Summary

  • Purchase price allocation is not a referendum on absolute ownership. Under acquisition accounting, the acquirer identifies assets and liabilities obtained, measures them at acquisition-date fair value and records the residual as goodwill or, in unusual cases, a bargain purchase gain. A transferable registration position can qualify for that exercise even if an RIR describes number resources as licensed, contractual or non-freehold.
  • The public record now contains material examples. Cogent assigned $458 million to IPv4 addresses acquired with the Sprint wireline business and described the amount as a novel valuation based on recent auction prices adjusted for market uncertainty. Uniti's 2025 Windstream merger included a preliminary $186.4 million IPv4 value after a $20.8 million measurement-period increase.
  • These values affect more than a note labelled “intangibles”. They can reduce goodwill, enlarge a bargain purchase gain, create or change deferred tax balances, determine future amortisation or impairment testing, and influence debt, covenant and return-on-capital analysis.
  • Valuation must begin with the exact bundle that can be controlled: specified prefixes, a verified chain from the registered holder, transfer eligibility, the receiving party's ability to obtain registry recognition, operational credentials, reputation and the right to use or monetise the space within applicable policy. A count of addresses without those conditions is not an asset inventory.
  • Market evidence is real but heterogeneous. A clean legacy block, a fragmented set of small prefixes, a registry-constrained transfer, a routed block with poor reputation and an address portfolio supporting leases are not interchangeable. Published per-address prices cannot be converted into a global rate without transaction-level adjustments and known denominators.
  • Useful-life judgments expose a second disagreement. Cogent treats its acquired IPv4 asset as indefinite-lived and tests it for impairment; other public companies amortise purchased addresses over finite periods such as ten years. Neither choice should be copied without evidence about IPv6 substitution, demand, policy, transferability and the holder's operating model.
  • Number Resource Society can improve evidence without becoming a price setter or title authority. A voluntary transaction-evidence standard could preserve prefix counts, transfer conditions, registry status, valuation date, comparable adjustments, useful-life assumptions and subsequent outcomes while keeping confidential deal terms protected.

The price enters through the accounting side door

Corporate vocabulary can preserve a disagreement for years. A registry says that addresses are public resources, registrations, licences for use or contractual rights rather than freehold property. A seller calls them assets. A buyer calls the payment consideration for a transfer of rights. A court approves a sale of specified interests. A network engineer wants the routing, reverse-DNS and security controls to work. Each institution uses language fitted to its mandate.

An acquisition closes despite that disagreement. The acquirer then has to prepare a balance sheet as of the date it obtained control. Cash, receivables, fibre, routers, leases, customer relationships, debt and tax positions are measured. If the acquired business controls millions of scarce IPv4 addresses that can support customers, avoid cloud charges, be leased or be transferred, omitting them can force their value into goodwill or distort a bargain purchase gain. Calling them “not property” does not make their economic effect disappear.

This is how market value entered merger accounting without a constitutional settlement. Accountants did not need to determine that an IPv4 address is an absolute thing good against the world. They needed to identify what the acquirer obtained, whether the position was separable from goodwill or arose from enforceable arrangements, whether future benefits were probable under the applicable framework, and whether fair value could be estimated. The answer can be an intangible asset described with careful boundaries.

The distinction is familiar outside networking. A licence, concession, customer contract, spectrum authorisation or landing slot can have substantial value while remaining conditional, regulated, time-limited or revocable. Accounting recognition does not erase the issuing institution. On the contrary, restrictions imposed by that institution enter the valuation. A licence that can be transferred only with consent is still economically relevant; the probability, timing and cost of consent affect its price.

IPv4 makes the tension unusually visible because the identifiers are globally coordinated and operationally dependent on distributed acceptance. The holder cannot compel every network to route a prefix. The RIR does not promise reachability. Yet a clean registration, established use and recognised control can save a buyer large sums and support revenue. The balance sheet records that bounded advantage, not ownership of the Internet.

This narrower reading matters. If accounting value were treated as proof of unqualified title, every estimate would become a legal weapon. If registry language were treated as proof that no asset can exist, audited accounts would have to ignore observable transactions. Both positions ask one institution to answer another institution's question. Purchase price allocation is useful precisely because it can recognise the economic position while exposing the conditions around it.

Purchase price allocation separates what the buyer acquired

Acquisition accounting begins with a total transaction and disaggregates it. The acquirer determines the consideration transferred and recognises identifiable assets acquired and liabilities assumed. Under IFRS 3, identifiable items are measured at acquisition-date fair value, with the remainder assigned to goodwill; an excess of net identifiable assets over consideration is recognised as a bargain purchase gain after the required reassessment. United States filings applying ASC 805 describe the same operating logic: record acquired identifiable tangible and intangible assets and assumed liabilities at fair value, then calculate goodwill or bargain purchase.

The allocation matters because the labels have different consequences. Goodwill is a residual attached to the acquired business. A separately identified IPv4 asset can be sold, leased, impaired or, if finite-lived, amortised on its own pattern. Investors can see that part of the price was attributed to a scarce operating input rather than undifferentiated expected synergies. Management must defend the amount and the subsequent accounting.

