Summary

  • No public global tape joins every IPv4 transfer to its consideration, fees, block quality and contractual terms. RIR transfer records are authoritative for specified administrative events, but their published fields do not supply transaction prices.
  • Public price series are compiled mainly by market entities. IPv4.Global publishes marketplace sales reports and price history from its activity; IPv4 Market Group publishes a chart and prices it says it has observed; IPXO publishes average lease prices on its own platform and republishes sale history identified as IPv4.Global data; Prefixx publishes market factors and current context rather than a single universal price.
  • These sources are valuable precisely because they observe trades that outsiders cannot see. Their limitation is coverage. Each venue sees its own customers, block sizes, regions, negotiation styles and disclosure choices, while private direct transactions and unsuccessful negotiations usually remain invisible.
  • Sample construction can move a benchmark without any underlying market-wide change. Per-transaction and per-address weighting answer different questions; listings, bids, signed deals and completed RIR transfers are different events; a package sale, a distressed sale or a clean /24 cannot be treated as interchangeable observations.
  • A contract-ready benchmark needs a defined interest, eligible inputs, block and region buckets, minimum observations, a weighting rule, treatment of fees and currencies, error and revision policies, coverage statistics, conflict disclosures, oversight, complaints and a cessation fallback.
  • Thin markets do not make benchmarking impossible. IOSCO's benchmark principles expressly contemplate low-liquidity markets, but require users to understand when transactions are limited public evidence and when firm bids, offers or expert judgment enter the calculation. Those principles are an analytical model here, not a claim that an IPv4 chart is a regulated financial benchmark.
  • Number Resource Society can make the market more legible by maintaining an open methodology registry, common transaction vocabulary and evidence-preserving conformance reviews. It should not set an official price, compel private disclosure, approve trades or turn a voluntary standard into a market-access gate.

The number looks more authoritative than the market beneath it

An IPv4 price is usually expressed with deceptive simplicity: a currency amount per address. That presentation invites the reader to imagine a homogeneous commodity, a common unit and a deep market in which many comparable trades establish one clearing value. The real market is more fragmented.

A /24 and a /16 do not merely contain different quantities. They appeal to different buyers, carry different total funding requirements and may face different routing, reputation and transfer considerations. A prefix registered in one RIR service region may be transferable to another only under compatible policies and procedures. A clean aggregate may be operationally useful immediately; a block with damaged reputation, disputed authority, fragmented use or difficult geolocation can require time and cost before deployment. A buyer needing one contiguous block is not in the same position as a buyer able to combine many smaller prefixes.

The transaction itself also varies. Some addresses sell through an online marketplace at a displayed price. Some are auctioned. Some are placed by a broker after bilateral negotiation. Some move within a larger corporate acquisition. Some are transferred between related parties. Some are leased, temporarily transferred or used under a service arrangement rather than sold. Some public registry changes reflect mergers or internal reorganisations without a market sale.

Compressing those differences into one line can still be useful. Airlines, metals, property and credit markets all use indices that simplify diverse activity. The simplification becomes legitimate when the method tells users what was included, what was excluded and how unlike observations became comparable. Without that method, a chart is not necessarily false. It is simply easy to ask it a question it was never built to answer.

The governance problem begins when a descriptive chart becomes prescriptive. A broker may publish recent experience to help a seller set expectations. An analyst may reproduce that chart to describe scarcity. An investor may then use it to value a portfolio. A lender may place it in a covenant. A lease may reset rent by reference to it. At each step the apparent authority increases, even though the underlying sample has not changed.

The first discipline is therefore linguistic. A platform average is a platform average. A broker's observed range is a broker's observed range. A registry transfer count is a registry transfer count. None becomes a global benchmark merely because it is widely copied.

The visible compilers are market entities, not a consolidated exchange

The most sustained public sale-price history is associated with IPv4.Global, a Hilco Global marketplace and brokerage. It publishes monthly marketplace reports, prior sales and commentary by block-size group. The reports are direct evidence about activity the firm observes. They can reveal changes inside a meaningful commercial venue, including differences between small, medium and large blocks.

The reports also make their venue perspective visible through language such as "we had" a stated number of transactions or "our monthly average." That is useful disclosure. It tells a careful reader that the count refers to the publisher's activity, not every transfer completed worldwide. The public pages do not present a full benchmark rulebook explaining every eligibility decision, private-sale inclusion rule, weighting choice, fee adjustment, correction threshold and historic restatement.

