractual compliance.

This arrangement can release capital and transfer residual risk. Continued use does not negate sale; leaseback exists precisely because the seller needs use after disposition. Registry recognition, accounting and law must still be analysed, but the buyer is not a mere nameplate.

Now consider a financing-shaped case.

The seller receives cash and leases the exact block for a non-cancellable term expected to cover its useful economic horizon. Rent equals the buyer's funding cost plus yield. The seller guarantees residual value, pays all costs, controls every route and registry decision, and must repurchase at original price. The buyer cannot lease, sell, substitute or use the addresses. Default leads to a payment claim but no practical right to remarket.

That bargain behaves like a loan secured by the block and its cash flows. Calling the investor "owner" and updating the registry may not change the economic result for accounting or tax.

A third case is evasive.

The purported buyer is affiliated or under side control, no real consideration remains at risk, authority documents are misleading, and the purpose is to defeat creditors or bypass a specific transfer rule. The seller keeps secret return rights. Here the issue is not merely financing classification but possible fraud or policy breach.

A fourth case transfers institutional risk to a specialist. The seller genuinely gives up the holder position but pays for time-bounded continued capacity and expert administration. This can be a real service if the specialist controls the holder role and residual decisions. It becomes cosmetic if every institutional consequence is passed back to the seller through indemnity and command rights.

The test handles all four without moral shortcuts.

Contract terms should expose rather than conceal the classification

A serious sale-and-leaseback agreement should make its economics readable.

The sale schedule identifies exact prefixes, price, allocation of transaction cost, existing uses and the registry event required for closing. A fair-value memorandum explains methods and assumptions without claiming a global price. Any difference between consideration and market indication is allocated to prepaid rent, financing, services or another identified component.

The lease schedule identifies the resource or capacity, exclusivity, permitted origins, geography if relevant, service level, abuse obligations, RPKI and route administration, reverse DNS, fees, term, renewal and termination. It says who may substitute addresses and under what equivalence standard.

The control schedule lists decisions reserved to the buyer, decisions delegated to the seller as operator or servicer, and decisions requiring consent. It covers sale, further lease, split, aggregation, RIR account action, RPKI, material route changes and enforcement. Delegation should be revocable where genuine buyer control requires it.

The option schedule consolidates every call, put, right of first offer, renewal, mandatory purchase, residual guarantee and affiliate right. Exercise price, valuation process and dispute mechanism are explicit. There should be no option hidden in a service fee or side letter.

The risk schedule allocates registry fees, policy change, transfer delay, route reputation, customer claims, abuse, sanctions, taxes, insurance, cyber incidents and force majeure. Allocation does not decide accounting by itself, but it reveals who bears benefits and burdens.

The default schedule distinguishes payment failure, operational misuse, registry impairment, false representation and insolvency. Remedies include cure, suspension, substitution, cash collateral, customer migration, termination and sale. A remedy should be operable without relying on a party that is already hostile.

The reporting schedule requires enough evidence for classification and continuity while protecting customer confidentiality. Registry records, route state, utilisation, lease revenue, disputes and options are reviewed periodically.

The accounting and tax schedule records each party's intended treatment but acknowledges that competent authorities and auditors apply governing standards. It requires cooperation if treatment changes. It does not ask the RIR to endorse the conclusion.

Transparent terms do not guarantee a true sale. They make the conclusion testable.

Operational control must be separated from registered control

The seller often remains the network operator. That fact is commercially logical and analytically dangerous.

The operator may originate the prefixes, configure services, answer abuse complaints and manage customers. The buyer may lack a network. If operation alone meant ownership, no ordinary lease of productive infrastructure could qualify as a lease. The question is whether the operator acts within a defined right of use while the buyer retains higher-level control and residual decisions.

Expected origin ASNs should be recorded. RPKI authority should match the lease and permit a safe transition at expiry. Internet Routing Registry entities and letters of authorisation need time bounds. Reverse-DNS control should be allocated. The buyer needs monitoring and emergency rights; the seller needs protection against arbitrary technical interference during a paid-up term.

Substitution is revealing. If the buyer can replace a prefix with operationally equivalent capacity after notice, it exercises asset management. If the seller can veto every substitute for reasons unrelated to service, it may control the exact asset. Yet strict equivalence can be legitimate where customers have embedded addresses in allowlists, certificates, firewalls or partner systems. Renumbering cost is an economic fact, not proof of sham.

