A buyer, a seller, a broker and a registry officer can look at the same IPv4 block and see four different things. The buyer sees continuity: customers that must remain reachable, cloud accounts that need stable addresses, access lists that cannot be rebuilt quickly, email reputation that has to survive a migration, and a way to postpone a redesign that would be technically correct but commercially expensive. The seller sees capital trapped inside an old registration record, sometimes attached to a network that no longer needs the same volume of addresses, sometimes attached to a corporate history that is more complicated than the current balance sheet suggests. The broker sees a spread of missing facts: who really controls the block, which documents will satisfy the registry, whether the route history is clean, whether the buyer has a deadline, whether the seller is distressed, whether the registry is likely to ask questions, and whether a rival intermediary has already failed to close the same file. The registry officer sees something narrower but more consequential: a ledger entry whose alteration must not corrupt uniqueness, reward false authority, damage operational services, or turn an administrative update into a public institutional dispute.
The transaction looks like brokerage. It is also governance.
IPv4 brokers exist because this market is expensive to search, hard to verify and risky to close. Scarcity has made address space economically material, while the registry system still looks more like an administrative membership framework than a fully standardised transfer infrastructure. A buyer may know the going price for a clean prefix and still not know whether the seller has authority. A seller may know that it has surplus space and still not know whether a registry will accept the proposed path. A lessee may receive routed service and still not know whether a later dispute will contaminate reverse DNS, RPKI, abuse contacts or reputation. A registry may insist that it only maintains records and still become the place where commercial uncertainty is converted into delay, leverage or veto.
That is why the broker in the IPv4 market is not merely a salesperson. The broker reduces search costs, but that is the lowest layer of the job. The harder work is to translate registry uncertainty into transaction design: what evidence to collect, how to test authority, whether to stage closing, how to handle post-close operations, which conflicts to disclose, and when an apparently routine record update is likely to become a discretionary review. The broker does not own the registry. But when the registry layer is fragile, the broker may become the actor with the most practical knowledge about timing, documentary expectations, informal risk and failure modes.
AFRINIC makes the issue sharper because its registry layer has not been quiet background infrastructure. AFRINIC is the Regional Internet Registry for Africa and parts of the Indian Ocean. Its public service surface includes IPv4, IPv6 and autonomous system number administration, WHOIS and RDAP publication, reverse DNS, the Internet Routing Registry and RPKI. Those functions are not decorative. They are the points at which a brokered transaction becomes operationally real. A buyer who cannot update route-origin authorisations, a seller whose contacts remain exposed, a network whose reverse DNS is stranded, or a prefix whose registration record is distrusted has not received the economic object it thought it was buying.
The recent AFRINIC record also supplies uncomfortable evidence. Public reporting in 2019 by MyBroadband and KrebsOnSecurity described alleged manipulation and sale of African IPv4 space tied to dormant or weakly monitored records. Litigation involving Cloud Innovation turned high-value IPv4 holdings into a prolonged test of resource review, contract interpretation and court remedy. Court-appointed receivership, board discontinuity and disrupted election processes in 2023-25 showed that the registry function can become entangled with legal continuity. Later efforts to reconstitute governance do not erase the market lesson: for counterparties, the registry layer itself can become a risk surface.
The question is not whether brokers are good or bad. That frame is too moralistic and too simple. Brokers will appear wherever scarcity, private information, registry process and counterparty distrust make direct exchange costly. The serious question is how to make intermediation legible: who represents whom, what conflicts exist, which authority chain is being relied on, which evidence supports a requested registry action, who operates the block during and after closing, what happens if a document proves false, and how the file can be audited later. A fragile ledger does not eliminate brokers. It makes their role more powerful.
Brokers exist because direct exchange is too expensive
A perfectly transparent IPv4 market would need few brokers. Holders would publish reliable supply. Buyers would know what they needed. Registry records would show current authority. Transfers would follow predictable stages. Disputes would be noted precisely without poisoning unrelated operations. Price would reflect prefix size, reputation, routeability, terms and urgency, not the private knowledge of whoever had spoken to the right seller, buyer or registry-facing representative first.
That is not the market that exists. IPv4 address space is finite, operationally embedded and unevenly distributed. Some holders use every address intensively. Others carry allocations that reflect older network plans, corporate reorganisations, acquisitions, dissolved subsidiaries, dormant services or addresses assigned when the Internet's economics were different. Some blocks have clean routing history, current contacts, usable reverse DNS and ordinary abuse records. Others carry spam history, stale geolocation, uncertain corporate authority, old sub-allocations, or a chain of route announcements that requires explanation before a serious buyer will touch them.
The seller's side is not transparent either. A holder with surplus space may not want to advertise supply because publicity can invite speculative buyers, customer concern, creditor attention, internal politics or registry questions. A company may want to test value without committing to a sale. A liquidator, successor entity or board may not know what authority the registry will require. A technical contact may know the block is idle but lack the corporate authority to transact. A seller that seems easy to find may still be hard to verify.
