AFRINIC and the economics of fees, reserves, and incentives

The fee table is not the economy

AFRINIC is often introduced by its formal role: the Regional Internet Registry for Africa and parts of the Indian Ocean, incorporated in Mauritius, administering IPv4, IPv6 and autonomous system numbers. That is a correct description, but it is also a constitutional description. It says what function the institution claims. It does not say how incentives change when a member association that was built for coordination sits above scarce, tradeable and business-critical IPv4 resources.

The more revealing documents are prosaic. Fee schedules, renewal calendars, service rules, reserve statements, court reports and public arguments over legal spending say more about the institution than mission language does. They show a registry whose ordinary invoice is attached to an extraordinary economic object. A yearly membership charge that looks modest beside the value of an IPv4 block can become a lever when payment status, transfer eligibility, record updates, reverse DNS, resource standing and confidence among counterparties meet in one institutional relationship.

The question is not whether AFRINIC needs money. Every critical coordination body needs revenue, staff, systems, audit, insurance, legal advice and a cushion against shocks. A registry with no reserves would be a bad public-interest outcome. The question is what type of institution member fees create when the revenue model was designed for administrative service but the resources under administration now behave like capital. In that setting, fees, reserves and legal budgets stop being neutral accounting lines. They become part of the registry's governing machinery.

Heng Lu's public notes supply the sharpest economic frame for this problem. They argue that mandatory registry fees resemble a tax on connectivity because members cannot move the same number resources to a competing authoritative registry at will; they also argue that IPv4 scarcity turns registry recognition from a low-stakes record function into a high-consequence control layer. Those notes come from an interested market participant and AFRINIC adversary, so they should not be read as disinterested findings. Their value is analytical: they describe a mechanism that can be tested against fees, budgets, litigation and market behaviour.

Official AFRINIC, NRO, RIR and ICANN materials are useful here as factual exhibits. They show dates, service descriptions, fee categories, exhaustion phases, receivership statements, continuity concerns and public positions. They cannot by themselves settle the economics. An institution will naturally describe its fees as cost recovery, its reserves as prudence, its policies as stewardship and its legal spending as defence of the registry. Those descriptions may contain truth. They are also self-interested. The price of registry power is better found in incentives than in vocabulary.

From clerical invoice to dependency charge

AFRINIC's public fee material reads like ordinary nonprofit administration. The organisation charges members to support operations. Fees may change with operational costs and financial health, and changes are validated through formal channels. Approved applicants pay allocation or assignment charges. Annual membership fees depend on the aggregate billable resources a member holds. There are different categories for resource members, end sites, IPv6 and special cases such as qualifying academic, research or critical-infrastructure applicants.

On a narrow reading, none of this is remarkable. A registry must pay for databases, WHOIS and RDAP, reverse DNS, Internet Routing Registry functions, RPKI services, member systems, support, security, audits, meetings and administration. It must maintain records and publish reliable information. It must be able to withstand system failures, staff turnover, payment delays and legal shocks. No serious operator should want an address registry living hand to mouth.

The economic reading is different. The invoice is not merely a bill for optional service. It is one of the conditions through which a member remains in recognised standing with the institution that controls the authoritative record. AFRINIC's billing calendar is therefore not a footnote. Renewal invoices are issued on November 1. Early settlement may earn a 5% discount. Invoices fall due on January 31. A 28-day moratorium follows. Late-payment penalties rise through March, April and May. A closure process begins on June 1.

Those dates matter because the registry relationship is not comparable to a subscription to software that can be replaced over a weekend. A member's standing can affect access to support, processing of record changes, confidence in transfer dealings, reverse DNS, public contact data and the perception that resources are cleanly recognised. The fee may be small against network revenue, but non-payment can become large in consequence. The relevant price is not only the dollar amount. It is the dependence attached to the dollar amount.