End-of-term planning should begin before the final day. The parties decide whether routes stop, ROAs change, reverse DNS moves, customers renumber or a renewal occurs. A buyer with no practical ability to recover use after expiry may hold only a payment claim.

Abuse handling also allocates risk. The seller controls traffic and should carry primary contractual responsibility for its use. The buyer bears residual reputation and registration exposure and therefore needs audit, notice and termination rights. Overbroad buyer control of every packet would undermine the seller's operation; no control would make residual ownership hollow.

Observed BGP does not settle the bargain. A route origin can be the seller, transit provider, mitigation service or managed host. Registry data does not settle operation either. The contract and technical evidence must be read together.

This separation supports legitimate leasing. It also makes hidden retention visible. Registered control, operational control, economic control and incident control can sit with different parties, but the allocation must be coherent.

Insolvency is where the label becomes expensive

If the seller-lessee fails, the buyer will claim that the addresses belong outside the estate and that it may terminate the lease or find another user. Creditors may argue that the transaction was secured financing and that the addresses or rights remain estate property subject to insolvency priorities. If the buyer fails, the seller will want continued use and protection from a sale to an incompatible purchaser.

The answer depends on governing law, transaction substance, perfection, registry recognition, lease terms and court orders. A contractual label is evidence but not final.

The parties should prepare for both insolvencies. If treated as a true sale, the buyer needs a clean authority chain and evidence that consideration was paid. The lease should address assumption, assignment, cure and continued service to the extent law permits. Customer dependencies need a transition plan.

If treated as financing, the buyer-lender needs an enforceable security interest, proper filings where applicable, proceeds protection and a path to seek registry recognition after a lawful enforcement disposition. A back-up security grant and cooperation covenant can reduce the gap, though they cannot override insolvency law or registry procedure.

Commingling is dangerous. Sale proceeds, rent, customer receivables and registry fees should be traceable. A special-purpose vehicle can isolate assets, but control by the seller and inadequate separateness can weaken the intended result. Independent governance should be substantive, not a name on stationery.

The leaseback should also survive buyer distress long enough to avoid abrupt outages, subject to law and payment. Escrowed operational records, alternate contacts, route-transition procedures and a cure period can protect users without guaranteeing the seller perpetual possession.

Registry staff should respond to authenticated court and party evidence under published procedures. They should not decide insolvency priority from a press account. A status hold may preserve the record while the competent forum acts, but indefinite administrative limbo can destroy value for every claimant.

Insolvency is the best reason to classify honestly at closing. The parties cannot wait until cash is scarce and incentives diverge to discover whether they sold an asset, granted security or created an unworkable hybrid.

Neither "property" nor "not property" completes the analysis

The IPv4 debate often collapses into a word.

RIPE NCC's Standard Service Agreement says registration does not constitute property or confer ownership. ARIN has also historically used contractual language denying that number resources are property. Market entities, courts and accounting records may nevertheless recognise valuable, transferable bundles of rights. The commercial fact of value does not force every jurisdiction to call the resource property; the registry's vocabulary does not erase the economic bundle.

Sale-and-leaseback makes this tension visible. If nothing of value can be transferred, there is no sale. Yet RIRs administer specified transfer processes, buyers pay consideration, operators lease use, and Cogent has disclosed financing secured by addresses and related cash flows. The market is not imaginary.

The accurate entity is often a bundle: recognised registration, contractual service rights, practical control, transfer eligibility, routing-related authorities, customer use, revenue and residual opportunity. Different laws attach different consequences to each strand.

A sale can transfer the buyer's position in that bundle subject to registry rules and retained lease rights. A financing can encumber the seller's interest. A court can analyse value without deciding that addresses are ordinary chattels. An auditor can test control without settling constitutional property theory.

This middle language is less satisfying to advocates because it resists a grand victory. It is more useful to contracts. Parties can warrant specific rights, condition closing on specific records, allocate specific risks and design specific remedies.

The registry should not weaponise "not property" to pronounce a lawful commercial arrangement invalid when policy permits the transfer. Investors should not weaponise "asset" to promise seizure beyond the rights actually acquired. Precision protects both coordination and capital.

The sale-and-leaseback test therefore asks what moved, what stayed, who controls decisions, who bears residual risk and what options reverse the movement. The answer can be rigorous without forcing one universal noun.

Evidence should be sufficient for assurance, not public surveillance

Classification requires evidence, but not every term belongs in a public database.