The buyer's side is equally costly. A buyer may know it needs addresses but not whether it should buy, lease, renumber, use provider address space, rely more heavily on NAT, acquire a company, or wait. It may not know which RIR policy applies, whether a cross-registry path is feasible, whether a prefix can be routed at a practical size, whether the seller's chain of authority is coherent, whether abuse reputation can be cleaned, or whether a low price is a sign of hidden trouble. The buyer is not only searching for numbers. It is searching for usable continuity.
Scarcity turns those frictions into economic rents. A person who knows where idle supply sits, which holders are under pressure, which documents usually satisfy a registry, which buyers can pay quickly, which blocks carry invisible risk and which closing structures tend to fail has valuable information. The broker monetises that information by reducing search time, packaging evidence and making both parties believe that a transaction can actually close.
This does not make brokerage parasitic. Search is a real cost. In a fragmented market, a competent broker can make transactions possible that would otherwise fail under suspicion or complexity. A seller may not want to negotiate with dozens of speculative buyers. A buyer may not want to reveal demand to competitors or suppliers. A network operator may need confidentiality because a public purchase would signal expansion, distress or customer migration. A broker can screen seriousness, reduce noise and protect negotiations from premature disclosure.
Yet search-cost reduction is also where power begins. The broker sees the order book before the market does. It learns which buyers are desperate, which sellers are naive, which registry files are slow, which blocks have reputation discounts, and which mandates are weak. In exchange-traded markets, standardised disclosure and settlement infrastructure reduce the private value of this knowledge. In IPv4, the value remains high because the market is episodic, bilateral and dependent on registry recognition. AFRINIC-related space magnifies the point: the broker is not only finding supply. It is sorting between files that may survive registry scrutiny and files that may become institutional trouble.
The broker as transaction architect
The old image of a broker is simple: find a willing seller, find a willing buyer, negotiate a price and collect a commission. That image understates the IPv4 broker's real function. In this market the broker often designs the transaction before the transaction can be sold.
Transaction architecture begins with the asset. An IPv4 block is not useful merely because it appears in a registry database. It must be globally unique, routeable at a practical prefix length, reputationally acceptable, correctly recorded, supported by current contacts, compatible with the buyer's intended use and capable of surviving the chosen transfer or lease path. The broker must ask whether the block is better sold outright, leased with operational controls, bundled with managed IP services, staged through escrow, separated into smaller routed units, kept under the seller's account for a transition period, or rejected because the evidence chain is too weak.
The next layer is authority. The named holder may be a company that changed its name, merged, dissolved, moved jurisdiction, lost records or delegated operations to a different entity. The person speaking for the seller may have technical access but not corporate authority. A reseller may have a contract but not enough registry-recognised power. A broker that fails here can turn a commercial negotiation into a false-authority problem. A careful broker asks for board authority, signatory evidence, historical records, registry contacts, account control, route-history explanation, abuse-contact control and a chain from legal entity to operational use.
The third layer is registry process. Each RIR has its own vocabulary and habits: allocation, assignment, registration, membership, transfer, resource review, utilisation, need, policy compliance and service status. Some of this is written. Much of it is institutional practice. The broker must know which part of the file is a formal requirement and which part is likely to become discretionary questioning. A seller may have a clean block but a weak account. A buyer may have a legitimate need but an intended use that attracts scrutiny. A registry may accept a holder change but delay related reverse-DNS or RPKI maintenance. The broker's value lies in anticipating such complications before money moves.
The fourth layer is operating continuity. Closing is not the end of risk. The buyer wants to announce the prefix, update reverse DNS, create route-origin authorisations, maintain abuse contacts, preserve geolocation quality, satisfy cloud and carrier onboarding, and protect customers from disruption. The seller wants payment certainty, release and limited downstream liability. A thin broker treats these as post-close details. A serious broker treats them as part of the transaction design.
Registry uncertainty changes the broker's identity. Where records are reliable and process is predictable, the broker can remain a market intermediary. Where the registry layer is fragile, the broker becomes an interpreter of administrative risk. It translates the buyer's demand into a form that the registry may accept, translates the seller's record into evidence the buyer can trust, and translates institutional uncertainty into price, timing, conditions, indemnities, holdbacks and operational safeguards.
The danger is that this translation may become opaque. If only the broker understands the registry file, the broker can shape each party's perception of risk. It can overstate urgency, hide conflicts, privilege one counterparty, steer a buyer toward inventory in which it has an undisclosed interest, or present registry discretion as unavoidable when it is merely anticipated. Transaction architecture is necessary. It must not become private law written by an intermediary.