IPv4 Market Group publishes another long-running view. It describes its own transaction experience, gives self-reported totals for transactions and addresses moved, and presents a price chart and current prices by prefix-size group. This is valuable first-party evidence from a broker that has participated in the market since its early development. It is not an independent census of competitors' transactions.

IPXO publishes average IPv4 lease prices across transactions on the IPXO platform, with filters for time, RIR and mask. That is a more explicit product and population: lease activity on one platform. Its separate price-history analysis identifies the historical sale data it uses as IPv4.Global data and cautions that the information comes from one independent brokerage and may not reveal administrative fees. The attribution is important. Two websites displaying the history do not necessarily create two independent samples.

Prefixx takes a different public approach. It states that no single fixed IPv4 price exists and identifies block size, RIR region, reputation, transfer type, timing and contiguity as price factors. It offers current market context and transaction support rather than presenting its page as a universal daily fixing.

The RIPE NCC's 2025 State of IPv4 report adds an institutional analysis based on registry data and interviews carried out by an external consultancy. It discusses indicative prices and says an estimated market-value series used publicly available information from various sources involved in transfers. The report is informative, but its public explanation does not turn those sources into a complete transaction tape or a reproducible benchmark available at trade level.

These compilers are not interchangeable and should not be ranked by a claim that one has the "true" price. They observe different slices. Governance begins by naming the slice.

The RIR logs answer who and when, not how much

Regional Internet Registries possess the strongest public evidence that specified administrative transfers occurred. Their records are indispensable to market analysis. They do not, however, operate a price-reporting system.

The NRO transfer log format provides a common structure through which RIRs publish resource transfers. APNIC's earlier transfer log specification illustrates the relevant fields clearly: resource type and prefix, source and recipient organisations and economies, source and recipient RIRs, delegation date and transfer date. There is no price, currency, commission, tax, payment date, leaseback, warranty, escrow term or condition of the block.

The RIPE NCC transfer statistics publish original and transferred blocks, parties, country codes, transfer type and date. They distinguish policy transfers from changes connected with business structure. Again, no consideration field appears. ARIN makes cumulative transfer data available through its public statistics directory, but the common log is designed to record resource movement, not market economics.

This division is sensible. An RIR must maintain accurate registration and process transfers under applicable policy. Requiring it to determine the economic price, classify every side agreement or publish confidential consideration would expand its role and could discourage accurate registration. The lack of a price field is not a registry failure. It is a limit on what analysts can infer from registry data.

A transfer log can supply a denominator for some questions: how many listed records meet a defined transfer type, how many prefixes moved, or how much address space those records contain. It cannot supply the denominator for paid arm's-length sales unless the analyst can distinguish sales from reorganisations, related-party movements, gifts and other changes. Nor can it reveal how many private leases occurred without a transfer entry.

Joining a broker's prices to the RIR log does not solve the problem automatically. A marketplace transaction may involve multiple prefixes, complete on a different date from the commercial agreement, or be registered through entries that do not map one-to-one to the invoice. A prefix may be split before transfer. A corporate package may include addresses alongside other assets. Matching by date and size can create confident but incorrect associations.

The public evidence therefore has two complementary layers. RIR records provide broad administrative visibility without prices. Market entities provide narrower economic visibility with prices. A defensible benchmark must state how, if at all, it connects them.

Define the interest before collecting the observations

The first page of a benchmark method should define the economic interest it seeks to measure. "The IPv4 price" is not sufficient.

One possible interest is the cash consideration per address for completed arm's-length permanent transfers of specified prefix sizes in a stated RIR region, excluding transaction fees and taxes, measured at commercial agreement date. Another is the buyer's all-in acquisition cost at RIR completion. A third is the midpoint of firm executable offers for clean, transfer-ready blocks. A fourth is the recurring monthly lease charge for identified prefixes with a stated package of routing and operational services.

These interests can move differently. A sales series is not a lease series multiplied by an assumed yield. A listing series measures seller aspiration or immediate availability, not necessarily execution. A completed-transfer series measures deals that survived diligence and registry processing, but may lag negotiations by weeks or months. An all-in buyer cost includes items that a seller-net series excludes.

The time stamp must also be defined. Agreement date captures the market when price was negotiated. Deposit date captures commitment. Registry completion captures administrative finality. Invoice or cash-settlement date captures payment. A volatile period can place one deal in different months depending on the selected event.