Auditors, tax advisers, lenders and courts may need the integrated agreements, valuations, option schedules, cash-flow model, authority evidence, registry receipts and servicing arrangements. The RIR needs the information required for its transfer and continuing registration functions. Operators need current contact and routing authority. The public may need accurate holder and abuse information.

Purchase price, rent, investor return, customer identity and tax analysis can remain confidential unless disclosure law requires otherwise. Publishing them as a condition of registration would expose strategy and discourage accurate filings. It would also tempt registry staff to judge matters outside their mandate.

A compact assurance statement can identify the classification process without revealing the bargain. It can say that independent advisers reviewed transfer control, lease term, options and residual risk as of a date; that no undisclosed mandatory repurchase was identified; and that the registry record changed. The statement must identify scope and cannot guarantee future treatment.

Evidence should be versioned. Amendments to rent, term, control or repurchase can change the analysis. The assurance process should trigger on material changes, not only at closing.

Parties should preserve the raw agreements and registry receipts. A hash-linked evidence index can prove which versions were reviewed. Access can be granted under confidentiality to a successor auditor, insolvency officeholder or authorised dispute forum.

Aggregated research can improve the market if denominators are honest. A study may report that, among a defined voluntary sample, a stated number had fixed-price calls or terms above a stated threshold. It must not extrapolate to all IPv4 transfers without a sampling basis.

Privacy and transparency are not opposites. The goal is transparent classification logic and verifiable evidence boundaries, not exposure of every commercial term. A system that collects less but defines it well will produce more trust than a compulsory warehouse of contracts no public administrator can competently interpret.

NRS can publish the test without becoming the judge

Number Resource Society can turn the principles into a common transaction profile.

The profile should record the facts that drive classification: prefixes, recognised holder before and after, closing event, identified leased resource, exclusivity, non-cancellable term, renewal assumptions, substitution rights, seller and buyer decision rights, calls, puts, mandatory repurchase, option pricing method, residual-value allocation, registry-risk allocation and servicing powers.

Each field should state whether it is public, shared with authorised reviewers or retained privately with a cryptographic commitment. The profile can produce a human-readable summary and a machine-readable evidence index. It should not publish rent or customer lists by default.

NRS can define several non-binding analytical outcomes: evidence supports a substantive sale and leaseback; evidence indicates financing characteristics; evidence is mixed and requires jurisdictional review; or evidence is incomplete. The outcome should include reasons, not a badge of virtue.

Independent accountants and lawyers remain responsible for formal conclusions. Courts remain responsible for disputes. RIRs remain responsible for registration. NRS should not approve the transfer, set tax treatment, certify perfection, rank investors or require its profile before an address can be routed.

It can also publish a back-testing protocol. Did the buyer exercise real control? Were substitutions possible? Did the seller repurchase? Who bore losses after abuse or market decline? Were renewal and residual outcomes consistent with the initial classification? Results from consenting entities can improve the framework.

NRS's own public materials discuss holder interests and monetisation through sale or lease. Those are first-party statements of mission and market view. They do not prove independence, adoption or the success of the proposed test. Credibility requires open governance, conflict disclosure, versioned criteria and competing implementations.

The positive case is strong because the alternative is worse. Today, parties can choose self-serving labels while critics choose self-serving moral categories. A common evidence vocabulary would make disagreement more precise without centralising permission.

NRS should be the keeper of the questions, not the owner of the answers.

A pilot should test documents against outcomes

A useful pilot would recruit a small voluntary set of structures: a direct sale-and-leaseback, an IPv4-backed loan, a securitisation, a portfolio sale with partial leaseback, and a managed capacity agreement that is not intended as a sale. The aim is comparison, not market census.

At closing, reviewers complete the common profile and record their classification with confidence and reasons. They preserve the evidence index. No private price or customer data is published without consent.

During the term, the pilot records material events: registry change, substitution, lease amendment, option grant, route incident, reputation impairment, payment default, renewal, repurchase and sale to a third party. It tests whether the initial control map predicted who actually made decisions and bore losses.

At expiry or after a defined period, the pilot compares expected and realised residual outcomes. Could the buyer remarket? Did the seller renew at market terms? Was a repurchase effectively inevitable? Did the registry relationship constrain a remedy? Did customers experience continuity?

