Summary

  • Registration uncertainty is not a single risk. A lender must distinguish the borrower's authority, the registry's recognition, continuing account compliance, route and lease performance, third-party claims, and the ability to realise value after default.
  • Public Cogent filings show that IPv4-backed finance is no longer hypothetical. The disclosed structure uses reserve cash, coverage and utilisation triggers, representations, indemnification for defective or ineffective asset transfers, recordkeeping duties, accelerated amortisation and the ability to substitute IPv4 addresses under stated leverage conditions.
  • A representation or warranty allocates the risk of a fact being wrong; it cannot bind a registry. A condition precedent protects funding until specified evidence exists; it should not be confused with a promise that registration will never later change.
  • Maintenance covenants price time. They require accurate records, fees, notices, lease administration, route-authorisation hygiene and early reporting of disputes. Borrowing-base exclusions and haircuts then convert deterioration into less credit rather than an immediate moral judgment about the borrower.
  • Substitute performance is the decisive protection. A borrower should be able to replace an impaired prefix with eligible addresses, cash or another agreed asset, subject to equivalence tests. The lender receives value preservation without acquiring a private right to force an RIR to accept a transfer.
  • Number Resource Society can make this market safer through a portable evidence vocabulary, event receipts and conformance tests. It should not certify title, approve loans, set advance rates or become a parallel registry gatekeeper.

The lender is financing a chain, not a database row

An IPv4 block looks deceptively simple on a collateral schedule. It has a prefix, a size and a registered organisation. A valuation may attach a price per address. The security agreement may place it within a broad category such as general intangibles, contract rights, Internet addresses or number-resource rights. A financing statement may then describe the collateral in language wide enough to catch whatever interest the borrower can grant.

None of that tells the lender what happens on the morning after default.

Value depends on a chain. The borrower must have authority to deal with the registration. The relevant registry must recognise the present holder and, if enforcement requires a transfer, accept the proposed recipient and process. Corporate records must support the signer's authority. No undisclosed lease, route authorisation, court order, competing security interest or account dispute should obstruct the remedy. The addresses must remain operationally usable enough for a buyer or lessee to pay. The lender or its enforcement vehicle must be able to satisfy any recipient requirements.

Registry records, RPKI objects, Internet Routing Registry entries, reverse DNS and customer arrangements may need coordinated changes.

A break in one link does not necessarily destroy the asset. It changes the probability, timing or cost of realisation. A stale contact can be cured. A policy hold may delay a transfer. A pre-existing lease may reduce immediate sale flexibility while preserving revenue. A route-reputation problem may lower price for one class of buyer but not another. A court dispute can make the timing genuinely unknowable. Contract design should react proportionately to each condition.

This is why a lender cannot solve IPv4 risk by declaring addresses to be property, nor by accepting a registry's contrary vocabulary as the end of the economic inquiry. The lender advances money against expected cash repayment and a package of secondary protections. It needs an enforceable claim against the borrower, truthful information, continuing controls, and realistic remedies. Registry recognition is one necessary external dependency in that package. It is not the whole package, and it is not under the borrower's complete control.

The financing document earns its keep precisely where control is incomplete. It maps who can promise what, defines evidence for the facts that matter, and specifies what the borrower must do when an external institution does not cooperate. Uncertainty becomes a set of priced obligations instead of a slogan.

Registry discretion is real, but it is not limitless or uniform

The word discretion can be abused in two directions. Borrowers may use it to suggest that no representation about a number resource can ever be meaningful. Registries may use administrative language to imply that their characterisation should settle every commercial question. Neither position is useful.

RIRs publish rules, agreements and procedures. ARIN's current transfer materials identify source and recipient requirements, officer acknowledgement, fees, demonstrated need in specified cases, dispute status and coordination for inter-RIR transfers. Its quick guide says that payment of the processing fee does not guarantee approval and that an approved transfer still requires the applicable Registration Services Agreement, fees and coordination before resources are transferred. Those are observable stages, not a promise of automatic recognition.

RIPE policy permits a legitimate resource holder to transfer complete or partial blocks that meet the policy conditions and states that the RIPE NCC completes the transfer by updating registration records. The RIPE NCC Standard Service Agreement also says registration does not constitute property or confer ownership, restricts assignment of rights under the agreement without consent, and contemplates deregistration under applicable rules. APNIC likewise describes transfer criteria for source and recipient entities and, in applicable transfers, demonstrated need.