Quality belongs in the definition. If the benchmark claims to measure "clean /24s," clean must have an observable meaning: specified reputation checks, no unresolved registration dispute, transfer eligibility, routeability testing and disclosed geolocation status. A marketing label without a test makes inclusion discretionary.

The rights included in the sale need definition as well. Does consideration cover only the transfer of the recognised registration position? Are RPKI administration, reverse DNS, route objects, transition services, customer leases, receivables or warranties included? A higher package price may reflect services rather than a higher address value.

The method should state what it does not measure. It may exclude legacy-resource disputes, corporate mergers, bankruptcy packages, related-party transfers, temporary transfers and leases. Exclusion is not evidence that those activities are unimportant. It protects the coherence of the chosen interest.

Without a stable definition, changes in composition masquerade as changes in price. A series can rise because it included more small clean blocks this month, not because comparable blocks became more expensive. The method must make that distinction available to the user.

Venue bias is the unavoidable first bias

Every voluntary venue attracts a non-random group of users. A marketplace designed for smaller standardised blocks may see more /24 and /22 transactions than a private broker serving holders of very large estates. A leasing platform may attract operators that value rapid recurring access and holders willing to delegate operational use. A restructuring adviser may see distressed or strategic portfolio sales that never appear on an open listing page.

Geography compounds the effect. Brokers have different registrations, languages, customer networks, banking relationships and familiarity with RIR procedures. A venue strong in ARIN and RIPE NCC transactions may not represent conditions in APNIC or LACNIC. Inter-RIR compatibility can widen the pool for some resources but does not erase regional documentation, demand or operational preferences.

Client selection matters. A provider known for high-touch diligence may observe cleaner blocks and customers willing to pay for certainty. A low-cost listing venue may attract more price-sensitive entities. A firm with a large institutional seller relationship can produce a temporary concentration of supply. None of those samples is wrong; each is conditional on how the venue acquired its flow.

Publication adds another layer. A broker may know the consideration in every completed transaction but publish only standard marketplace sales. Large private deals may be omitted for confidentiality. Unusual prices may be held back because they could identify a party. Failed negotiations, withdrawn listings and bids that never met a reserve may not appear in a sales table. The visible sample can therefore differ from the provider's own full book.

This bias cannot be cured by calling the provider experienced. Experience can improve execution and interpretation, but it does not make one firm's customer population equal to the world market. Nor can several series simply be averaged if their observations overlap. A secondary analyst must check whether one site republishes another's data, whether brokers co-brokered the same deal and whether a registry entry represents the same transaction counted by two intermediaries.

The correct response is disclosure, not dismissal. The publisher should describe the venue, eligible transaction channels, geographic coverage, share of observations from its own platform, treatment of co-brokered deals and any material concentration by customer or seller. If it cannot disclose identities, it can still disclose concentration in bands.

A user then knows what the benchmark represents: not "all IPv4," but a defined window onto IPv4 commerce.

Completion bias hides the negotiations that did not clear

A completed-sales series observes success. It does not observe every price at which buyers refused, sellers withdrew or registry conditions prevented closing.

That can create an optimistic or pessimistic picture depending on the market. If sellers anchor above available bids, only competitively priced inventory clears; completed trades may appear lower than listings. If urgent buyers pay premiums while patient buyers wait, completions may appear higher than the broader willingness to pay. If difficult blocks fail diligence, the observed sample becomes cleaner than offered inventory.

Time to completion changes the sample again. Deals requiring complex corporate authority, inter-RIR coordination, legacy documentation or remediation may close later than simple transfers. A monthly series based on completion can over-represent easy transactions during a period of congestion. The missing deals are not random; difficulty is related to value.

Cancelled transactions are informative even when their prices remain private. A methodology can publish aggregate counts of eligible observations, withdrawals, failed diligence and incomplete registry processing without naming parties. It can state whether the calculated point uses only settled deals or includes signed agreements awaiting administrative completion. It can publish the median lag from price agreement to the selected timestamp if the data support it.

Firm bids and offers can help in a thin period, but only if they are genuinely executable for a defined block and duration. An indicative email, a stale listing and a funded bid are not equal inputs. The method should describe verification, validity windows, duplicate removal and whether the submitter had authority to trade.

This is where a benchmark can become vulnerable to strategic behaviour. A entity with inventory benefits from higher published indications. A buyer benefits from lower ones. If non-transactional submissions enter the calculation, the administrator needs controls against selective submission, wash indications, related-party coordination and repeated quotes unsupported by capacity.