Metrics should include completeness of evidence, time to classify, amendment detection, time to resolve conflicting records, substitution success and operational continuity. They should not treat a "true sale" outcome as inherently better than financing. The objective is accurate classification and resilient performance.

The report must preserve denominators. If six transactions participate, conclusions concern six transactions. If only market providers volunteer, selection bias is explicit. Failed negotiations can be included anonymously because they may reveal more about unusable terms than successful closings.

Conflicts require control. A provider promoting sale-and-leaseback should not alone certify its own deals. Seller, buyer, operator, accountant, insolvency expert and registry observer should have representation. Criteria and change logs should be public.

The pilot would not prove global legal treatment. It would show whether the evidence profile captures the decisions that matter and whether transaction labels predict outcomes. That is a practical contribution within NRS's competence.

What can be concluded, and what cannot

Several conclusions are supported.

First, IPv4 sale-and-leaseback is a publicly offered commercial structure. Second, IPv4 resources and lease cash flows have supported disclosed institutional financing. Third, RIR policies make registry recognition a material closing and continuity event. Fourth, established accounting and legal frameworks distinguish sale, lease and financing through control, term, repurchase and economic risk rather than contract title alone.

Several stronger claims are not supported.

There is no public global count of completed IPv4 sale-and-leasebacks. There is no denominator for the share recharacterised as financing, challenged by an RIR, impaired by reputation or tested in insolvency. There is no universal fair lease term, residual percentage, option price or advance rate. There is no evidence that one provider's public offering has transferred registry risk successfully in every completed case.

The IFRS and FASB rules are accounting standards, not RIR policies. The UCC rule cited concerns goods and is used as an analogy within stated limits. IRS Notice 2005-13 concerns United States federal tax and a particular abuse pattern, not the global IPv4 market. Registry agreements differ, and jurisdiction matters.

The article's scenarios are analytical applications, not reports of undisclosed transactions. The proposed NRS profile and pilot are recommendations, not deployed infrastructure. A positive institutional role is possible but unproven.

These limitations prevent moral overreach. They also prevent promotional overreach. A provider cannot cite the existence of accounting guidance as automatic validation of its product. A registry cannot cite missing market statistics as proof that every leaseback is evasive. An analyst cannot invent a prevalence rate from transfer logs that do not disclose leasebacks.

The disciplined conclusion is narrower and more durable: any IPv4 sale-and-leaseback can be tested by mapping transferred rights, retained use, control decisions, term, options, cash flows, residual exposure, registry events and remedies. Weak evidence produces a weak conclusion. Strong evidence makes classification possible.

Registration is evidence; economics is the verdict

An IPv4 operator may sensibly want cash without renumbering customers. An investor may sensibly want residual exposure to scarce address capacity. A specialist may be better equipped to manage the holder relationship. A sale-and-leaseback can connect those needs.

The structure does not become illegitimate because the seller continues to use the addresses. Continued use is the leaseback. Nor does it become a true sale simply because the RIR records the buyer. Registration is essential evidence that one institutional relationship changed; the complete transaction determines what else changed.

The buyer must receive more than a name in a database. It needs meaningful control outside the seller's defined use, exposure to residual value, and an operable remedy. The seller must receive less than an inevitable return of the same asset. Term, renewal and repurchase should leave genuine uncertainty where a sale is claimed. Cash should not merely travel in a circle designed to pay itself.

Accounting already has a language for this. If control did not transfer, keep the asset and recognise financing. If it did, recognise the retained right of use and only the gain on rights actually transferred. Treat above-market proceeds honestly. Tax and secured-transactions law add jurisdiction-specific tests. None requires a registry to become a moral tribunal.

RIRs should verify the facts within their mandate and publish reasoned decisions under their rules. Investors and operators should disclose the integrated bargain to their professional reviewers. Courts should examine substance when disputes arise. NRS should make the evidence comparable and portable while refusing a gatekeeping role.

The result may be sale, financing or a mixed structure. Mixed is not failure if the documents, accounts and remedies describe it accurately. Evasion begins when parties need one institution to see the sale, another to see the loan, and neither to see the retained control.

The test is therefore simple to state and demanding to apply.

Who controls the block after closing? For how long does the seller retain use? What can the buyer do during and after that term? Who bears market, registry, reputation and customer risk? Can the seller force or economically guarantee repurchase? Where do the cash flows go if performance fails?

Answer those questions with evidence, and the registry's moral label becomes unnecessary.

Sources