These texts matter because they define institutional dependencies. They also show why a single global warranty such as "the addresses are freely transferable" is usually too broad. Eligibility, documentation, contractual status, regional rules and registry action can differ. A representation can accurately say that the borrower is the current registered holder, has not granted an undisclosed conflicting right, has paid due fees, has supplied specified records, and is not aware of a dispute. It cannot honestly say that every registry will approve every future enforcement transfer to any nominee.

Discretion should therefore be translated into bounded events. Did the registry acknowledge the request? Did it confirm that the source is the current registrant? Did the recipient obtain pre-approval where available? Were the required agreements executed? Was the record updated? Is a stated hold, appeal or dispute pending? Each event can carry a contractual consequence.

This approach also protects the registry from invented duties. A lender should not represent a transfer pre-approval as a credit guarantee or demand that staff opine on legal title. It can rely on the registry for what the registry actually controls: its account status, its procedures and its authoritative record. The loan contract must absorb the gap between those administrative facts and the lender's broader recovery objective.

Public finance has already moved beyond theory

The strongest evidence that contract architecture can price IPv4 uncertainty is not an analogy to mortgages. It is a disclosed IPv4 financing.

Cogent Communications reported that a bankruptcy-remote subsidiary issued $206.0 million of secured IPv4 address revenue notes in May 2024, with a 7.924% rate and an anticipated repayment date in May 2029. It later issued $174.4 million of a second series in April 2025, with a 6.646% rate and an anticipated repayment date in April 2030. By the end of 2025, the two disclosed series totalled $380.4 million in principal. The collateral package included specified IPv4 addresses, customer IPv4 leases, customer receivables and related assets.

Those figures describe one issuer and two series. They are not a global IPv4 lending rate, not a valuation per address, and not proof that another portfolio would obtain similar terms. The coupons incorporate the complete credit, structural, market and timing profile presented to those investors. Public filings do not isolate the price of registry risk from every other risk.

The disclosed controls are more revealing than the headline amount. Cogent described a liquidity reserve account, optional and mandatory prepayment provisions, indemnification if transfers of pledged assets were defective or ineffective, recordkeeping and information-access covenants, rapid amortisation linked to debt-service coverage, and utilisation thresholds tied to the proportion of addresses leased. The 2025 disclosures also described amendments permitting disposal and substitution of IPv4 addresses subject to a pro forma leverage condition and other requirements.

This is what it means to finance an uncertain institutional asset. The structure does not need an RIR to guarantee repayment. It surrounds the address portfolio with cash-flow assets, a special-purpose issuer, reserves, ratios, information rights, representations, remedies and substitution. If one part of the collateral chain weakens, the contract has intermediate responses before terminal enforcement.

The filings do not reveal every negotiated definition, legal opinion, registry communication or enforcement path. The notes were offered in exempt transactions, and the public descriptions summarise rather than reproduce the entire bargain. It would be wrong to infer that investors solved every question about perfection, registry recognition or insolvency. But it would be equally wrong to say lenders cannot structure around those questions.

The lesson is narrow and important: address value can support institutional credit when the contract measures more than the registry line. The collateral is a managed system of registered resources, leases, receivables, records and remedies. That system can be covenanted.

A broad collateral definition is only the opening move

Financing documents often begin with breadth. They include all assets, all general intangibles, contract rights, proceeds, substitutions and supporting obligations. Broad language prevents an accidental omission when an asset does not fit a familiar category. It does not establish that the debtor has transferable rights, that a security interest is perfected against every claimant, or that an external administrator must act for the secured party.

One SEC-filed financing amendment from 2019 illustrates both the ambition and the problem. It defined "Domain Assets" to include Internet domain names, Internet protocol addresses and related rights, assets and agreements. It also defined a Domain Name Control Agreement involving the borrower, collateral agent and applicable domain-name registrar. Later covenants focused that control-agreement machinery on domain names and registrars.

The drafting is useful evidence that lenders have placed IP addresses inside collateral definitions. It is not evidence that a domain registrar controls an IPv4 registration, or that the same tripartite control device works with an RIR. Domain names and Internet number resources have different administrators, contracts, transfer processes and technical dependencies. Grouping them in one defined term does not merge those systems.

Under the model rules of UCC Article 9, a security interest generally requires value, debtor rights in the collateral or power to transfer rights, and an authenticated security agreement or another specified basis. The broad residual category of general intangible can capture personal property not assigned to another category. But the UCC is enacted through state law, its application is jurisdiction-specific, and classifying a borrower's number-resource interest requires legal analysis.