IOSCO's Principles for Financial Benchmarks recognise that a low-liquidity market can still support a benchmark and that firm bids and offers may supplement transactions. They also demand transparency about the hierarchy and expert judgment. Applied by analogy, the lesson for IPv4 is straightforward: missing trades do not justify inventing certainty. The series should mark a period limited public evidence when its input threshold is not met.

Weighting can reverse the apparent direction

Suppose a period contains many small-block transactions and one very large transaction. A simple average per transaction gives the large deal one vote. An address-weighted average lets it dominate because it contains far more units. A median transaction price may largely describe the small-block segment. Each measure can be calculated correctly and tell a different story.

For an operator buying one /24, a median across similarly situated /24 transactions may be more relevant than an address-weighted global figure dominated by a portfolio sale. For an investor valuing millions of addresses, the large-block market may be more relevant, but a single strategic disposal still may not represent the price at which the entire portfolio could liquidate. For a lender, a conservative executable value under a time constraint matters more than a headline average.

Prefix splitting complicates counts. One commercial bargain can become several registry records. Counting records as transactions inflates activity. Combining all transfers between the same parties on one date may wrongly merge separate bargains. The benchmark needs a trade identifier or a documented aggregation rule that does not expose confidential identity.

Package trades create another problem. A buyer may acquire several block sizes for one total consideration. Allocating the package price equally per address assumes away size premiums and discounts. Allocating by an external curve uses the benchmark to create the observations that then support the benchmark. The circularity must be avoided or disclosed.

Outlier rules can also change direction. A fixed statistical trim may remove a genuine market break. A discretionary trim may remove an inconvenient trade. The method should say whether it winsorises, excludes for quality reasons, separates a special situation or publishes the observation with a flag. It should record who made the decision and why.

No single weighting rule is universally correct. The answer is a family of measures tied to uses: transaction median by prefix bucket, address-weighted mean within the bucket, observation count, address volume and dispersion. Publishing only one point conceals the composition needed to judge it.

Coverage should accompany every point. If a large-block bucket contains too few independent transactions, it should say so. Carrying forward last month's value without a stale flag turns absence into false stability.

A price per address is not a quality-adjusted price

Two prefixes of equal size can impose different costs from the first day of use. Reputation systems may associate addresses with spam, fraud, proxies or other prior activity. Geolocation providers may place the range in the wrong country. Reverse DNS may need delegation. Route objects and RPKI authorisations may need change. The block may have a history of fragmented announcements or operational claims that require resolution.

The public market sources recognise several of these differences. Prefixx identifies reputation, RIR region, transfer type, timing and contiguity as price factors. IPv4.Global distinguishes block-size segments and offers diligence and transition services. The existence of those services is itself evidence that nominal address count does not describe the whole bargain.

A benchmark can handle quality in three ways. It can narrow eligibility to a defined standard, stratify observations by quality attributes, or adjust price with a model. The first is clearest but covers less market activity. The second requires enough observations in each category. The third reaches broader coverage but introduces judgment and model risk.

Whatever choice is made, the benchmark should not silently mix remedied and unremedied blocks. If the transaction price includes a warranty, indemnity, replacement promise or remediation service, the method should say whether the value of that protection is part of the measured interest. A low bare price plus an expensive service package may cost more than a higher clean-block price.

Contiguity needs care. A buyer that requires a single aggregate may discount a collection of fragments because routing and management become more complex. Another buyer may prefer smaller units for deployment. "Large blocks are cheaper" is not a law; it is an observed relationship that depends on funding, buyer population and available inventory.

Transfer readiness is another quality. A seller with complete authority documents and a block already eligible for the intended path offers lower execution risk than a seller still resolving corporate succession. A completed-sales benchmark selects for deals that eventually cleared, but the time and professional cost of reaching completion may not appear in the stated price.

A credible series therefore publishes an attribute table beside its price. It does not need to reveal the prefix or party. It needs to reveal enough composition for a user to know whether the sample resembles the resource being valued.

Region, currency, fees and taxes cannot remain footnotes

A price quoted in dollars per address may have been negotiated in euros, pounds or another currency. Converting at agreement date, invoice date or publication date can produce different results. The benchmark must select a source and timestamp for foreign exchange and preserve the original currency for audit.

Fees also change perspective. The seller may pay a broker commission from gross proceeds. The buyer may pay platform, escrow, legal, diligence and RIR charges. A published transaction price may be gross consideration, seller net or buyer all-in. Those are not interchangeable inputs.