AFRINIC turns evidence into a market service
Scarcity gives evidence value. When a resource is abundant, imperfect records are an administrative nuisance. When the resource is scarce, imperfect records become market constraints. IPv4 crossed that line years ago. AFRINIC's exhaustion regime, including its soft-landing limits, means that new demand cannot simply be satisfied by ordinary large allocations from a deep free pool. Demand must be met through existing holders, transfers, leases, provider inventories, conservation or design compromises. A record that once looked like administrative residue may now support a material transaction.
The economic object being traded is not a metaphysical ownership theory. It is operational reliance. IPv4 addresses support servers, customers, data centres, cloud platforms, security rules, payment flows, access lists, email reputation, logging, routing and network reachability. Whatever label is used in the registry context - allocation, assignment, registration, right of use or resource holding - counterparties are paying for the ability to use scarce public identifiers without being surprised by the recordkeeper. The market is not buying philosophy. It is buying a reduced probability of disruption.
That makes evidence a product. A broker with a credible file can create value before price is negotiated. The file may show current holder status, historical assignment, corporate continuity, authorised signatories, absence or scope of disputes, route history, abuse remediation, reverse-DNS control, RPKI status, transfer eligibility and the intended post-close operating model. A buyer pays attention because the evidence reduces the probability that a registry, carrier, cloud platform, lender or customer will reject the arrangement later.
AFRINIC's history makes the evidence service unusually important. MyBroadband's 2019 investigation reported that AFRINIC's public database of IP block assignments had been manipulated, with valuable resources associated with dormant or weakly monitored entities moving into market channels. KrebsOnSecurity described the same pattern as a large African IPv4 address heist, focusing on allegations that dormant blocks and internal access had become part of a commercial theft surface. The narrow lesson for broker-market governance is not that every old AFRINIC record is suspect. It is that scarcity turns stale authority into an attack surface.
In such a market, a broker who can identify false authority is valuable. So is a broker who can reconstruct corporate succession, spot route anomalies, detect stale contacts and understand why an apparently quiet block may have an invisible claimant. The evidence work can protect buyers and honest sellers. It can also protect the registry from bad updates. Without such work, many counterparties would lack the expertise to assemble a usable file.
But evidence services are vulnerable to capture. The broker who collects documents may decide what to show, what to omit and what to call immaterial. A seller may rely on the broker to explain the buyer's risk tolerance. A buyer may rely on the broker to explain the seller's authority. The registry may rely on packaged documents without seeing the commercial incentives behind them. Each reliance creates a blind spot.
The governance answer is not to ban broker knowledge. It is to make the evidence trail auditable. A market can accept that brokers prepare files. It should not accept that the file becomes a black box. A registry and the parties should be able to see who provided which document, what authority it supports, what claim it does not support, and who bears responsibility if the evidence proves false. Evidence should be valuable because it is careful, not because it is selectively disclosed.
Information asymmetry is the broker's asset
Information asymmetry is the central economic fact of IPv4 brokerage. Sellers know more about their urgency, authority problems and historical use than buyers do. Buyers know more about their demand, budget, timing pressure and fallback options than sellers do. Brokers often know more than both about current market sentiment, comparable blocks, registry practice, counterparty reputations and which files are likely to stall. The registry knows more than the parties about account status, past communications, internal review concerns and policy interpretation. No actor sees the whole picture.
This matters because IPv4 is not a homogeneous commodity. A /24 with clean history, current contacts, stable route origin, working reverse DNS and no dispute is different from a /24 with the same size but a past spam reputation, stale holder records and uncertain authority. A larger block may be less attractive if it cannot be split cleanly, if it carries route-history problems, or if a registry review could delay recognition. A smaller block may command more confidence if it is clean and operationally ready. Price opacity is therefore not only about quoted numbers. It is about hidden quality.
The broker's informational role is double-edged. The broker can reduce asymmetry by telling the buyer why one block is safer than another, telling the seller what documentation is missing and telling both parties what registry process is likely to require. This is productive brokerage. It converts private confusion into shared diligence.
The same position can be abused. A broker can keep buyer and seller apart so neither learns the other's constraints. It can describe a registry delay as ordinary when it reflects a specific weakness in the file. It can use fear of AFRINIC uncertainty to push a buyer toward more expensive inventory. It can assure a seller that a buyer is committed while using the seller's documents to solicit alternatives. It can obtain a mandate that looks exclusive but actually serves the broker's own inventory strategy. It can turn market opacity into margin.
AFRINIC-related transactions increase the value of private knowledge because a registry under legal or governance stress produces fewer simple signals. Counterparties ask whether staff can act, whether board authority is stable, whether court orders affect service, whether an account review could become broader than expected, and whether other institutions might intervene. The broker who has recent experience can tell a story about timing. That story may be accurate. It may also be self-serving.
Information asymmetry also affects the registry. Registry officers see files, not always the full commercial arrangement. A broker may present a transfer as ordinary holder-to-buyer movement when the underlying relationship is a financing structure, leaseback, option, default remedy, aggregation strategy or undisclosed market-making position. The registry may not need to judge the business model, but it does need to know who has authority to request a change and who will operate the resource. If the broker controls that narrative, registry blind spots become market infrastructure.