A filing against "all general intangibles" cannot manufacture rights that the debtor lacks or compel a registry to recognise an enforcement buyer.

The prudent schedule therefore describes the collateral in layers. First comes the registered prefix and holder record. Second comes the borrower's contractual relationship with the registry. Third come leases, receivables and service rights. Fourth come operational authorities such as RPKI, routing records and reverse DNS. Fifth come proceeds, claims, indemnities and replacement assets. The security grant covers the borrower's interest in each layer to the extent legally available.

Precision does not weaken the lender. It prevents a false sense of control. The lender can then ask for the correct evidence and remedy for each layer rather than discovering after default that its "all assets" clause was an inventory list without an execution route.

Warranties price facts, not future institutional behaviour

A warranty shifts the economic consequence of an inaccurate statement. If the borrower says it is the registered holder, has authority to enter the financing, has disclosed all active leases, and knows of no adverse claim, the lender can rely on those facts in setting availability and price. If the statement is materially false, the contract supplies a breach, indemnity, draw stop, reserve or event of default.

The warranty should be drafted around facts the borrower knows or can verify.

The holder warranty identifies each prefix, registry, account, legal entity and registration status as of a date. It distinguishes direct registration from downstream use. It states whether legacy status or a special agreement applies. It avoids the metaphysical phrase "good and marketable title" unless counsel has established what that means under the governing law and registry relationship.

The authority warranty covers corporate approval, signatory power and the absence of a contractual prohibition on the grant. A separate compliance warranty addresses fees, account standing, required contacts and material submissions. A claims warranty identifies known disputes, notices, transfer locks, sanctions constraints, court orders and competing demands.

The use warranty maps active leases, assignments, route-origin arrangements, letters of authorisation, reverse-DNS delegations and customer dependencies. The reputation warranty should be carefully bounded: it can require disclosure of known blocklists, unresolved abuse campaigns or material routing incidents, but it should not promise a permanently clean reputation across every private list.

The cash-flow warranty reconciles customer contracts, billed receivables and collected amounts to the pledged portfolio. Where financing depends on lease revenue, a prefix that is registered but not lawfully available to the stated customer is a collateral defect even if the registry record itself is untouched.

Future institutional behaviour belongs elsewhere. The borrower can covenant to cooperate with a transfer and maintain eligibility. It can represent that no known fact currently prevents a request. It should not warrant that an RIR will never change policy, never ask a further question, never delay, never correct a record and always accept the lender's chosen vehicle.

Knowledge qualifiers, materiality and disclosure schedules should not become escape routes. The contract can separate objective facts, such as the name displayed in the record, from knowledge-based facts, such as an unfiled third-party claim. Repeating key representations at each draw converts fresh credit into fresh disclosure. The lender pays less for uncertainty when the borrower must keep statements current.

Conditions precedent buy evidence before exposure

A condition precedent answers a different question: what must be true or delivered before the lender is obliged to fund, release a reserve, permit a substitution or allow a disposition?

For initial funding, the lender should require a verified resource schedule, relevant registry agreements, corporate authority documents, evidence of current account standing, current public registration records, material lease schedules, lien searches, a legal analysis of the security grant, and the operational credentials or escrow arrangements needed to preserve services. Where a registry offers recipient pre-approval, that may be useful for a contemplated transfer. It is not equivalent to completed transfer approval.

If proceeds fund an acquisition, money should not move merely because the buyer and seller signed. The closing condition can require the registry event appropriate to the bargain: acceptance of the request, execution of the recipient agreement, or completed record update. The correct trigger depends on who bears interim risk. Escrow and holdbacks can bridge the period between commercial signing and authoritative registration.

For subsequent advances, the key representations should remain true, no defined default should continue, coverage tests should be satisfied, and the collateral report should be current. This is ordinary credit discipline applied to a specialised asset. Public credit agreements routinely make truth of representations and absence of default conditions to each extension. IPv4 finance needs the same architecture with resource-specific evidence.

Conditions should be objective enough to prevent discretionary drift by the lender. "Satisfactory to the lender in all respects" may be unavoidable for a narrow document at the first closing, but it is a poor permanent test for portfolio eligibility. A borrower cannot plan liquidity if every routine registry question permits an unlimited funding refusal. The agreement should list acceptable record states, cure periods, evidence sources and escalation routes.