This is the first institutional-economics chain. Mandatory fee becomes standing condition. Standing condition becomes service access. Service access becomes transfer confidence and operational continuity. Transfer confidence and continuity become asset value. Once that chain exists, the registry's revenue model is no longer safely separated from resource economics. A fee dispute or billing classification can travel into the market value of an address block even if no packet is blocked by the invoice itself.

The usual defence is that billing discipline is necessary. It is. A member-funded body cannot let non-payment become normal. But the proportionality question is unavoidable. Late penalties, collection steps and limits on discretionary services are one category. Threats that cloud recognition, transferability or live operational continuity are another. In an ordinary association, the difference is inconvenience. In a registry, it can be a discount applied to capital-like resources.

That is why the fee schedule must be read together with the powers around it. If the registry charges only for narrow administrative service, then the consequences of non-payment should be narrow and proportionate. If the registry uses standing to influence transfer, status, dispute posture or resource recognition, then the fee has become part of a larger permission system. AFRINIC's core problem is that the invoice is written in service language while the surrounding environment is now scarcity economics.

Billing tiers expose the value gap

AFRINIC's IPv4 fee table is resource-weighted but not value-weighted in any capital-market sense. A resource member holding from a /22 up to below a /20 sits in the Extra Small band at $1,400 a year. A /20 to below a /18 is $2,200. A /18 to below a /16 is $6,400. A /16 to below a /14 is $12,800. A /14 to below a /12 is $22,500. A /12 to below a /10 is $30,000. Holdings at or above a /10 are $38,400.

There are variations. End-site fees are lower. Academic and research institutions may qualify for a 50% discount. Critical-infrastructure applicants may receive a full discount if they satisfy the applicable conditions. IPv6 has its own logic. Actual bills depend on status, resource mix and current rules. Yet the pattern is clear enough: the table is an administrative membership schedule, not a mechanism for pricing the economic value beneath the record.

The mismatch becomes stark once IPv4 market value enters the calculation. The Internet Governance Project reported in 2021 that IPv4 transfer-market prices had climbed from roughly $8 per address in 2017 to roughly $30 by 2021. At $30 per address, a /16 is worth close to $2 million. Under the AFRINIC table, a member whose holdings place it in the /16-to-/14 band pays $12,800 a year before any other applicable charges. A /14 contains more than 262,000 addresses. At the same 2021 price, it implies value above $7.8 million, while the annual category begins at $22,500.

Those arithmetic comparisons should not be mistaken for a clean valuation model. Blocks differ by size, reputation, transferability, route history, contamination risk, legal status, geographic perception and registry risk. Prices move. Not every address is liquid at the same number. But the order of magnitude is the point. AFRINIC's invoice is association-scale. The asset value linked to recognition can be balance-sheet-scale.

IGP made the same point in a more controversial setting when discussing Cloud Innovation. It reported that the company had acquired nearly 7 million AFRINIC IPv4 addresses, paid AFRINIC about $10,000 a year in registry fees, and leased addresses to customers at $2 to $3 per address per year. These figures were part of IGP's analysis of arbitrage and should be attributed as such. Still, they illustrate the structural issue: a fee schedule built around administrative categories can sit beneath commercial use cases whose revenue and valuation are far larger than the membership charge.

The obvious reply is that the registry should not tax asset value. AFRINIC is not selling property. It charges for registry service. It should not become a royalty collector on every address and every transaction. That reply is serious. If annual registry fees were tied to full market value, they could become extractive, unpredictable and hostile to small networks. A registry that priced itself like a sovereign taxing scarce capital would invite even more conflict.

Yet the reply cuts both ways. If AFRINIC charges only for service, its reach should look like service. It cannot plausibly say the fee is merely administrative when money is collected, then treat membership standing, transfer approval, need assessment, regional retention or status conversion as instruments over high-value commercial movement. Either the registry is a low-cost utility maintaining records, or it is a capital gatekeeper whose decisions move liquidity and risk. The fee table says the first. Much of the present controversy arises because members experience elements of the second.