The process also prevents double counting. If address value is captured in a customer-relationship valuation because customers depend on IPv4-enabled service, and then the full value is separately assigned to addresses, the same cash flow may appear twice. If a lease portfolio is valued using income from addresses and the underlying address asset is also valued through the same lease cash flows, the valuer must separate contributory charges or choose a consistent unit. The allocation is a system, not a list of optimistic appraisals.

An acquired address position can enter through several transaction forms. In a share acquisition, the legal holder may remain the same while control of the company changes. In an asset acquisition, specified prefixes and related business assets move to a new holder. In a merger, the registered organisation may disappear into a successor. In an internal reorganisation, accounting may continue at carrying values rather than acquisition-date fair value because common-control rules differ. The registry event and the accounting event therefore need not occur in identical form.

Measurement is made at a date, not for all time. A price observed before a transfer-policy change, a large cloud-provider purchase or a market dislocation may not represent conditions at closing. The valuer should use information that market entities would have known at the acquisition date, then distinguish later evidence that confirms those conditions from a later event that changed them. This discipline is especially important in a thin market where a few disclosed trades can dominate a benchmark.

The result is not a certificate saying “these addresses are worth this amount everywhere”. It is an estimate for a particular holder, bundle, date and accounting purpose. The same numerical block can have a different value in another transaction because its registration history, route reputation, prefix composition, service region, transfer lane, operating use and transaction costs differ.

Identifiability does not require a metaphysical answer

Under IAS 38, an intangible asset is an identifiable non-monetary asset without physical substance. The IFRS Interpretations Committee has restated the two routes to identifiability: the asset is separable, meaning capable of being sold, transferred, licensed, rented or exchanged, or it arises from contractual or other legal rights. The rights do not need to look like land title. A right can be conditional and still be identifiable.

IPv4 positions can present evidence on both routes. A functioning transfer market shows that specified blocks and associated registration rights can be separated from one holder and moved to another under regional policy. Registry service agreements, legacy status, transfer approvals and transaction contracts define enforceable relations among parties. Even where an RIR denies freehold ownership, it may acknowledge exclusive registration, use and policy-compliant transfer rights.

ARIN's 2022 explanation is unusually direct: Internet number resources are not freely held property, but they constitute a bundle of contractual rights created on issuance.

That statement narrows the accounting entity rather than destroying it. The buyer is not valuing the integers in isolation. The buyer is valuing an exclusive and administratively supported position concerning specified integers: the ability to be recognised as registrant, maintain records, operate reverse DNS, obtain applicable routing-security services, use the addresses in a network, and transfer or monetise the position subject to policy and contract.

Legacy resources complicate the contractual route because some were issued before modern RIR agreements. They do not automatically become valueless. Evidence can include the original registration, uninterrupted use, public records, court orders, corporate succession, past registry treatment and the transfer agreement that the recipient will sign. Separability may be demonstrated by actual exchange even where the historical source did not begin with a current contract.

Control also needs precision. An acquirer controls an economic resource for accounting purposes when it can obtain benefits and restrict others' access in the relevant sense. A globally unique prefix has practical exclusivity because conflicting use creates routing failure and registry conflict. But exclusivity is not perfect. Hijacking can occur; networks can reject routes; an RIR can revoke or deregister under defined conditions; sanctions or court orders can interrupt service. Those risks resemble impairment and enforceability factors, not automatic disproof of control.

The recognition analysis should record the limiting facts. Was the transfer approved or merely expected? Does the buyer have authority over the registry account? Are the prefixes actually part of the acquired legal entity? Are any subject to dispute, lease, encumbrance, route-origin authorisation or customer assignment? Can the buyer use them in its service region? Does the seller retain a right of use? The asset exists in the quality of these answers.

This is why an acquisition agreement that says “all IP addresses” is inadequate valuation evidence. The phrase can include customer endpoint data, internal private ranges, provider-assigned addresses and public number resources. Only a reconciled schedule of transferable public prefixes tied to authority and registry records establishes the unit being measured.

Nortel established a market fact in 2011

The 2011 Nortel transaction remains the starting point because it made a previously informal scarcity visible in a supervised sale. Nortel agreed to sell 666,624 legacy IPv4 addresses to Microsoft for $7.5 million, or $11.25 per address. The transaction was presented in the Delaware bankruptcy proceeding, and the registry relationship was addressed before completion. It was not an ordinary allocation and it was not a theoretical estimate. A sophisticated buyer paid a specified amount for a specified address inventory connected to a transfer of registration rights.

The transaction did not create a universal price. It involved legacy space, a distressed seller, a large buyer, a particular block composition and a market at the moment of exhaustion. The court process affected sale conditions. Microsoft had its own operational need. ARIN's involvement affected delivery. The per-address arithmetic is historically useful but economically incomplete.

Nortel nevertheless changed the burden of argument. After the sale, an acquirer with similar space could no longer say that no observable exchange existed. An auditor could ask whether omitted address value was material. Creditors could ask whether an estate had investigated the value of unused prefixes. Boards could ask whether a sale or retention decision was rational. Brokers and later transfer facilitators gained a public reference point.