Nor should a condition force unnecessary disclosure to the registry. The lender may need the private financing documents; registry staff generally need only the material required by policy and agreement for the requested service. Financing risk does not create a public interest in loan pricing, investor identity or customer revenue.

A well-drafted condition buys the lender time and information. It does not transfer adjudicative power to the lender or registry. If the condition is not met, the immediate consequence is that the specified credit action does not occur. The parties then use cure, substitution, cash collateral or termination rights already negotiated.

Covenants turn a static block into maintained collateral

Addresses do not preserve their financing value by remaining written on a schedule. They require administrative and operational care.

The first covenant is record maintenance. The borrower keeps legal name, authorised contacts, abuse contacts and other required registry information current. It pays applicable fees, responds to legitimate account notices and preserves copies of material submissions and decisions. Any change in account status, transfer eligibility or contractual relationship is reported promptly to the lender.

The second is negative control. Without consent, the borrower does not transfer, pledge, lease beyond permitted limits, grant conflicting route authority, divide, aggregate, return or request deregistration of pledged space. Consent standards should reflect ordinary portfolio management. A blanket prohibition can destroy value by preventing beneficial leases or technical changes. A permitted-disposition regime is usually better: the borrower may act if coverage remains adequate, proceeds enter controlled accounts and records are updated.

The third is revenue administration. Lease invoices, collections, defaults, renewals and concentration are reported against the prefix schedule. Customer payments may flow through controlled accounts. Contract amendments that materially weaken term, termination rights or assignment should require notice or consent.

The fourth is operational integrity. The borrower maintains reasonable routing authorisation records, avoids conflicting origin authority, monitors hijack or leak events, coordinates reverse DNS, and keeps abuse handling reachable. A lender should not operate the network, but it should know if the collateral has become commercially impaired.

The fifth is evidence continuity. Registry messages, transfer receipts, lease records and change histories are retained in an exportable form. Personnel turnover should not erase the authority chain. Credentials require controlled access, recovery and revocation procedures rather than one employee's mailbox.

The sixth is notice. A dispute, court order, registry hold, material policy inquiry, loss of account access, RPKI incident or threatened customer termination reaches the lender before the next quarterly certificate if it can materially affect value. Early notice is itself a credit protection because it expands the cure set.

These duties price time more accurately than a one-day legal opinion. The opinion can describe enforceability at closing. Covenants address what happens during the five years in which the borrower changes staff, customers, routes and corporate structure while policies also evolve.

Haircuts and eligibility rules convert doubt into availability

The most disciplined response to collateral uncertainty is often not a default. It is an eligibility adjustment.

A borrowing base includes assets that satisfy stated tests and applies an advance rate below their assessed value. In IPv4 finance, eligibility can require direct registration to a pledged entity, an acceptable registry agreement, no known dispute, a complete authority chain, compliant block size, accurate contacts, permitted lease status, usable routing history and an acceptable concentration profile. The lender then applies haircuts for risks it is willing to finance but not at full value.

An address block under a long customer lease may have strong recurring revenue but less immediate sale flexibility. A block with a recent transfer may have clean documentation but limited performance history in the new portfolio. A legacy block may have valuable characteristics while carrying a different contractual path. A prefix with a correctable reputation issue may still qualify at a lower advance rate. A block subject to an active ownership dispute may be excluded until resolution.

This is pricing in the contractual sense. The risk changes how much can be borrowed, what reserve must be held, which concentration limit applies and when cash is swept. It does not require a public price for "registry discretion" as a standalone commodity.

The Cogent disclosures offer a concrete version of dynamic protection. Utilisation and debt-service coverage affect amortisation and access to restricted proceeds. Address substitution and disposal are tied to leverage conditions. The public materials do not disclose a universal borrowing-base formula, yet they show that portfolio performance and replacement can be linked to creditor protection.

Valuation should also separate three numbers: market sale indication, present value of contracted lease cash flow, and stressed recoverable value after delay and cost. They may diverge substantially. The lender should state which one drives each covenant. A sale valuation does not prove that an enforcement transfer can close immediately. A lease valuation does not survive customer termination automatically.

No responsible article can supply a global haircut or default frequency. Public RIR transfer logs do not identify financed transactions, failed enforcement attempts, private waivers or recovery proceeds. Public issuer disclosures cover only selected borrowers. The denominator of IPv4-backed loans is unavailable. Contract design should acknowledge that data gap by using conservative scenarios, independent valuation and periodic review rather than a fabricated market average.