Tax treatment is jurisdiction-specific and often private. A benchmark need not estimate every party's tax. It should state that tax is excluded unless the measured interest explicitly includes it. It should not compare a tax-inclusive public auction price with a tax-exclusive private sale without adjustment.

Regional segmentation is not only a currency question. RIR policies define available transfer paths, holding periods, agreements and recipient conditions. The RIPE NCC report notes that regional policies can affect inter-RIR movement and that registry transfer counts include some business-structure changes. A global series that treats every registered movement as a sale will misstate both volume and coverage.

Country codes in registry records also require caution. APNIC's analysis of IP addresses through 2025 explains that RIR country fields can refer to holder headquarters or intended use under different practices. They are not a reliable substitute for the commercial location of demand, the currency of the bargain or the place of deployment.

A contract that references a "global IPv4 price" should specify which regional bucket applies if policies or market conditions diverge. It should also specify the conversion rule, whether fees are included, and what happens when the relevant bucket has limited public evidence observations.

The benchmark administrator should publish both original and normalised values where confidentiality permits. Normalisation enables comparison; original terms make the normalisation contestable. A single converted number without its conventions creates an illusion of precision.

Commercial interest does not disqualify a compiler, but it must be visible

The firms with the best price information are often the firms that broker, list, lease, finance or advise on IPv4 transactions. Excluding every commercially interested administrator would remove much of the available evidence. Pretending the interest does not exist is worse.

A broker may earn more commission when a deal completes. A marketplace benefits from liquidity and from being perceived as the reference venue. A lessor benefits from confidence in recurring lease value. A holder benefits from higher sale benchmarks. An adviser may publish a bullish view while seeking mandates from sellers. None of these facts proves manipulation. Each shapes the controls a user should demand.

The publisher should identify its businesses, how fees are calculated in general form, whether employees or affiliates can hold relevant inventory, whether contributed data come from clients, and whether sales staff can influence benchmark determinations. It should separate marketing commentary from the mechanical result.

Oversight should include people who do not report to the revenue line responsible for the observed transactions. Material inclusion or exclusion decisions should leave a reasoned record. Related-party trades should be excluded, separately labelled or subjected to a defined arm's-length test. Co-brokered activity should not be double counted.

The public output should disclose concentration without breaching confidentiality. Examples include the share of address volume supplied by the largest contributor, the share of transactions from the administrator's own platform, and the number of independent buyers and sellers represented in bands. If one portfolio sale drives a point, users should know that concentration occurred.

Forecasts require an even clearer boundary. A current benchmark should be based on defined current inputs. An outlook is an analyst's judgment about the future. Publishing them on the same chart without distinction lets the forecast borrow the benchmark's authority.

IOSCO's principles focus heavily on administrator governance, conflicts, submitter controls and accountability because benchmark integrity depends on more than arithmetic. An IPv4 methodology can adopt those disciplines voluntarily without claiming regulatory status. Commercial expertise and open governance can coexist.

The minimum method is longer than a chart legend

A contract-ready IPv4 benchmark should publish a method with at least four layers.

The first layer defines scope. It names the administrator, the interest measured, eligible transaction types, prefix buckets, regions, quality standard, currencies, fee treatment, timestamp, publication frequency and intended uses. It states whether the point measures listings, firm quotes, signed agreements, settled cash or completed registration.

The second layer defines inputs. It describes accepted sources, contributor eligibility, verification, duplicate controls, related-party treatment, confidentiality, late data and the hierarchy among executed transactions, committed orders, firm bids and offers, and expert judgment. It states the minimum number of independent observations and any concentration limit required to publish.

The third layer defines calculation. It identifies weighting, aggregation, outlier treatment, package allocation, currency conversion, rounding, stale values, model use and quality adjustments. It gives examples with synthetic data so users can reproduce the arithmetic without exposing a real party.

The fourth layer defines governance. It sets oversight, conflict controls, record retention, complaints, independent review, correction thresholds, revision windows, methodology consultation, emergency publication, cessation and contractual fallback.

Every published point should then carry a compact coverage statement: number of eligible transactions, independent buyers and sellers in non-identifying bands, total addresses, prefix-size distribution, region distribution, share of executed versus non-transactional inputs, concentration, dispersion, stale status and any expert adjustment.

Confidentiality is not a reason to omit all coverage. Small cells can be suppressed or banded. Publication can be delayed. A trusted reviewer can inspect protected evidence and publish an assurance conclusion. What cannot be done is to claim representativeness while refusing to describe the sample at any level.