The important institutional outcome was coexistence. The court could approve a sale of the debtor's interests; ARIN could preserve its policy and agreement; Microsoft could obtain a recognised registration position; and none of those acts had to announce a global law of property. The economic transfer was real because the entities coordinated their different powers.

That pattern foreshadowed purchase price allocation. A valuer does not need an abstract asset detached from institutions. The valuer needs evidence that a market entity can acquire an enforceable and operationally useful position. Nortel provided transaction evidence. Later RIR procedures made the path more regular. Scarcity and continued IPv4 dependence created recurring demand. Public-company accounts then began to show the value at balance-sheet scale.

Nortel should still be used carefully in modern analysis. A 2011 price is not a current comparable without time adjustment. A large legacy portfolio may deserve a block-size or quality adjustment. The transfer's distress context can pull price down, while the novelty and strategic buyer can push it in another direction. The record does not disclose every assumption needed for a modern appraisal. It proves a market fact, not a universal denominator.

Cogent-Sprint made the address line material

Cogent Communications' acquisition of the Sprint wireline business on 1 May 2023 is the clearest public example of IPv4 value changing a major purchase allocation. The transaction was unusual: the seller provided substantial consideration to the buyer, and Cogent recognised a material bargain purchase gain. In that setting, every large acquired asset and liability affected an already sensitive accounting result.

Cogent's filings state that it recorded a $458 million intangible asset for acquired IPv4 addresses after management determined both the quantity for which title was transferred and the valuation approach. The company retained certified valuation specialists. It described the asset as novel and the transaction as a distressed business combination, two facts that increase rather than reduce the need for transparent judgment.

The valuation used recent auction prices and a factor reflecting uncertainty about how the IPv4 market would function in the future. This is a market approach with an explicit adjustment, not a claim that every address equals the latest headline price. The quantity, comparables, uncertainty factor and condition of the acquired portfolio jointly support the amount.

The wording about determining the quantity for which title was transferred is important but should not be inflated. It is management's acquisition and accounting description. It does not bind ARIN, another RIR, a court in an unrelated jurisdiction or every network operator. It shows that the company would not record the asset until it believed the transfer perimeter was established. That is good acquisition control: quantity follows evidence, not the seller's spreadsheet.

Cogent judged the IPv4 asset to have an indefinite useful life and therefore did not amortise it. Instead, the company tests for impairment annually and more often if indicators arise. Its public impairment discussion considers market data for address sales and leases and cash flows associated with the asset. “Indefinite” in accounting does not mean immortal. It means that, based on the current analysis, no foreseeable limit to the period of net cash inflows can be determined. IPv6 adoption, policy change, technological substitution or market decline can alter that judgment.

The purchase allocation also interacted with tax accounting. Cogent disclosed that measurement-period adjustments, including the newly recorded IPv4 amount and changes to other acquired values, increased the net deferred tax liability. The address appraisal therefore did not simply move a label inside total assets. It affected the recognised difference between book values and tax bases and the resulting bargain purchase calculation.

The acquired portfolio later supported a business model beyond internal network use. Cogent leases IPv4 space and has financed a group of IPv4 address assets and related lease cash flows through secured notes issued by a special-purpose subsidiary. That later financing does not prove the acquisition-date fair value, but it corroborates the proposition that the acquired position can generate separately observable benefits. It also creates subsequent evidence against which original assumptions can be tested.

Cogent is one company, not the market. Its scale, portfolio, lease programme, tax position and distressed acquisition are distinctive. The case should be used to identify the mechanics and questions, not to apply $458 million or an inferred per-address rate to every telecom merger.

Uniti-Windstream shows that the number can move goodwill

The 2025 merger between Uniti and Windstream supplies a second large and more conventional purchase allocation. Uniti's 2025 annual report describes an acquisition-method allocation of $2.3766 billion in merger consideration to Windstream's identifiable assets and liabilities. The preliminary table included $186.4 million for IPv4 addresses, alongside customer relationships, trade name, spectrum licences, rights of way, property and goodwill.

The amount was not static. During the fourth quarter of 2025, the company increased the estimated fair value of the IPv4 addresses by $20.8 million as a measurement-period adjustment. Other asset and liability values also changed. The combined effect included a $6 million reduction in goodwill. The company said third-party appraisals and related tax impacts were still being finalised at year-end.

This is what purchase price allocation looks like in practice. A preliminary balance sheet is not a ceremonial table produced on closing day. Management collects prefix inventories, validates legal and operational facts, selects methods, receives specialist work, resolves tax bases and updates estimates within the permitted measurement period. An IPv4 adjustment can be large enough to move goodwill visibly.

The case also warns against treating goodwill as a harmless plug. If address value is omitted, goodwill rises. If a later appraisal identifies the value, goodwill falls, subject to related tax effects and other changes. Analysts comparing acquisitions can misread “strategic goodwill” if one acquirer separately values address space and another buries the same economics in the residual.

Uniti's filing provides a total value but not every denominator needed for a public per-address calculation. The article therefore preserves the missing information. It does not divide $186.4 million by an assumed count, infer a global price or compare the result mechanically with Cogent. A useful comparison would require the exact transferable inventory, prefix sizes, reputation, utilisation, regional status, lease encumbrances, expected use, valuation date and method.