Substitute performance prevents one prefix from holding the loan hostage

The pivotal clause is the right and duty to substitute.

Suppose a pledged /16 becomes subject to a registry dispute unrelated to credit performance. If the agreement treats any impairment as an immediate terminal default, the lender may accelerate into the very uncertainty it fears. If it ignores the event, collateral coverage may silently deteriorate. Substitution creates a middle path.

The borrower delivers one or more replacement prefixes that meet defined eligibility tests. The collateral agent verifies registration, authority, liens, lease status, reputation, operational records and valuation. Coverage and concentration tests are recalculated. Necessary security documents are executed. The impaired block remains covered until the substitute becomes effective unless the lender receives cash collateral or another agreed bridge.

Equivalence should not mean identical prefix size alone. A replacement can contain the same address count yet be less valuable because it is fragmented, encumbered, difficult to route, concentrated in one customer, subject to a weaker authority chain or registered in a region whose transfer path does not fit the enforcement plan. The test should compare stressed value, cash flow, transfer pathway and operating condition.

Cash is a valid substitute where a clean prefix cannot be delivered promptly. The borrower can deposit the deficiency amount into a controlled account, pay down debt or purchase eligible replacement space within a bounded period. A third-party guarantee or letter of credit may cover temporary uncertainty, though it introduces separate counterparty risk.

Substitution also disciplines the lender. It cannot use a curable registration issue as a pretext to seize an otherwise performing business if the borrower supplies equivalent protection. The clause defines what adequate cure looks like and when consent may be withheld.

Public structured-finance agreements outside the IPv4 field commonly condition substitutions on eligibility, coverage, no continuing default, delivery of certificates and application of sale proceeds. Cogent's public filing goes further for this market by describing amendments that expressly permit substitution of new IPv4 addresses under a leverage test. It does not disclose every operational step, but it confirms that replacement is not a theoretical remedy.

The substitute clause is where institutional humility becomes bankable. The lender does not need power over the registry if it has enforceable power over the borrower's collateral pool and cash. It protects value by changing the input rather than commanding the external institution.

Indemnities should cover controlled defects, not every public decision

Indemnification can look like a complete answer: if registration fails, the borrower pays. In practice, an indemnity is only as valuable as the indemnitor's capacity and the clarity of the covered loss.

The strongest indemnity targets defective facts or actions within the borrower's control. Examples include a false authority statement, an undisclosed competing lease, a forged officer acknowledgement, failure to pay a fee the borrower covenanted to pay, an unauthorised route permission, or a transfer document that the borrower executed defectively. The loss can include reasonable correction cost, delayed collections, replacement expense and specified enforcement cost.

An indemnity should not casually make the borrower guarantor of every policy change, cyber incident at a registry, court order, geopolitical restriction or discretionary decision by an institution it does not control. Those risks are better allocated through eligibility, reserves, insurance where available, substitution, force-majeure analysis and negotiated risk sharing.

Cogent's filings report indemnification obligations where transfers of pledged assets are defective or ineffective in stated ways. The qualification matters. Public disclosure does not say the issuer indemnifies investors for any decline in the value of any address or for every registry act. The contract identifies defects that trigger a defined payment obligation.

Caps, baskets and survival periods should match the risk. Fraudulent authority may justify uncapped recourse. An administrative error cured in five days may not. A special-purpose issuer can offer limited recourse by design, making reserve and replacement mechanics more important than a large nominal indemnity against an entity with few unpledged assets.

The lender should also avoid circular recovery. If an indemnity payment comes from the same lease revenue already supporting debt service, it may only relabel the shortfall. Cash reserves, sponsor support or separate collateral can make the remedy real.

In this market, a promise to compensate is not a substitute for a method to continue operating. The best sequence is correct the record, preserve customers, replace the impaired collateral, use reserves, and only then quantify residual loss. Contractual damages sit behind continuity, not in place of it.

Enforcement begins before a transfer request

A weak security package imagines enforcement as a form sent to the registry after default. A strong one begins months or years earlier.

The lender knows the exact legal entities holding the resources. It holds current authority evidence, registry agreements and contact maps. It understands which leases survive a change of holder, which require consent and which can be terminated. It knows the operational providers needed to maintain announcements, RPKI and reverse DNS. It has a qualified transferee vehicle or a sale process that can identify one. It has contractual power to obtain records and direct proceeds.