The remedy is not full public disclosure of every commercial term. IPv4 transactions often require confidentiality. Prices, customer names, financing terms and deployment plans may be sensitive. The remedy is structured disclosure of authority and conflicts. The registry and counterparties should know whether the broker represents the buyer, the seller, both, neither or itself; whether it has a financial interest in the inventory; whether it may allocate the block to another customer; whether it holds documents only as representative; and whether it will provide operational management after closing. The market does not need every secret. It needs to know who is speaking for whom.
Conflicts travel through mandates
The broker's conflict problem is structural because the broker sits at the intersection of search, evidence, price, timing and registry process. That position creates legitimate value and persistent temptation. A broker can be adviser, introducer, mandate holder, reseller, market maker, documentation consultant, leasing manager, abuse-contact facilitator and post-close operator. Each role may be defensible in isolation. Trouble begins when the parties do not know which role is being performed.
A pure buyer's broker should seek the safest usable space on the best terms for the buyer. A pure seller's broker should maximise value and certainty for the seller. A neutral marketplace should disclose that it is not advising either side. A market maker may buy or control inventory and resell or lease it for its own account. A managed-service provider may care more about continuity and recurring revenue than a one-time commission. A registry-facing consultant may be paid to get a file accepted. These are different incentives, and they should not be hidden behind the same title.
In IPv4 brokerage the lines often blur. A broker may claim to be neutral while earning more if the buyer accepts a specific block. It may advise a seller on price while privately knowing that a buyer is willing to pay materially more. It may discourage direct contact because direct contact would reveal the spread. It may present a registry concern as a reason to use its preferred structure. It may take a mandate from a holder, then use the existence of that mandate to create scarcity in the buyer's mind.
AFRINIC makes these conflicts more consequential because registry risk can be used as an argument for dependence. A broker may say, accurately, that the buyer needs specialist guidance through AFRINIC process. It may also say, less innocently, that only its channel can manage the risk. The difference is hard for a buyer to test. If the registry layer is perceived as unpredictable, the broker's claim to special knowledge becomes a moat.
The broker can also become a conflict channel for the registry. A registry should not favour one intermediary over another, yet in practice some brokers may understand staff expectations better, have a history of successful files, know which language avoids review, or have relationships that make communication easier. This does not require corruption. Institutional familiarity alone can create unequal access. In a scarce market, unequal access is economic power.
Conflicts cannot be eliminated. They can be governed. Engagement terms should state representation, compensation, inventory interest, exclusivity, sub-broker arrangements, confidentiality duties and post-close role. Transaction files should carry relevant conflict disclosure with the registry-facing submission where it affects authority or operations. If a broker is both representing a buyer and supplying inventory from an affiliated pool, the buyer should know. If a broker is paid by both sides, both sides should know. If the broker has a continuing role in routing, reverse DNS or abuse handling, that role should be separated from the transfer mandate.
Registry process can reinforce this discipline without becoming a commercial regulator. AFRINIC does not need to approve brokerage fees or police every advisory duty. It can require clear evidence of authority from the registered holder and the intended recipient. It can ask whether the submitting party is a representative and for whom. It can record the source and scope of authority. It can reject unsigned, ambiguous or recycled mandates. It can maintain audit logs that make later reconstruction possible. The broker's conflict temptation grows when registry process is both powerful and opaque. Making process legible reduces the private value of hidden access.
Chain of custody is the hidden infrastructure
The phrase "chain of custody" sounds legalistic, but in IPv4 markets it is operational infrastructure. It answers a simple question: how did this block move from a recognised holder to the current proposed use, and who had authority at each step? Without a credible chain, a transaction can close commercially while remaining fragile administratively.
The chain begins with historical registration. Who first received the resource, under what category, and for what broad purpose? It continues through corporate changes, contact updates, resource reviews, prior transfers, leases, sub-allocations, route announcements, reverse-DNS delegations, abuse-contact changes, RPKI objects and dispute notations. It also includes negative evidence: no known competing claimant, no forged mandate, no unresolved dissolution, no undisclosed court order, no hidden registry hold. The absence of a problem is still a claim that requires a basis.
Public reporting on alleged African address theft shows why this matters. Scarce IPv4 made dormant and poorly maintained records attractive. Where an organisation had disappeared, merged or stopped paying attention, someone with access or knowledge could attempt to move the resource into a market channel. Even if later reporting disputes details or assigns blame differently, the structural lesson remains: an address ledger with stale authority is a theft surface. Brokers can either repair that surface by demanding proof, or exploit it by treating weak records as cheap supply.