The measurement-period change is itself evidence. It indicates that the first estimate was subject to refinement as appraisals and facts developed. That is normal in a complex acquisition, but it raises an audit question: what new information related to conditions at the acquisition date justified the increase? A good disclosure distinguishes newly obtained evidence about existing conditions from a later market rise that should not be backdated into the allocation.

For registry governance, the transaction shows why merger records matter economically. A delayed or disputed update can affect which prefixes qualify for recognition, who can administer them and whether the buyer controls the expected bundle. Registry processing is not responsible for setting the accounting value, but the certainty and speed of the process can affect the risk discount embedded in it.

Standalone purchases confirm that value is not merely a merger residual

An objection to purchase price allocation is that management can assign optimistic numbers among categories while total consideration remains fixed. Standalone address purchases provide a useful cross-check because cash is paid directly for the resource position rather than inferred as one element of a whole company.

MongoDB disclosed that it purchased $24 million of intangible assets for IP addresses during the three months ended 31 January 2025. It said the addresses were expected to reduce future cloud-infrastructure costs and amortised them on a straight-line basis over ten years. The disclosure identifies consideration, purpose and useful life. It turns the economic benefit into an avoided-cost thesis: owning the address position reduces what the company would otherwise pay in its operating environment.

DigitalOcean's public filings similarly describe purchased intangible assets consisting of Internet Protocol addresses and source code, amortised over an estimated ten-year useful life. Its 2025 disclosures show a separate historical-cost and accumulated-amortisation line for IP addresses. These are not merger allocations, but they demonstrate recurring recognition by businesses whose services depend directly on Internet infrastructure.

The United Kingdom's Department for Work and Pensions offers a different institutional example. Its public accounts have treated IP addresses as a specific intangible sub-category and valued them by reference to the income they could return in the emerging market, with reclassification to assets held for sale when criteria are met. A public body using government accounting rules thus recognises market value without needing to claim an unrestricted freehold estate.

These examples should not be flattened. MongoDB's avoided cloud cost, DigitalOcean's service infrastructure and DWP's potential disposal are different uses. Useful lives and valuation premises can differ. The phrase “IP addresses” may also be broader than IPv4 in some filings unless the surrounding disclosure is explicit. Evidence quality requires preserving the issuer's exact description.

Together they weaken the idea that IPv4 value exists only because a merger accountant needed to reduce goodwill. Cash purchases, operating savings, public-sector valuations, leases and sales provide independent economic channels. They also improve the comparable set for a merger appraisal, subject to necessary adjustments.

The strongest analysis reconciles all channels. Does the price implied by market transactions fit the buyer's avoided-cost model? Do expected lease cash flows support the carrying amount after contributory charges and operating cost? Does a subsequent sale validate or contradict the original estimate? Are accounting useful lives consistent with the company's IPv6 and infrastructure plans? Convergence is more persuasive than one headline comparable.

Three valuation approaches answer different questions

The market approach begins with observed transactions for identical or comparable address blocks. It is intuitive and has the strongest connection to the transfer market. The valuer derives a unit price or range, then adjusts for block size, prefix fragmentation, registry region, legacy status, transaction date, reputation, routing condition, transfer restrictions and other differences.

The method becomes weak when the denominator is crude. One /12 is not economically identical to thousands of /24s. A clean block with coherent registration may command a different price from a block with spam history or disputed authority. A confidential broker quote may reflect an asking price rather than a completed transfer. An auction can include distress or strategic effects. The valuer needs transaction-level evidence, not a chart of anonymous averages.

The income approach estimates the present value of cash flows attributable to the address position. For a lessor, that can mean lease revenue less vacancy, bad debt, abuse handling, registry cost, technical administration, expected price decline and contributory charges for staff, systems and customer relationships. For an operator, the benefit may be revenue enabled by serving IPv4-dependent customers or costs avoided by not renting addresses or purchasing cloud connectivity that includes them.

Attribution is the central risk. Customers buy connectivity, hosting or cloud service, not an address in isolation. The model must separate value contributed by fibre, data centres, software, brand, support and customer relationships. If the entire customer margin is assigned to IPv4, the estimate will overstate the resource and duplicate other intangibles.

The cost approach asks what it would cost to replace the service capacity. There is no newly manufactured equivalent to scarce IPv4, so direct replacement cost is often unsuitable. A functional substitute may combine IPv6 deployment, carrier-grade NAT, application changes, customer migration, support and residual IPv4 purchases. That analysis can illuminate avoided cost but may not represent a market entity's exit price if the substitute has different risk and utility.

IFRS 13 defines fair value as an exit price in an orderly transaction between market entities at the measurement date, using market-entity assumptions including risk. For a non-financial asset, highest and best use must be physically possible, legally permissible and financially feasible. In the IPv4 context, “legally permissible” and operationally feasible require attention to transfer policy, agreements and registry recognition; the appraiser cannot assume away the institutional environment.

A robust valuation normally triangulates. Market evidence anchors scarcity. Income evidence tests whether the acquired business can earn or save enough to support that price. A substitution analysis tests the ceiling created by IPv6 and address-sharing alternatives. Differences are explained rather than averaged mechanically.