This is the central mechanism. The fee schedule and IPv4 market value are misaligned. The schedule does not capture asset value; the institution's control points can nevertheless affect asset value. That gap creates incentives for every actor. Members seek to minimise fees and maximise optionality. The registry seeks financial stability and relevance. Litigants fight harder because the downside is large. Markets price the uncertainty. What looks like a small invoice becomes the entrance to a larger bargain.

Reserves are not just rainy-day money

A reserve fund is usually presented as boring prudence. It covers payroll shocks, system failures, currency movements, payment delays, emergency vendors, audit, security incidents and legal surprises. For a regional registry, continuity money is a public good. The Africa region would be worse off if AFRINIC's records, RPKI, reverse DNS, member systems or support functions failed because the institution had no financial cushion.

But reserves also change incentives. Cash gives management time. It lets a board litigate, negotiate, resist, rebuild, delay or survive a period of institutional stress. It can keep essential services running during receivership. It can also soften the discipline that members might otherwise impose. The same dollar can insure the registry function or entrench the institution that controls it. The difference is governance.

AFRINIC's crisis made that distinction visible. IGP reported in August 2021 that the Supreme Court of Mauritius provisionally froze up to $50 million of AFRINIC funds held at two banks after Cloud Innovation sought urgent relief. That figure should not be read as proof that AFRINIC had $50 million in free reserves; it reflected the litigation claim and court process described in that reporting. Its importance is institutional. Cash, banking access and legal process became weapons and vulnerabilities in a dispute over registry power.

The NRO's 2023 statement on the appointment of an official receiver used preservation language. It described the receiver's role as maintaining the status quo of AFRINIC's assets, preserving the value of the business, overseeing the election process and helping restore a board and chief executive. That official statement is useful as a factual exhibit. It shows that assets, governance machinery and registry continuity had become linked through court-supervised financial preservation.

For members, the reserve question is therefore not only "how many months of operating cost should AFRINIC hold?" It is "what may member-funded reserves be used for?" Payroll, record integrity, security services, disaster recovery, escrow, audit and continuity are one category. Strategic litigation, discretionary enforcement disputes, public-positioning campaigns and prolonged institutional fights are another. The first is insurance. The second is political economy.

Heng Lu's public note on registry power and liability cites AFRINIC's 2021 annual report as showing membership-fee revenue of about $5.98 million and expenses of about $4.12 million. Those numbers should be checked against the underlying accounts before being used for legal conclusions. As an economic signal, they are useful: an association-scale operating budget can sit above private and public consequences many times larger than annual revenue.

The reserve fund then becomes a map of incentives. If reserves are large enough to fund prolonged conflict but not large enough to compensate foreseeable harm, the institution has more capacity to fight than to make whole. If reserves are too small, the registry may be too fragile to resist pressure or maintain service during litigation. The right answer is neither starvation nor a war chest. It is a ring-fenced continuity reserve, paired with clear disclosure of extraordinary spending and independent review of high-consequence legal strategy.

Reserves also affect member behaviour. A member that believes fees are funding a protected service layer may accept the charge as infrastructure cost. A member that believes fees are building a fund for discretionary control will treat the same charge as a levy for institutional self-preservation. Trust, not cash alone, determines whether reserves are stabilising.

Legal budgets reveal who bears the downside

AFRINIC's recent history cannot be understood without litigation cost. Public reporting describes a registry that operated without a board for years, entered receivership, attempted elections, saw voting disputes and annulment, later elected directors and remained entangled in matters involving Cloud Innovation, ICANN, Smart Africa, member rights, bylaws and attempted winding-up proceedings. These are not side costs. They are the operating environment.

The Register reported in March 2026 that AFRINIC accused Cloud Innovation, LARUS and associated advocacy campaigns of trying to paralyse the organisation through litigation and procedural roadblocks. AFRINIC said delays and rising legal costs obstructed initiatives such as training and research. Heng Lu replied to The Register that the deeper issue was structural: high-consequence registry power over economically critical resources without legal and financial liability matching the harm that could result.