Chain-of-custody dependence also changes the economics of legitimate transactions. A clean chain reduces due-diligence cost, negotiation time and holdback demands. A weak chain reduces price, narrows the buyer pool or forces a broker to design a risk-heavy structure. A block that appears cheap may become expensive after legal review, registry delay, abuse cleanup and operational uncertainty. A block that appears expensive may be cheaper if its evidence lets the buyer close and use it without friction.
The broker's role in building the chain is therefore important. The broker may collect corporate registry documents, old invoices, registry correspondence, route histories, letters of authority, account-access evidence and operational attestations. It may identify gaps and advise the seller to repair records before marketing the block. It may warn the buyer that a proposed bargain is not bankable. In this sense, good brokerage is a private audit function.
The problem is that private audit without accountability can become private certification. The broker may say a chain is clean without showing the basis. It may rely on screenshots or recycled documents. It may treat access to a registry account as proof of legal authority. It may accept a mandate from a person whose authority is unclear because the block is valuable. It may fail to preserve documents once a dispute emerges. It may sell confidence without preserving the trail that justifies confidence.
AFRINIC's governance discontinuity makes chain-of-custody discipline more valuable because later reconstruction may fall to different officers, a receiver, a new board, a court or a successor operator. If transaction records are thin, every later dispute becomes a credibility contest. If records are robust, the registry can correct errors without destroying unrelated operations. A proper chain should distinguish evidence, assertion and conclusion: the document or verified action, the party's claim about what it means, and the decision based on it. Mixing those categories lets brokers and registries launder judgment into fact. Keeping them separate makes errors reversible.
Timing risk becomes broker leverage
Timing is one of the least visible forms of market power. In IPv4 brokerage, timing can decide price, counterparties and risk allocation. A buyer that needs addresses before a customer launch pays more. A seller facing cash pressure accepts less. A registry that processes requests quickly lowers uncertainty. A registry that delays, asks new questions or enters institutional crisis gives intermediaries more room to shape the outcome.
The broker's timing power comes from experience. It may know which files move quickly, which evidence gaps add weeks, when policy changes are being discussed, whether staff are overloaded, whether a court or election event may interrupt ordinary service, and whether competing buyers are waiting. This knowledge can help clients plan. It can also be used to manufacture urgency.
In a fragile registry environment, time becomes less predictable and therefore more valuable. AFRINIC's period of receivership, litigation and election discontinuity created the possibility that routine actions might be delayed or recast as governance questions. The broker who claims to understand the current state of the registry can charge for that understanding. A buyer may accept a less favourable block because the broker says it can close faster. A seller may grant exclusivity because the broker says delay will kill demand. A market maker may hold inventory through uncertainty and release it when fear rises.
Timing power also affects evidence. The broker may advise a party to submit a file before a policy change, wait until after a governance step, avoid a path that could invite review, or collect additional documents before contacting the registry. Such advice can be prudent. It can also hide weak authority or steer the transaction into a channel favourable to the broker.
The registry itself can unintentionally magnify broker timing power. If service expectations are unclear, if reasons for delay are vague, if similarly situated requests receive different treatment, or if dispute notations do not specify what is blocked, brokers fill the information gap. They become informal forecasters of institutional behaviour. The market then pays for access to practical knowledge that should have been public procedure.
AFRINIC can reduce this power without banning brokers. It can publish clearer processing stages, evidence requirements, expected timelines, escalation paths, reasons for pause and appeal mechanisms. It can separate routine record maintenance from contested resource review. It can make dispute status precise enough that counterparties know whether a transfer, contact update, reverse-DNS change or route-origin action is actually blocked. It can preserve audit logs that later show whether a delay was justified.
For brokers, the corresponding discipline is to distinguish forecast from fact. "The registry usually asks for this evidence" is different from "the registry has required this evidence in your file." "This transaction may face delay because of institutional uncertainty" is different from "the registry will reject it." "We have experience with similar files" is different from "we have special access." Timing claims move money. They should be documented with the same care as price and authority.
Registry fragility is evidence, not the whole story
AFRINIC's recent history is useful for broker-market governance precisely because it should not be reduced to a generic crisis story. The relevant point is narrower. A registry that has faced alleged address-record manipulation, high-value litigation, court-appointed receivership and disrupted election processes gives counterparties reason to treat registry confidence as a market variable.
Different observers draw different conclusions from that history. Some emphasise institutional repair through courts, receivership and board recovery. Some emphasise governance failure and the cost of operating without ordinary leadership. Some emphasise member conflict, commercialisation or external pressure. Some emphasise the need to protect the African registry function from any single litigant, faction or outside institution. A broker-market analysis does not need to settle every political interpretation. It needs to ask what market participants learn.
They learn that registry confidence is not automatic. A buyer of AFRINIC-administered space may ask whether the record will be respected by carriers, clouds, lenders and customers. A seller may ask whether a transaction will close before another legal interruption. A broker may ask whether a file with clean documents can still be delayed by institutional scrutiny. A registry officer may ask whether approving a transaction could later be criticised as favouring one side in a broader conflict. Each question raises transaction costs.