The unit of account is a rights bundle, not 32 bits

Valuation fails early if the entity is misdescribed. An IPv4 address is mathematically one of roughly 4.3 billion possible values in a 32-bit space, but not every value is globally unicast, available, transferable or operationally equivalent. Reserved ranges, private-use space and addresses held by others are not part of the buyer's inventory. The accounting unit begins with specified public prefixes tied to an authority chain.

The prefix structure matters. Routing tables operate on prefixes, not loose piles of individual addresses. Regional transfer policies commonly impose minimum sizes. A larger aggregate can be operationally attractive because it reduces route announcements and administration. Fragmentation can widen the buyer pool by producing affordable pieces, but it can also increase transfer, filtering and reputation work. The direction of the adjustment depends on the market and intended use.

Registration status is another component. The valuer should reconcile seller records, RIR data, corporate names, legacy status, contractual coverage and pending requests. A block associated with a dissolved predecessor may still be recoverable through documented succession, but the time and uncertainty deserve adjustment. A disputed block should not be valued as clean inventory merely because it appears in an internal database.

Operational control includes account access, reverse-DNS authority and routing-security capability where available. A transfer that updates the registrant but leaves old credentials, route-origin authorisations or nameserver dependencies unresolved is incomplete from the buyer's perspective. These elements may be services around the asset rather than separate assets, but their condition affects value.

Reputation is economically observable even if it is not a registry title question. Address ranges can appear on blocklists, geolocation databases, fraud systems and allowlists. Remediation takes time and may fail. A block suitable for backbone infrastructure may be unsuitable for email-intensive service. The valuer should use evidence appropriate to the buyer's expected market-entity use rather than one generic “clean” label.

Encumbrances include leases, customer assignments, contractual promises, security interests, court restrictions and transfer locks. A seller may control the registration but not have immediate vacant use. Conversely, a portfolio of enforceable leases can add an income stream while reducing flexibility. The acquired asset and associated contracts must be modelled consistently.

The bundle also contains institutional obligations. Annual fees, accurate-directory duties, policy compliance and renewal conditions are not incidental. They are the costs and constraints of maintaining the position. The balance-sheet asset is net of this context, not evidence that the holder has escaped it.

Transfer friction is part of fair value

If two address blocks are technically identical but one can be transferred in weeks with complete documents and the other requires a year of succession research, they do not present the same acquisition risk. Transfer friction affects when the buyer can deploy, monetise or finance the resource and whether closing will occur at all.

The appraisal should separate ordinary transaction costs from asset characteristics in accordance with the applicable accounting framework. Yet even where transaction costs are excluded from a fair-value measure, market entities' assumptions about policy restrictions, probability of approval, timing, legal exposure and the condition of the bundle remain relevant. A restriction inherent in the asset is not the same as a fee paid to complete one sale.

Regional policy can affect the buyer population. Inter-regional compatibility, recipient qualification, holding periods, minimum sizes and contractual requirements determine which market entities can bid. A block cannot be valued on the assumption of a global unrestricted auction if the actual position can only move through a narrower lane. The principal market must be evidenced.

Corporate acquisition routes create their own friction. In a share purchase, the registered entity may remain unchanged, but a later integration or merger can require documentation. In an asset purchase, the prefixes must be expressly scheduled and transferred. A reorganisation can qualify for a succession process rather than an arm's-length market transfer. Accounting should reflect the transaction that occurred, while valuation considers what a market entity could do thereafter.

Registry processing time can create a working-capital and integration cost. The buyer may need temporary leases, dual operations or delayed customer migration. If completion is uncertain, escrow and closing conditions allocate risk but do not make it disappear. A contract price contingent on registry approval can provide cleaner evidence than a price paid before any recognised transfer path exists.

The RIR should not be made the appraiser. Its role is to apply policy consistently, authenticate parties and update records. But it can improve valuation reliability by publishing clear procedures, status definitions, aggregate processing times and reasons for refusal. Unexplained discretion becomes a discount imposed on every entity.

This is one place where market development and registry governance align. More predictable evidence requirements do not commodify the address space by themselves. They reduce the premium paid for uncertainty and help ensure that accounting values relate to genuine transferable positions rather than speculative claims.

Useful life is a governance forecast disguised as an accounting estimate

Once the asset is recognised, management must decide whether its useful life is finite or indefinite under the relevant framework. The decision changes reported earnings. A finite-lived asset is amortised over its expected useful life and tested for impairment when indicators arise. An indefinite-lived asset avoids scheduled amortisation but faces periodic impairment testing and annual reconsideration of the life assessment.

Cogent's indefinite-life conclusion rests on its view that no foreseeable limit currently constrains the period of benefit. MongoDB and DigitalOcean use ten-year lives for purchased addresses. Other telecom disclosures have used still different finite periods. The diversity is not necessarily inconsistency. Business models, acquisition dates, expected use, technology plans and accounting judgments differ. It does mean that “the industry uses ten years” or “IPv4 is indefinite” is not an adequate memorandum.