The method must also have a version. If a publisher changes block buckets, adds private sales or moves from transaction weighting to address weighting, the historic series may break. Users need the effective date, rationale, impact analysis and whether history was restated.

This is more work than drawing a line. That is precisely why a line should not govern cash flows until the work is done.

Revision policy is part of the price, not an editorial detail

IPv4 transactions can be reported after the period closes. A broker may discover that a package was incorrectly allocated. A registry transfer can complete after a signed deal fails or changes size. Currency or fee treatment may be corrected. A benchmark that never revises may remain wrong; one that revises without rules can rewrite contractual history.

The administrator should distinguish routine late-data updates, corrections of error and methodological restatements. Each needs a threshold and window. A small correction below the publication precision may be logged without changing the headline. A material error should trigger a revised point, notice and machine-readable record of old and new values.

Contracts must say which value controls. If rent resets on publication day, does a later correction alter invoices already paid? Does the corrected value apply prospectively? Can either party challenge a calculation? The benchmark administrator cannot answer every commercial question, but its revision policy gives the contract a stable input.

Historic comparability matters too. When methodology changes, the administrator can run old and new methods in parallel for a period, restate history where evidence permits, or mark a series break. It should not splice two methods into one continuous line without disclosure.

Cessation deserves advance planning. A broker can exit the business, stop publishing, be acquired or lose enough observations to make the series unrepresentative. A website disappearing is not an acceptable contractual event definition. The administrator should publish a cessation policy and notice where feasible.

Users need fallbacks that do not create circular discretion. A contract might move to a second independently specified series, use a panel with documented quotes, obtain a qualified valuation, or carry forward the last value only for a short defined interval before renegotiation. "A reasonable market price chosen by the lessor" is not a neutral fallback.

IOSCO's transition principle asks benchmark users to plan for material change or cessation and to include robust fallback provisions. IPv4 contracts should do the same even when the referenced series is outside financial-benchmark regulation.

Revision governance protects both sides. It prevents an obvious error from becoming permanent and prevents a publisher or contracting party from changing history when the result becomes inconvenient.

Thin periods should produce less confidence, not more modeling

The temptation in an illiquid bucket is to fill the gap. The previous value can be carried forward, a neighbouring prefix size can be adjusted, a listing midpoint can replace a trade, or an expert can estimate where a deal would clear. Each technique can be reasonable if disclosed. None is equivalent to an executed observation.

The input hierarchy should place verified arm's-length transactions closest to the measured interest at the top. Firm executable bids and offers can follow. Indicative quotes, comparable buckets and models can sit below. The published result should identify how far down the hierarchy it travelled.

Minimum data requirements should include independence, not merely count. Ten prefixes sold in one package to one buyer are not ten independent price discoveries. Five submissions from affiliates are not five market views. The administrator should test contributor concentration and common control.

Expert judgment needs boundaries. The decision maker should record the observable facts, adjustment and reason. Repeated judgment in one direction should be reviewed. Sales staff should not be able to alter a point informally because they believe the market feels stronger.

Some periods should have no benchmark. "Limited public evidence eligible data" is useful information. It warns a contracting party that the chosen index may not suit frequent resets. Publishing a precise modeled point can encourage leverage and contractual dependence unsupported by the market.

Dispersion matters alongside the central estimate. A wide range may reflect quality, region or genuinely uncertain clearing conditions. A narrow point derived from a model does not eliminate that uncertainty. Confidence labels should respond to sample size, independence, concentration and input type.

The benchmark can publish a hierarchy flag, such as transaction-supported, quote-supported or model-supported, without pretending those categories have equal strength. Users can then set their own rules. A lender may accept only transaction-supported observations. A valuation may use model-supported points with a wider range. A lease may defer a reset when data are limited public evidence.

Market maturity is shown by the ability to admit uncertainty. The benchmark's legitimacy should not depend on publishing every month at any cost.

Contracts must not outsource judgment to an undefined webpage

An IPv4 purchase agreement can use a benchmark to adjust price between signing and completion. A lease can index renewal rent. A financing agreement can test collateral coverage. A portfolio manager can mark holdings. Each use requires a different fit.

The contract should name the benchmark administrator, series, version, region, block bucket, currency, fee basis and observation date. It should archive or attach the applicable method. Referring only to "the prevailing IPv4 price" or a homepage leaves the decisive term open.

The contract should state how the subject block maps to the series. A /19 should not automatically use a /24 point. A damaged block should not receive a clean-block benchmark without an agreed adjustment. A package should not be split by unilateral judgment after signing.