The alleged address-theft history supplies one lesson: records can become the object of economic attack. When dormant blocks acquire high value, chain-of-custody discipline matters. When internal access or institutional weakness is alleged, counterparties cannot rely on database entries alone. They need documentary provenance and credible correction mechanisms.
The Cloud Innovation litigation supplies a different lesson. Without taking a position on the merits of either side's claims, the dispute showed that high-value IPv4 holdings can become the place where resource review, contract interpretation, regional policy, commercial use and court remedy collide. Litigation can protect rights. It can also make the registry layer a live risk surface for operational assets.
Receivership and election discontinuity supply a third lesson. They show that governance continuity itself can become part of transaction risk. A registry can keep technical services running while counterparties still worry about who has authority to decide contested files, how long routine service will take, and whether institutional memory will survive a handover. Later board recovery can reduce that risk, but it does not remove the need for auditable files created during the period of uncertainty.
These episodes explain why brokers gain power in AFRINIC-related transactions. The broker does not need to control the outcome. It only needs to be perceived as knowing how to navigate uncertainty. In a low-trust environment, navigation knowledge is marketable. The market-design challenge is to ensure that such knowledge is rewarded when it reflects diligence, not when it reflects privileged access, fear or unverifiable claims.
Mandate laundering enters through the file
Mandate laundering occurs when a narrow administrative role is converted into broader economic authority through language such as community, stewardship, region, stability or continuity. In the broker market, this conversion often enters through the transaction file. A question that should be about authority and evidence becomes a question about whether the registry approves of the commercial model.
A registry must maintain uniqueness and accurate records. It must reject false authority. It must prevent duplicate claims. It must preserve publication services such as WHOIS, RDAP, reverse DNS and RPKI. It must record disputes and respect valid legal constraints. These are registry functions. They are real and important.
But a registry can be tempted to go further. It may ask whether a holder is using addresses in the approved spirit, whether a lease is sufficiently regional, whether a buyer deserves scarce space, whether monetisation is acceptable, whether out-of-region use is politically offensive, or whether a broker's business model undermines community control. Some of these questions may have policy constituencies behind them. That does not make them part of a routine record update unless they have been clearly adopted as applicable rules, bounded by procedure and subjected to independent review.
The broker is often the first actor to translate informal registry concern into commercial terms. It may tell the buyer that AFRINIC dislikes certain uses. It may advise the seller to frame the transaction in a way that avoids scrutiny. It may recommend a lease because a transfer would trigger questions, or recommend a transfer because a lease is politically sensitive. It may turn the registry's broad unease into private deal structure.
This can be practical risk management. It can also hide the laundering of mandate. If a registry has no clear authority to regulate a business model, but brokers structure transactions around fear that it might do so anyway, the market begins to behave as if the authority exists. The registry's informal preference becomes a private tax on transactions. The broker collects the tax by selling compliance architecture.
AFRINIC is vulnerable to this pattern because its recent disputes have mixed several narratives: regional representation, scarcity fairness, alleged misuse, resource review, litigation, continuity and institutional survival. In such an environment, the line between protecting the ledger and protecting the gatekeeper can blur. A broker who understands that blur can profit from it. A registry that tolerates the blur increases dependence on brokers.
The discipline should be simple. For each transaction, distinguish the registry's question from the market's question. The registry's question is whether the requested record action is authorised, accurate, unique, technically coherent and not barred by a specific rule or independent decision. The market's question is whether the buyer, seller, lessor, lessee, lender or operator accepts the commercial risk. The broker may advise on the market question. The registry should not convert it into a discretionary permission system. If a rule exists, apply it. If fraud exists, correct it. If a court order binds the registry, follow it. If a dispute exists, record it precisely. The objection is to using control over records to smuggle broader economic judgments into the file.
Settlement exposure is not solved by escrow branding
Settlement is one of the places where broker-market governance becomes concrete. Money may move before the registry acts. The registry may act before the buyer can route. A seller may withdraw cooperation after payment. A buyer may refuse release because a post-close service is not working. A court order or dispute notice may arrive between signature and update. A broker may control documents, contacts or operational credentials during the gap. These are not theoretical risks. They are the daily mechanics of a market whose asset is recognised by an administrative record but used through live network operations.
Escrow helps, but it does not solve the whole problem. It can hold funds, define release conditions and reduce payment fraud. It cannot by itself prove authority, preserve reverse DNS, maintain RPKI, clean abuse reputation, decide who controls route announcements, or determine what happens if the registry asks for more evidence after funds are committed. Treating escrow as the main answer overvalues the payment layer and undervalues the authority layer.