The analysis should begin with technical substitution. IPv6 deployment expands, but it does not make every IPv4-dependent customer, application, network and counterparty disappear at one date. Dual-stack operation can extend demand. Translation and sharing technologies can reduce addresses required per customer while introducing cost and service limitations. The useful life follows expected economic benefit, not a slogan about protocol succession.

Policy risk matters. A transfer market can remain active while registries tighten particular procedures, or it can become easier through inter-regional compatibility. Annual fees and contractual terms can change holding cost. Courts and regulators can alter enforceability. A useful-life model should identify which institutional events would shorten or preserve benefit.

Market behaviour also matters. Rising lease cash flows can support a long life even if ultimate substitution remains expected. Falling sale prices, vacancy, abuse cost or customer resistance can indicate impairment. Management should use a coherent forecast across valuation, strategic planning and public statements. It cannot tell investors that IPv6 will rapidly remove dependence while using perpetual scarcity to avoid amortisation without reconciling the two.

Indefinite does not mean no decline. An asset can be indefinite-lived and still impaired if fair value falls below carrying amount. Conversely, a finite ten-year life does not forecast that addresses become worthless on the final day; it allocates depreciable value over the expected benefit period, with residual value considered where appropriate.

Auditors should test the life as an estimate with significant assumptions. Evidence includes customer protocol mix, address utilisation, lease renewal, market transactions, infrastructure road maps, policy developments and impairment outcomes. A board-approved label unsupported by operating data is not enough.

Tax balances reveal the allocation's second-order effects

Book fair value and tax basis often differ in a business combination. The accounting asset may be stepped up to acquisition-date fair value while the tax basis carries over, follows an agreed asset allocation or receives different amortisation treatment. The resulting temporary difference can create a deferred tax liability. That liability changes recognised net assets and therefore goodwill or bargain purchase gain.

Cogent's acquisition disclosures make the interaction visible. The IPv4 recognition occurred alongside an increase in net deferred tax liability among measurement-period adjustments. Uniti's Windstream allocation similarly included a substantial deferred tax liability and said appraisal changes involved related tax effects. Neither filing supplies a universal tax formula for IPv4. They show that tax facts must be resolved as part of the allocation.

The effect can appear circular to a non-accountant. Recognising a larger intangible asset can create a larger deferred tax liability; the liability reduces net identifiable assets; the residual goodwill can rise relative to a no-tax calculation, or a bargain purchase gain can change. The exact result depends on transaction structure and jurisdiction. This is why valuation specialists, tax teams and financial-reporting teams cannot work in isolation.

The purchase agreement can also allocate consideration for tax purposes. In a United States applicable asset acquisition, buyer and seller may report classes on Form 8594 using the residual method. A whole-company share purchase can produce a different tax basis from an asset purchase unless an election or local rule changes treatment. An address value accepted for financial reporting is evidence but not automatically the tax allocation.

Deferred tax is not cash tax. It records expected future tax consequences of recovering or settling book amounts under enacted rules. Cash effects depend on amortisation, impairment, sale, jurisdiction, elections and taxable income. Public analysis should not convert a deferred tax liability into a claim that an authority has declared absolute property ownership.

Still, the tax line raises accountability. A high address appraisal can enlarge assets and a related deferred tax liability while changing goodwill. Management should explain method, basis assumptions and sensitivity when material. Investors need to know whether the amount creates future deductions, no deduction, recapture or taxable gain on disposal.

The lesson extends to diligence. Buyers should map each prefix to the legal owner, accounting holder, tax owner and registry holder. Those can diverge after reorganisations. A consolidated balance sheet does not prove that the selling subsidiary owns the transfer position; a registry record does not prove tax basis; a purchase agreement does not by itself complete registry recognition. The reconciliation is the asset.

Auditors must test the estimate, not bless the narrative

IPv4 valuations contain several features associated with estimation risk: a thin and partly confidential market, heterogeneous blocks, management-selected adjustments, long useful-life forecasts, uncertain policy and possible material effects on goodwill or bargain purchase. Novelty increases the need for evidence. It does not excuse a round number.

PCAOB AS 2501 requires auditors of relevant issuers to obtain sufficient appropriate evidence for accounting estimates, including fair value. The auditor can test the company's process, develop an independent expectation, evaluate subsequent events or combine approaches. The standard directs attention to methods, data, significant assumptions, external-source reliability and management bias.

For an IPv4 asset, the existence procedure should begin below the valuation model. The auditor needs the complete prefix schedule, RIR records, transfer documents, corporate authority, subsequent registration state and evidence of control. Sampling addresses without reconciling prefix totals risks counting reserved, customer-assigned, duplicate or non-transferable space.

Rights and obligations require legal and registry evidence. Does the acquired entity have the right asserted? Did a successor receive it? Are agreements, disputes, leases or liens omitted? Did the transaction close subject to RIR approval, and was that condition met? An auditor need not decide global property law to test whether the company obtained the bounded rights it recorded.

Valuation testing should reconstruct comparable selection and adjustments. Are prices completed trades or offers? Do dates match market conditions? Are block-size, reputation and policy differences supported? If a broker supplied data, what controls establish completeness and independence? A confidential database can be legitimate evidence, but “market data” is not a substitute for provenance.