After default, the first objective is stabilisation. Freeze unauthorised dispositions, preserve account credentials, continue essential fees, notify critical operators, collect lease revenue and prevent contradictory route instructions. A receiver, collateral agent or servicer needs defined authority under the loan documents and applicable law. The registry may require separate proof before recognising any account action.

The second objective is classification. Is the problem payment default, covenant breach, fraud, insolvency, registry suspension, customer loss or operational misuse? Each supports a different remedy. An immediate sale may be rational after payment default with a clean portfolio. It may be destructive during a temporary registry-access incident.

The third objective is market preparation. Buyers need due-diligence evidence and may need regional pre-approval. Existing lessees need continuity assurances. Address reputation and route records must be reviewed. Sale lots should reflect routing and transfer constraints, not merely maximise theoretical address count.

The registry request is one step in that process. The lender should submit only through an authorised party and with the evidence required by the applicable RIR. It should not assume that a UCC foreclosure document, court order or private power of attorney will be treated identically in every region. Legal counsel must connect the enforcement sale to the registry's recognised transfer route.

The final objective is reconciliation. Registration, RPKI, routing records, reverse DNS, lease notices, collection accounts and debt balances must agree. A record update without cash settlement or customer continuity can still leave a recovery shortfall.

Contractual readiness lowers expected loss even if enforcement is never used. It also reduces the temptation to pressure a registry into an improvised decision. The parties arrive with a documented, legally reviewed request rather than asking an administrative institution to repair years of neglected collateral governance.

Cross-regional finance needs a matrix, not a global fiction

A portfolio can span ARIN, RIPE NCC, APNIC and other service regions. The financing agreement may be governed by one law, while registry contracts, borrowers, customers and routes sit elsewhere. Calling all addresses one asset class does not erase those differences.

The collateral schedule should record the RIR, registration type, governing registry agreement, direct holder, transfer restrictions, inter-RIR compatibility, applicable recipient conditions, minimum transferable size, account status and known legacy characteristics for each block. The legal analysis then maps local security and insolvency rules to that administrative path.

Inter-RIR transfers add a second institution. Approval by one side may not complete the other side's requirements. The loan should define when a disposition is treated as effective, who bears delay, and whether proceeds remain blocked until both authoritative records reconcile. A substitute intended for another region should not qualify merely because the source registry released it.

Policy change requires a measured response. A covenant can require notice and good-faith adaptation when a rule materially affects eligibility. The lender may adjust availability after an independent determination, not after every mailing-list debate. Existing transactions need grandfathering analysis rather than automatic panic.

Jurisdiction also shapes the meaning of security. UCC concepts can guide a United States transaction, but they are not a global law of number resources. A RIPE agreement is governed by Dutch law. APNIC operates under Australian corporate arrangements. Borrowers may be organised elsewhere. The agreement should not advertise a universal perfection conclusion where counsel has only analysed one state.

This complexity argues for modular documents. A common credit agreement defines coverage, reporting, substitution and remedies. Regional schedules add local registration and legal conditions. The lender can then compare risk without forcing every registry relationship into ARIN terminology or every enforcement path into a New York filing.

The absence of a global legal label does not make finance impossible. International lending routinely handles assets subject to local registries, licences and transfer restrictions. The difference is that mature asset classes have tested opinions, custodians and enforcement practice. IPv4 finance is building that infrastructure. Honest matrices are more valuable than premature uniformity.

The due-diligence file should be designed for the next adverse event

Good diligence is not a photograph taken for closing. It is an evidence system that a replacement employee, auditor, receiver or buyer can understand under pressure.

The resource file begins with a machine-readable schedule of prefixes and a human-readable explanation of holding structure. Each entry links to the registry record, agreement, account identifier, authority documents, acquisition history and relevant transfer receipt. Hashes or signatures can protect integrity without publishing confidential materials.

The claims file contains lien searches, financing statements, disclosed leases, purchase agreements, court orders, settlement obligations and correspondence about competing authority. A clean representation means little if the evidence cannot be found.

The operations file maps expected origin ASNs, RPKI authorisations, route objects, reverse-DNS control, geofeed or geolocation dependencies, abuse contacts and critical service providers. It records changes and exceptions. This does not make the lender a routing operator; it makes impairment observable.

The revenue file maps each lease and receivable to the covered prefixes, customer, term, termination provisions, concentration and collection account. Personal and commercially sensitive information can remain access-controlled. Investors need assurance, not a public customer list.