A serious brokered transaction should describe which event releases money. Is it signature, registry acknowledgement, completed holder change, working reverse DNS, successful route announcement, creation of route-origin authorisations, or a defined period of stable operation? Each answer allocates risk differently. The seller wants release once it has done what only it can do. The buyer wants assurance that the resource is usable. The broker may prefer a trigger that closes quickly and protects commission. The registry may not recognise any of the parties' commercial definitions as dispositive.
Settlement exposure also includes documentary control. If the broker holds letters of authority, registry credentials, abuse mailbox access or routing coordination, the parties need to know when those items move and under what conditions. A broker that controls evidence can become indispensable at the point of maximum leverage. A buyer may discover that it paid for a block but depends on the broker for a critical maintenance function. A seller may discover that a mandate continues to be used after the intended transaction failed.
Auditability is the antidote. The file should show the requested registry action, the authority for that action, the payment trigger, the operational transition, the release of credentials, the preservation of evidence, and the procedure if a representation proves false. It should state who can withdraw a request, who can instruct escrow, who controls routing during the gap, who maintains abuse contacts, and what happens if a dispute notation appears. These terms need not be public, but they should be reconstructable by the parties, the registry where relevant, and an independent forum if the transaction fails.
AFRINIC's history gives this issue extra weight because institutional delay or litigation can widen the gap between agreement and recognised use. In a fast registry environment, settlement gaps are short and predictable. In a fragile environment, the gap can become the transaction's centre of gravity. Brokers who understand the gap can reduce risk. Brokers who obscure it can turn settlement into leverage.
Blind spots sit between the record and the network
Every registry has blind spots. It sees records, accounts, tickets, policies and services. It does not see every side agreement, financing arrangement, lease, customer dependency, beneficial owner, undisclosed resale, operational delegation or broker incentive. It may not know that a seller is distressed, that a buyer is using a front entity, that a broker has promised the same block to another party, or that a route-origin change is part of a broader transaction. The registry's knowledge is formal. The broker's knowledge is transactional. A 2026 WHEREIS study on IP address registration geo-consistency makes the same general point from a measurement angle: registry data is operationally consequential, yet post-allocation use and registration facts can diverge in ways that matter to operators and policy debates.
This difference is normal. The registry should not know everything. A registry that demands full visibility into every commercial arrangement risks becoming the market enforcer it should not be. But blind spots must be understood because brokers can exploit them.
One blind spot is beneficial control. A registry may see the registered holder and authorised contact, but not the economic party directing the transaction. If the broker controls that party, has an undisclosed option, or finances the transaction, the registry may process a file that looks like holder choice while practical control sits elsewhere. This matters when intermediaries aggregate supply or steer scarce blocks through affiliated entities.
Another blind spot is mandate scope. A letter may authorise the broker to discuss a sale but not to sign transfer documents. It may authorise a lease but not a transfer. It may authorise one buyer but not another. It may expire before closing. If registry staff see only a broad mandate, the broker's practical control may exceed the holder's intention. Precise authority language reduces this risk.
A third blind spot is ongoing operation. The registry may approve a holder change while reverse DNS, RPKI, abuse handling and routing remain controlled by a broker or seller. That may be intentional and harmless. It may also mean the buyer does not truly control the operational resource it paid for. Registry records and operating control can diverge. Broker governance should make that divergence explicit.
A fourth blind spot is dispute motivation. A party may frame a commercial dispute as a registry-authority defect to gain leverage, or frame an authority defect as a mere commercial dispute to avoid scrutiny. Brokers can help classify disputes accurately, but they can also choose the classification that helps their client. AFRINIC's history of litigation and governance conflict makes precise classification essential.
The market should not expect AFRINIC to read minds. It should expect AFRINIC to ask narrow, repeatable questions: who is the recognised holder, who is authorised to request this action, what action is requested, what evidence supports it, what representative role is declared, what services are affected, what dispute specifically constrains the action, and what audit trail will survive? These questions reduce blind spots without expanding the registry's mandate.
Brokers should also accept that registry blind spots create responsibility. If a broker knows a file is misleading, silence is not neutrality. If it knows authority is weak, packaging the file as clean is misconduct. If it knows operational control will not pass, the buyer should be told. The broker's knowledge is the reason it is paid. It is also the reason it should meet a higher standard of documentation.
Broker capture is not ordinary brokerage
A market needs brokers. It does not need broker capture. The difference is whether intermediaries compete to reduce transaction costs or position themselves to control access to scarce resources and registry process.
Ordinary brokerage is plural. A buyer can compare advice. A seller can seek alternative mandates. Registry requirements are public enough that no broker can claim mystical access. Evidence standards are clear enough that counterparties can challenge weak files. Brokers compete on diligence, network, reputation and execution. Mistakes are traceable.
Broker capture is narrower and more dangerous. It appears when a small number of intermediaries control supply visibility, dominate registry-facing knowledge, hold exclusive mandates over significant inventory, influence policy language, maintain privileged relationships with operational staff, or make themselves indispensable to routine file movement. Capture does not require formal corruption. It can arise from asymmetry, habit and the market's fear of registry uncertainty.