Publication timing matters. If the series appears after month end, the contract should identify the first publication or final revised value. It should handle holidays, delayed publication, limited public evidence data, material method change and cessation. It should also state whether a correction reopens settled amounts.

Conflict protection is especially important when one contracting party or its broker contributes data to the benchmark. The contract can exclude the parties' own transaction from a reset, require independent calculation, cap the effect of non-transactional inputs or allow an evidence-based challenge.

A benchmark should not be the sole proof of fair value. The subject resource has its own reputation, authority, encumbrances, services and timing. A benchmark supplies context; diligence supplies the transaction-specific facts. The more unusual the block or package, the less weight a broad series should carry.

Valuation users should also distinguish orderly-exit value from forced-sale value and portfolio value from marginal trade value. Selling one /24 at the displayed market is not evidence that millions of addresses can liquidate at the same point without moving the market.

The strongest contractual clause contains a fallback ladder and an obligation to exchange supporting evidence. It makes the benchmark useful without turning a publisher into an unappointed arbitrator.

Investors need coverage and liquidity, not a heroic line

An investor evaluating an IPv4 strategy asks more than whether the latest point rose or fell. It asks how much inventory could transact near that point, how long execution takes, what costs sit outside the quote, how value differs by block and region, and what happens when IPv6 or policy changes demand.

A price history from one venue can support scenario analysis if its limits are preserved. It can show that the venue observed changing conditions. It cannot prove that an unobserved portfolio shared the same return. Backtests that apply one small-block series to every historical holding create artificial performance.

Liquidity should be measured separately from price. Registry logs can help describe administrative transfer activity, but, as the RIPE NCC report warns, those logs can include business-structure changes. Broker counts describe their own activity. Neither alone supplies the total value of arm's-length global sales.

Bid-ask spreads, time on market, withdrawal rates and concentration would improve investment analysis, but public denominators are incomplete. The responsible conclusion is not that liquidity is high or low by a universal percentage. It is that users need venue-specific evidence and stress assumptions.

Valuation should use ranges and scenarios. A base case can use comparable executed observations. A downside case can allow for time, fees, reputation remediation and market impact. An upside case can consider scarcity and strategic demand without treating advocacy as a realised bid. Each scenario should identify the source and sensitivity.

The same discipline applies to collateral. A lender should not advance against a headline per-address price without testing enforceability, registry completion, buyer depth and sale costs. A benchmark can trigger review, but it does not deliver a buyer after default.

There is no public denominator for the share of all IPv4 bargains captured by any cited price publisher. There is no public global count of failed sales, private consideration or lease terms. This analysis does not estimate those rates. It treats the absence as a limit on confidence, not a licence to fill the gap.

Number Resource Society can govern the method without governing the price

Number Resource Society has a credible positive role if it resists the attraction of an official number. Its public materials advocate transparency, accountability and broader operator participation in number-resource governance. Those aims support an open price-methodology layer, but they do not establish that NRS currently operates an audited benchmark or possesses a complete transaction dataset.

The first useful contribution would be a public vocabulary. It could define permanent transfer, temporary transfer, lease, listing, firm bid, signed transaction, registry completion, gross consideration, seller net, buyer all-in cost, prefix bucket, quality attributes and related-party deal. Common terms would let publishers disagree in the open rather than present unlike measures under the same label.

The second contribution would be a methodology registry. Any broker, platform, analyst or academic group could deposit a versioned method, coverage template, conflict statement and revision history. NRS could check whether the publisher followed its own stated method and publish the review evidence. It would not decide whether the price is high or low.

The third contribution would be a confidential contribution protocol. Parties could submit protected transaction facts to independent custodians using a common format. Only aggregate, non-identifying coverage and results would be public. Participation should remain voluntary unless a competent legal authority imposes a specific duty. Absence from the dataset must not be treated as suspect.

The fourth contribution would be interoperability. Multiple administrators could publish competing series using comparable coverage labels. Users could select the series that fits their contract and detect overlap. A decentralised system is more resilient than one official fixing and makes methodological competition possible.

The safeguards are as important as the service. NRS should disclose funding and affiliations, separate advocacy from review, include buyer, seller, operator and independent measurement perspectives, publish complaints and corrections, and permit external audit. It should not require an NRS benchmark for an RIR transfer, certify a private trade as morally acceptable or demand private contracts as the price of market participation.