Both positions can be partly true. A litigant can impose heavy costs on a registry. A registry can also create the incentive for litigation when it threatens action that could damage a member's business far beyond ordinary fees. The practical question is not which public statement is more righteous. It is how the cost of fighting is allocated.

In a member-funded registry, legal bills are never abstract. They are paid from fees, reserves, deferred projects, staff attention or future fee pressure. NRS, an advocacy group associated with Lu's reform campaign, has publicly claimed that $3,289,408 of member-funded legal expenditure was disclosed for 2022-2025 and urged members to demand invoices, authority records and answers before approval. That claim is partisan and should not be treated as audited fact without the underlying documents. It is still a signal of member anxiety: who authorised the spending, what did it buy, and which members benefited?

Legal expenditure is delicate because AFRINIC's members are not aligned. Some may want the organisation to fight Cloud Innovation aggressively. Others may fear precedents around revocation, regional-use enforcement, transfer limits or discretionary review. Some may simply want services restored and costs contained. If all members fund one institutional strategy, the legal budget becomes a collective levy for a disputed theory of control.

That is not unique to AFRINIC. Associations often litigate in the name of a membership whose interests diverge. Scarcity raises the stakes. When the dispute concerns resources that can support large commercial revenue, a registry's legal posture can affect transfer prices, leasing confidence, due diligence and financing terms. Members are not merely paying lawyers to defend a nonprofit. They may be financing positions that alter the value of their own holdings.

This is moral hazard in institutional form. If management can pursue high-risk enforcement while members pay the bill and liability remains capped or uncertain, the institution may litigate too readily. If a well-resourced member can use courts to freeze operations, that member may litigate too readily. The answer is not to pretend law can be avoided. It is to align legal authority with disclosure, proportional remedies, member-visible spending categories and review before disputes become existential.

Legal budgets therefore require a different disclosure norm from ordinary operating expenses. Members do not need privileged legal memos. They do need decision-useful categories: how much was spent on core service continuity, how much on corporate governance, how much on elections and bylaws, how much on resource disputes, how much on outside interventions, and what contingent exposure remains. Confidentiality can justify redaction. It cannot justify opacity about purpose and authority.

The liability gap makes every fee more sensitive

The fee debate is often framed as if members are simply arguing over price. That understates the problem. A member does not merely buy an annual service. It accepts a dependency relationship in which a registry action, delay or dispute can affect resources whose business value is far larger than the invoice. The fee is therefore paired with risk. The member pays into the institution but may have limited financial protection if the institution makes a damaging decision.

Heng Lu's public criticism of RIR liability turns on this gap. His notes argue that registry bodies can exercise high-consequence power over IPv4 recognition while operating on association-scale budgets and with legal exposure that may be much smaller than the losses a holder could suffer. That is an adversarial argument and should be read as such. Yet the economic point is broader than any one dispute. A body that can change the liquidity, usability or perceived legitimacy of a resource should not be analysed only by its operating cost. It should be analysed by the size of the downside it can create.

Ordinary markets have ways to price such downside. A custodian has duties. An exchange has rule books, risk controls and capital expectations. A bank has prudential supervision. A vendor with mission-critical responsibilities may carry insurance, service credits or contractual liability. None of those analogies maps neatly onto a Regional Internet Registry, and the internet-numbering system has its own legal architecture. But the absence of a perfect analogy does not erase the problem. If members must rely on registry recognition, they need some substitute for the discipline that liability, competition or exit would create elsewhere.

This is why a low fee can still feel expensive. The member may be paying a modest sum for a service that is cheap to perform in the narrow sense: maintain records, answer tickets, publish data, operate reverse DNS and support routing-security functions. But the same member is exposed to a much larger loss if the institution uses standing, policy interpretation or litigation posture in a way that clouds a block's commercial usefulness. The annual fee is small; the option value surrendered to the registry may not be.