The income model should reconcile to budgets and actual operating metrics. Lease rates, utilisation, churn, abuse cost, registry fees, IPv6 migration and contributory assets need support. Avoided-cost models should use realistic alternatives rather than the most expensive imaginable replacement. Sensitivity should reveal which assumptions drive the amount.

Subsequent transactions can corroborate or contradict the estimate. A sale soon after acquisition at a very different price deserves investigation. Lease performance below forecast can indicate impairment or an optimistic original model. Later evidence is not automatically backdated, but it can expose whether management's acquisition-date assumptions were reasonable.

Finally, the auditor should test presentation. Calling the asset simply “IP addresses” may be too broad if the material value is a specific IPv4 portfolio. Users need the amount, method, useful life, impairment policy and uncertainty necessary to understand the line. Novel material assets deserve more than a generic intangible footnote.

Registry non-property language still has a legitimate purpose

Accounting recognition can tempt holders to demand that registries abandon every statement that number resources are not property. That would be a mistake. RIR language protects coordination goals: uniqueness, conservation where still relevant, needs-based issuance, policy compliance and the ability to correct or deregister records under defined conditions. It also prevents an allocation fee from being misread as a sale of sovereign territory in the address space.

APNIC's policy states that delegation and registration do not confer ownership and that globally unique unicast space is licensed for use rather than owned. ARIN says resources are not freely held property while acknowledging a bundle of contractual rights. RIPE policy has used the principle that address space should not be considered freehold property. AFRINIC's service agreement likewise describes number resources as non-property.

These statements are not identical and should not be collapsed into one global doctrine. Legacy resources can sit outside some modern contracts. National courts can characterise interests differently for insolvency, security, tax or remedies. Policies evolve. The safe conclusion is institutional: RIRs deny an unrestricted property model because their function depends on conditional administration.

Purchase accounting can coexist with that function because it values the holder's actual position, restrictions included. A revocable licence can be an asset. A conditional contract can be an asset. A regulated right can be an asset. The fair value should fall if conditions reduce market-entity benefit or transferability. The registry does not have to call the position freehold for the acquirer to recognise it.

The reverse restraint also applies. A company cannot cite a balance-sheet number to force a registry to approve a transfer contrary to published policy. Financial statements report the company's estimate under an accounting framework; they do not amend the registry agreement. If management assumed policy-compliant transferability and that assumption fails, the accounting response may be impairment, disclosure or correction rather than institutional surrender.

A stable vocabulary would help. “IPv4 address intangible asset” can be defined in accounts as the controlled economic benefits arising from specified address registrations and associated rights, subject to RIR policy, contract and operational acceptance. “Transfer” can mean the recognised change in registration rights. “Ownership” can be reserved for jurisdiction-specific legal conclusions. Precision lowers conflict without suppressing value.

The control surface is larger than the registry record

A buyer pays for useful control, not a line of text in one database. The operational surface includes routing, route-origin authorisation, reverse DNS, directory data, abuse contacts, customer assignments, geolocation and reputation systems. Each can lag the acquisition and reduce benefit.

The RIR record is foundational because it identifies the recognised holder and enables important services. But BGP acceptance is decentralised. A prefix with correct registration can still be filtered because of route objects, RPKI state, maximum-prefix controls, reputation or counterpart policy. The purchase allocation should not assume that registry approval guarantees instantaneous productive use.

Conversely, a block can continue routing under the seller's arrangements after legal closing even though the buyer lacks durable administrative control. Temporary packet flow is not proof that the asset was delivered. Integration plans should identify who controls credentials, who can change authorisations, when customers migrate and how conflicting announcements are prevented.

The impact mechanism reaches financial reporting. Delayed use reduces avoided cost or lease revenue. Reputation remediation increases expense. Fragmentation can increase routing and support overhead. Customer contracts can limit redeployment. A block used by an acquired service may be more valuable in combination with that service than in an immediate sale, but the model must reflect the cost of complementary assets.

The same surface creates evidence for impairment. Route utilisation, lease occupancy, incident volume, blacklisting, registry disputes, transfer requests and IPv6 substitution are observable. A holder that records hundreds of millions of dollars should maintain governance metrics proportionate to the carrying amount.

Acquirers also need a post-close control map. One responsible team should reconcile legal schedules, RIR accounts, routing-security credentials, reverse DNS, network configuration, customer use, accounting subledgers and tax basis. If those records disagree, the organisation can neither operate the asset safely nor defend its value.

This is where the phrase “number resource” earns its meaning. The value is not mystical scarcity alone. It is the capacity to coordinate a unique identifier across technical and institutional systems for productive use. Acquisition accounting should reward that demonstrated capacity and discount the absence of it.

A defensible allocation needs an evidence ladder

The first layer is identity and quantity. List every prefix, address count, registry, status, legacy designation and registered organisation. Reconcile the total to the acquisition agreement and the valuation model. Preserve unavailable denominators rather than filling gaps with estimates presented as facts.