AFRINIC's environment creates capture risk because uncertainty increases the premium on experienced intermediaries. If parties believe that only a few brokers can navigate AFRINIC successfully, mandates and demand may concentrate. Those brokers then gain more transaction data, more registry experience and more ability to shape expectations. The cycle reinforces itself.
Capture also appears through narrative. A broker can argue that it represents market efficiency against registry obstruction. Another can argue that it protects the community against speculators. Another can present itself as continuity guarantor. Some of these narratives may contain truth. The governance question is whether the narrative allows the broker to avoid ordinary conflict disclosure and evidence discipline.
The registry can accidentally encourage capture by communicating informally, applying requirements inconsistently or treating certain intermediaries as trusted channels without transparent criteria. A broker who becomes the practical route to predictable service has captured part of the registry interface. That is bad for buyers, sellers and the registry itself. It makes institutional trust dependent on private relationships.
The answer is open procedure. AFRINIC should not maintain a hidden broker class. If representatives can submit files, the requirements should be public. If certain documents are needed, list them. If a broker cannot act without a holder's authorisation, say so. If a service is paused by a dispute, describe the scope. If a file is rejected, give a reason that can be corrected or challenged through a defined path. The more predictable the process, the less valuable capture becomes.
The market also needs counterparty discipline. Buyers should ask whether the broker is principal or representative. Sellers should ask where their block will be marketed and whether the mandate can be used to support other deals. Both should ask what happens if registry process stalls. Brokers who resist basic role and evidence disclosure are asking the market to trust personality over infrastructure. Capture is not solved by denunciation. It is solved by reducing the private value of hidden information.
Legible intermediation is the realistic end-state
The IPv4 market will not become broker-free. Scarcity, uneven holder distribution, registry process, reputation history and counterparty distrust make intermediation economically rational. The question is whether brokers remain servants of market coordination or become private governors of access to registry-dependent assets.
AFRINIC's case shows why the distinction matters. In a stable registry, brokers reduce search and documentation costs. In a fragile registry, brokers also sell interpretation of institutional risk. That may be valuable, but it is dangerous if the broker can allocate fear without disclosure. The broker who can allocate fear can shape price, timing and counterparty choice. The broker who can shape those things without auditability is not merely facilitating a market. It is governing it.
Legible intermediation is not a demand for a heavy licensing regime. It is a demand that the market be able to reconstruct who did what, on whose authority, with what conflicts, and with what effect on the registry record. The role disclosure should say whether the broker represents buyer, seller, both parties as disclosed intermediary, or itself as principal. The authority file should connect the registered holder to the signatory, the signatory to the mandate, the mandate to the transaction and the transaction to the requested registry action. The evidence file should preserve documents in a form that later allows a registry, court, buyer, seller or successor operator to understand why a change was made.
Conflict notation should identify inventory interest, dual representation, financing involvement, post-close management role or a relationship with a registry-facing representative. Dispute notation should distinguish allegations from proven defects, payment disputes from authority defects, contract disputes from chain-of-custody problems, and governance disputes from prefix-specific constraints. Service-continuity terms should identify who maintains routing, reverse DNS, RPKI, abuse contacts and operational communication during and after closing. Reversible-error design should preserve the last verified state and isolate the disputed action so that correction does not become arbitrary destruction of running networks.
AFRINIC can support this by standardising representative authority, clarifying required evidence, publishing processing states, preserving audit logs and separating routine maintenance from contested enforcement. Brokers can support it by making their role, conflicts and evidence clear. Buyers and sellers can support it by refusing files that depend on vague authority stories. Courts and other independent forums can handle disputes that exceed registry administration.
This architecture will not eliminate bad transactions. No market design does. It will make bad transactions easier to detect, isolate and reverse. It will also reward brokers whose value lies in diligence rather than access. That is the right direction for a scarce infrastructure market.
The buyer, seller, broker and registry officer around the IPv4 block do not need to agree on every theory of ownership, stewardship or regional policy before a legitimate transaction can proceed. They need a narrower shared grammar. Who controls the resource today? Who has authority to request a change? What evidence proves it? What conflicts exist? What services must continue? What dispute, if any, limits action? What happens if the evidence is wrong? Who can audit the path later?
If those questions are answered clearly, brokers reduce transaction costs and the registry protects confidence. If they are hidden, brokers become information gatekeepers and registry fragility becomes market power. AFRINIC's broker-market governance problem is therefore a test of institutional maturity. The market does not need brokers to be saints. It needs them to be documentable. It does not need AFRINIC to approve every commercial judgment. It needs AFRINIC to keep the ledger narrow, accurate and continuous. In a scarce IPv4 market, that combination is not administrative tidiness. It is the difference between useful intermediation and a shadow governance layer built from fear, timing and unverifiable authority.