This role is positive because it creates public infrastructure from facts the market already produces. It is evidence-bounded because conformance to a method does not prove total coverage, fair value for every block or legal status in every jurisdiction.

A practical assurance label would expose rather than hide uncertainty

A useful conformance label could have five dimensions.

"Method disclosed" would mean the measured interest, inputs, calculation, quality rules and timestamp are public. "Coverage disclosed" would mean every point includes observation count, volume, independence, concentration and input hierarchy in non-identifying form. "Revisions controlled" would mean errors, late data, method changes and cessation follow published rules. "Interests disclosed" would mean the administrator and contributors identify relevant commercial roles and controls. "Calculation reviewed" would mean an independent reviewer tested a sample against the stated method.

The label should not say "correct price." Two compliant benchmarks can differ because they measure different regions, block sizes or venues. The purpose is to make the difference intelligible.

Assurance should be periodic and event-driven. A material methodology change, new data source, acquisition of the administrator or sharp concentration shift should trigger review. The reviewer should publish exceptions rather than quietly negotiate them away.

Open test files can improve reproducibility. The administrator can publish synthetic observations and expected calculations. Users can verify weighting, currency, outliers and revisions without access to protected trades. A cryptographic commitment to the protected input set could later show that the dataset was not silently changed, though it would not prove the truth of each submission.

Complaints should be available to contributors and users. A party can challenge duplicate counting, misclassification, an unannounced change or a calculation error. The process should publish outcomes while protecting transaction confidentiality.

Costs should be proportionate. A small broker publishing occasional commentary should not need the same control environment as a series embedded in large contracts. The label can identify intended use: commentary, valuation support or contractual reference. Responsibility rises with dependence.

The market would then have a way to distinguish a useful opinion from a governed reference without suppressing either. That is better than declaring one price official and better than accepting every chart as equivalent.

What the evidence supports, and what remains unavailable

The evidence supports four firm conclusions.

First, market entities publish meaningful IPv4 price information. IPv4.Global, IPv4 Market Group, IPXO and Prefixx each provide direct evidence from their roles. Their observations should not be dismissed merely because they are commercial.

Second, their public samples and methods differ. IPXO explicitly identifies its platform population for lease averages and attributes sale-history data to IPv4.Global. IPv4.Global's reports speak from its marketplace activity. IPv4 Market Group describes its own experience. Prefixx emphasises transaction-specific factors. Copying the same underlying data across sites does not create independent confirmation.

Third, RIR transfer data cannot supply missing prices. The published formats provide resource, party, region, type and date fields, not consideration. Transfer counts can contextualise activity only after analysts account for business-structure changes, record splitting and other non-price events.

Fourth, established benchmark-governance principles offer a practical design for methodology, input quality, conflicts, revisions, complaints, review and transition. They can be adopted voluntarily and proportionately without claiming that IPv4 is a regulated financial instrument.

Important denominators remain unavailable. No cited public source establishes the number or value of all arm's-length IPv4 sales worldwide. No source establishes what share each broker or platform captures. No complete public dataset joins every transfer to price, fees, quality, negotiation date and contract type. No public evidence supports a universal error rate for current price charts or a prevalence rate for manipulation.

Those gaps do not prevent price discovery. They prevent overclaiming. A transparent partial benchmark can be highly useful. An opaque claim to universality is fragile.

The benchmark earns authority after dependence, not before it

IPv4 commerce has developed enough public history to make prices discussable but not enough shared reporting to make one number self-proving. The market's practical experts see real transactions. The registries see real administrative changes. Analysts see public series and broader trends. Each holds part of the picture.

The next step is not a compulsory global price. It is a disciplined account of what every published price means. Method, coverage, revision and interest disclosure turn a number from assertion into evidence. They allow competing series to coexist and allow users to choose one suited to a particular block, region and contract.

Contracts should move last. Once a series changes payments automatically, its administrator exercises practical power over parties that may never have contributed data or agreed to a later methodology change. That power needs a rulebook, oversight and a fallback.

Investors should demand the same discipline. A line without address volume, independent observations and dispersion cannot carry a portfolio valuation by itself. Registries should keep publishing accurate transfer facts without being converted into price authorities. Brokers should be able to publish expertise while naming their sample. Analysts should preserve attribution and avoid counting republished data as independent.

Number Resource Society can help make those roles legible. Its strongest contribution is not to announce what an address is worth. It is to make every claimed reference price answer four questions: how was it made, what did it cover, what changed after publication, and who had an interest in the result?

Until a series can answer them, it is market information, not market law.

Sources