The second layer is authority. Assemble corporate succession, board approval, court orders where relevant, seller representations, registry correspondence, agreements and closing conditions. Identify which evidence establishes the seller's ability to convey and which establishes the buyer's recognised position. They are related but not identical.

The third layer is restrictions and encumbrances. Record transfer holds, regional compatibility, leases, customer assignments, liens, disputes, sanctions, annual fees, contractual conditions and technical dependencies. Value the actual bundle after these constraints, not an idealised vacant block.

The fourth layer is market evidence. For each comparable, retain date, region, prefix composition, size, condition, transaction type, price basis and source quality. Explain every adjustment. If a denominator or term is unavailable, say so and reduce weight. Do not manufacture a global average from a selective broker series.

The fifth layer is economic use. Link addresses to customers, services, avoided purchases, lease revenue, infrastructure plans and IPv6 transition. Reconcile cash-flow assumptions to other acquired assets to prevent double counting. State whether value is in use, exchange, lease or a combination.

The sixth layer is accounting consequence. Document identifiability, valuation method, useful life, residual value, impairment unit, deferred tax, goodwill effect and disclosures. Changes during the measurement period should identify the new evidence and its relation to acquisition-date conditions.

The seventh layer is subsequent verification. Track registry completion, deployment, lease performance, sale outcomes, impairment indicators and policy change. A purchase allocation should be replayable years later when an auditor, regulator, investor or board asks why the number was reasonable.

This ladder does not eliminate judgment. It makes judgment reviewable. It also disciplines sellers. A portfolio with clean schedules, authority and operational history should be easier to value than a claim supported only by a large address count.

Number Resource Society should standardise evidence, not price

Number Resource Society has a constructive role if it remains bounded. It can define a voluntary evidence format for transactions and acquisitions: hashed prefix schedules for confidential review, aggregate counts for publication, registry status, transfer lane, legal continuity evidence, encumbrance categories, valuation date, method, comparable adjustments, useful-life assumptions and subsequent outcome.

The format could help acquirers, auditors and RIRs communicate without asking any one institution to exceed its mandate. A registry could attest that a transfer completed under a named policy without attesting fair value. An auditor could verify quantity and rights without publishing confidential prefixes. A valuation specialist could identify which comparable attributes were known. Investors could receive aggregate disclosures with explicit limitations.

NRS should not publish an official per-address price. A benchmark backed by incomplete private data could become a self-reinforcing quote, invite manipulation and ignore quality differences. It should publish distributions only when denominators, transaction types and coverage are adequate, and it should preserve “limited public evidence evidence” as a valid result.

It should not certify ownership. The evidence standard can distinguish registration, contractual coverage, court order, legacy status and operational control. Jurisdiction-specific legal conclusions belong to competent courts and advisers. Tax classification belongs to tax law. Accounting recognition belongs to the reporting framework and responsible management.

It should also avoid becoming a mandatory gate. Participation can improve diligence and comparability, but access to registry services should not depend on buying an NRS valuation. The society's legitimacy would come from reducing information asymmetry, not collecting rent at the transfer boundary.

A useful pilot would compare completed acquisitions with later outcomes. Did the prefixes transfer as assumed? Were valuation discounts associated with delay or reputation? Did finite useful-life assumptions match observed price and utilisation? Did impairment follow policy or technology change? Such evidence would improve future estimates without pretending that one decade predicts the next.

The positive case is therefore institutional modesty. NRS can make a market that already exists more legible, protect confidential evidence and expose uncertainty. It cannot make all addresses equivalent or settle ownership by taxonomy.

The balance sheet has already moved ahead of the vocabulary

The argument over whether IPv4 addresses are “really assets” is now too coarse. Public companies have paid cash for them, assigned hundreds of millions of dollars to them in acquisitions, changed goodwill when appraisals moved, recorded deferred tax consequences, amortised some portfolios, tested others for impairment and financed lease cash flows. Those are observable institutional facts.

The equally coarse response is to say that accounting has proved absolute property. It has not. Acquisition accounting recognises identifiable economic resources and rights under a reporting framework. It can value a conditional, regulated and contract-dependent position. The more conditional the position, the more those conditions should appear in the estimate.

The productive question is therefore not “asset or no asset?” It is “what exactly passed, what can the buyer do with it, what constraints remain, what evidence supports the amount, and how will the estimate be tested?” Cogent-Sprint and Uniti-Windstream show that these questions can move material financial-statement lines. MongoDB, DigitalOcean and DWP show that value is not merely a merger residual.

Registries should respond with clearer evidence and predictable procedure, not by becoming appraisers. Acquirers should respond with prefix-level reconciliation and explicit assumptions, not with a headline unit price. Auditors should test existence, rights, data, methods and bias. Investors should distinguish fair value at one date from a promise of perpetual scarcity.

The institutional settlement is available without agreement on every legal theory. Let RIRs preserve conditional stewardship and uniqueness. Let courts decide disputes within jurisdiction. Let tax authorities classify taxable events under statute. Let accounting measure the bounded economic position acquired. Then require the records to reconcile.

That settlement is less rhetorically satisfying than declaring either property or no property. It is more accurate. The market value is already in the accounts. Governance now has to make the path into those accounts visible, evidence-bounded and auditable.

Sources