The institution file records material registry tickets, fees, audits, notices, pending requests and decisions. It distinguishes a routine question from a restriction on transfer. A single status field such as "compliant" is too vague for credit decisions.

The continuity file contains credential-recovery procedures, alternate authorised contacts, servicer arrangements, incident playbooks and the sequence for substitution or enforcement. It is tested periodically. A plan that depends on an employee who left last year is not collateral control.

Reporting frequency should follow risk. Stable holder records may be checked monthly or quarterly. Route and lease data can change more often. A dispute requires event-driven reporting. The lender should sample underlying evidence rather than rely solely on management certification.

Privacy remains a constraint. The diligence file is not a public registry. Access is role-based, logged and limited to the credit purpose. Public output can identify aggregate portfolio characteristics without exposing customer contracts or personal credentials.

Designed this way, diligence reduces both lender risk and registry burden. When a legitimate request arises, the parties can present coherent evidence quickly. They do not ask registry staff to reconstruct the borrower's corporate and commercial history from fragments.

A practical covenant matrix

The following matrix is not a standard form and cannot replace transaction-specific legal advice. It shows how separate tools answer separate risks.

Risk Evidence at closing Continuing covenant Trigger First remedy
Borrower lacks authority Corporate approvals, holder record, agreements, acquisition history Preserve entity and authorised contacts; no unauthorised disposition False statement or competing authority Cure evidence, suspend advances, substitute or prepay
Registry will not recognise a contemplated transfer Pre-approval where available, policy analysis, recipient readiness Maintain eligibility and report policy or status changes Rejection, hold or material delay Appeal or correction, replacement prefix, cash collateral
Lease revenue deteriorates Lease schedule, concentration and collection history Report renewals, defaults and utilisation; control collections Coverage or utilisation threshold breach Cash sweep, reserve increase, accelerated amortisation
Address is operationally impaired Route, RPKI, reputation and abuse review Maintain authorisations and incident response Hijack, conflicting ROA, material blocklisting Correct records, move customer, substitute affected space
Third-party claim emerges Lien and dispute searches, warranties, notices No further liens; prompt claim notice Court order, registry dispute, adverse claim Defence, holdback, indemnity, substitute or release
Portfolio changes Eligible-prefix schedule and valuation Permitted disposals only if coverage survives Proposed sale, split, aggregation or new lease Consent under objective test, apply proceeds, add collateral
Institution or service becomes unavailable Exported evidence, alternate contacts, continuity plan Maintain portable records and tested recovery Outage, account lock or organisational failure Alternate service path, temporary cash protection, orderly migration

The matrix makes one principle visible: the loan does not have one binary "registry covenant". It has an authority covenant, a registration covenant, a revenue covenant, an operational covenant, a claims covenant and a continuity covenant. A breach affects the layer it threatens.

That granularity improves pricing. A lender can offer more availability against a portfolio with complete evidence, diversified lease revenue, tested substitution and current records. A borrower that refuses information or cannot replace an impaired block should expect lower advance rates, more reserves or shorter tenor. The contract rewards controllable resilience rather than institutional favour.

It also improves conduct. Neither party benefits from declaring a minor administrative delay to be fraud or from concealing a serious dispute as paperwork. Defined triggers and remedies create a shared language for escalation.

NRS can standardise evidence without certifying the loan

Number Resource Society has a constructive role because this market lacks a portable vocabulary for credit-relevant number-resource events. That role must be ambitious about evidence and modest about power.

NRS can define a resource-finance evidence profile: prefix, direct holder, registry relationship, authority proof type, active-use category, lease status, expected origin context, dispute state, effective dates and last verification. Sensitive documents remain with the parties, while signed attestations and hashes make later substitution or enforcement easier to reconcile.

It can standardise event receipts. A receipt can distinguish request submitted, identity verified, recipient eligible, agreement executed, registration updated, correction pending and dispute noted. Lenders then stop treating an email saying "approved" as if it always meant final record change.

NRS can publish conformance tests for portfolio systems. Do records cover the full prefix without overlap? Does each lease map to a registered block? Are authority attestations current? Do substitutions preserve address count, stressed value, region compatibility and operational readiness? Can the evidence be exported to another servicer?

It can also support anonymised research. Participating lenders and borrowers could report time to cure, substitution frequency, cause categories and recovery sequence under a privacy-preserving protocol. The denominator must be explicit: participating facilities, reporting period and covered address count. Voluntary reports must never be presented as the entire global market.

NRS should not declare that a borrower owns an address, that a security interest is perfected, that an RIR must accept an enforcement buyer or that a loan deserves a particular rate. Those are legal, institutional and credit decisions. It should not maintain a compulsory lender whitelist or require financing terms as a condition of number-resource recognition.

Its public description as a membership organisation focused on number-resource holders supports an interest in this work, but that is first-party evidence of mission, not proof that a finance standard is deployed or trusted. A credible programme would publish the specification, invite lenders, operators, registries, insolvency practitioners and privacy experts, and allow independent implementations.

This positive role strengthens markets without replacing them. NRS makes evidence legible and portable. Competition decides credit terms. Courts decide legal disputes. Registries administer their records. Operators control routing. No institution needs to absorb every function.

What the evidence does not establish

The available public record is unusually informative and still incomplete.

Cogent proves that large disclosed IPv4-backed note transactions have closed and that public reporting can describe reserves, coverage triggers, utilisation tests, defective-transfer indemnification, substitution and creditor remedies. It does not reveal the full investor diligence, every indenture definition, every legal opinion or the probability assigned to an RIR action. One issuer cannot establish a market-wide default rate or risk premium.

SEC-filed credit agreements show how sophisticated finance uses representations, conditions precedent, covenants, borrowing-base tests and collateral substitution. Most such documents concern assets other than IPv4. They support the clause architecture, not an assertion that every term transfers unchanged to number resources.

RIR policies and agreements establish the administrative conditions and contractual language in force at the cited institutions. They do not determine secured-transactions law, tax ownership, accounting control or the rights of every insolvency estate. Nor are the five regional systems identical.

UCC Article 9 supplies a framework for attachment, debtor rights, proceeds and general intangibles in adopting jurisdictions. It does not itself label every IPv4 interest, override all contractual restrictions or bind a foreign registry. A competent opinion must analyse the actual debtor, agreement, forum and remedy.

Basel and EBA principles support sound credit granting, monitoring, collateral valuation and controls. They are not IPv4 term sheets. They explain why lenders should measure and monitor risk, not what advance rate any block deserves.

No public dataset provides the total number of IPv4-secured loans, the proportion containing substitution rights, the frequency of registry-related covenant breaches, or losses after attempted enforcement. Private facilities, waivers and failed negotiations are largely invisible. Any global percentage would be invented.

These limits do not weaken the article's conclusion. They define it. The case is for contracts that survive uncertainty, not for a claim that uncertainty has already been statistically conquered.

The best covenant does not ask the registry to become a lender

IPv4 finance will remain exposed to institutions outside the credit agreement. Registration, policy, courts, routing and customer behaviour cannot be reduced to one collateral-agent instruction. The lender's task is not to abolish that reality. It is to decide which uncertainty it will fund and on what terms.

Warranties make the borrower pay for false facts. Conditions precedent keep money behind a gate until specified evidence exists. Covenants preserve registration, revenue, operations and records through time. Borrowing-base rules and haircuts turn deterioration into measured availability. Reserves buy time. Indemnities address controlled defects. Substitute performance prevents one disputed prefix from dictating the fate of an entire facility. Enforcement planning connects legal remedy to actual registry and network steps.

This architecture is more demanding than writing "all IP addresses" in a collateral definition. It is also more respectful of institutional boundaries. An RIR is asked to maintain and update its record under its rules, not to guarantee a private loan. A court applies the law, not a transfer policy. A lender judges credit, not community virtue. A borrower carries the facts and duties it controls.

The public Cogent transactions demonstrate the commercial direction: IPv4 addresses, leases and receivables can sit inside a structured credit system with ratios, reserves, replacement and remedies. They do not settle every legal question, and their disclosed coupons are not a universal price. What they settle is the claim that registry uncertainty makes disciplined finance conceptually impossible.

NRS can help the next transactions by making evidence and event states portable. Its success should be measured by fewer ambiguous closings, faster cures, reliable substitutions and lower evidence-reconstruction cost, not by the number of loans it approves. It should make contracts easier to compare while leaving credit and law where they belong.

The registry's discretion is a dependency. It should appear in the contract with dates, evidence, boundaries and consequences.

Once it does, the lender no longer needs to pretend the dependency is certainty. Nor does it need to treat uncertainty as a reason to deny the economic reality of the resource.

It can lend against what the borrower can prove, maintain and replace.

Sources