Reserves do not solve this unless they are designed for the right purpose. A reserve that preserves systems during a shock is useful to all members. A reserve that allows the institution to fight for years without clarifying member authority may increase the liability gap. The member funds the war chest but may not receive compensation if the fight damages the member's own interests. That asymmetry explains why reserve policy and legal-spend categories matter so much. They are not just financial controls. They are partial substitutes for a missing market discipline.

The same logic applies to billing enforcement. If a missed or disputed invoice merely creates ordinary collection consequences, the liability issue is contained. If it can affect transfer confidence, essential record changes or operational continuity, then the registry has converted a small debt into a much larger risk surface. The sharper the consequence, the stronger the duty to provide notice, cure, reasons and independent review.

The public-interest answer is not to make AFRINIC liable for every commercial loss connected to a member's address business. That would be impossible and probably destructive. The answer is to narrow discretion, publish decision standards, ring-fence continuity money and make extraordinary legal spending visible before it consumes the budget. A registry that cannot carry balance-sheet liability for the value beneath its records should be especially careful not to act as if it owns that value.

Member standing is a financial control point

Fee tiers do more than raise money. They define who is small, who is large, who is an end site, who is a resource member, who receives a discount, who is critical infrastructure and who holds enough address space to become strategically visible. A member's place in the table is not only a billing category. It is a political category.

This matters because member-funded governance is often described as democratic. In principle, those who depend on the registry can participate in its corporate and policy life. In practice, participation is expensive. It takes procedural knowledge, time, mailing-list attention, travel, legal understanding and tolerance for conflict. The invoice is paid by many. The governance work is performed by few.

Heng Lu's first AFRINIC governance note argues that "community ownership" can conceal concentration by insiders, consultants and repeat participants who know the rooms and procedures. That is a viewpoint from a party strongly opposed to the current model. But the underlying principal-control problem is familiar. Low participation in a member association does not prove consent. It often proves that the cost of attention is high until a member sees a direct threat.

Fee tiers reinforce that pattern. Large holders pay more in absolute terms, but they also have more reason and capacity to monitor policy. Small operators pay less, yet the fee and compliance burden may consume more managerial bandwidth. End sites may see the registry as remote until reverse DNS, RPKI, transfer status or a billing issue affects them. Members interested in policy but not deeply dependent on resources may have different incentives again. The membership ledger is not a single public with one interest.

The legal form of membership adds another complication. The Register reported in May 2026 that an ISPA response to members argued that AFRINIC Resource Members may not automatically be Registered Members under Mauritian law, creating tension between bylaw rights and company-law rights. That claim belongs to a disputed legal environment, not a final settlement. It nonetheless shows why "member governance" can become fragile: the party paying an invoice, the party relying on resources and the party recognised for corporate-law purposes may not map neatly onto each other.

In a scarcity-era registry, that mapping matters. If members fund reserves and legal budgets but lack practical control over spending or policy risk, the fee becomes an extraction channel. If voting members can impose asset-impacting rules on inactive or differently situated holders, the vote becomes a control channel. If the board adjusts fees for financial health during litigation, the billing model becomes a way of transmitting crisis costs across the member base.

The solution is not slogans about one member, one community or one region. It is segmentation of authority. Routine registry cost recovery can be spread broadly. High-consequence spending and policy moves affecting transferability, status, revocation risk or commercial use should require stronger disclosure and, where feasible, consent from affected principals. Members should not discover after the fact that their fees funded a strategy they would have opposed had the economic consequences been explicit.

Standing should also be separated from essential continuity wherever possible. Billing discipline is legitimate. But using good-standing rules to block unrelated transfers, suspend important record correction or create leverage in disputes can turn routine administration into capital control. The harsher the consequence, the more process, notice and independent review are needed.

Scarcity changes the institution's survival incentives

AFRINIC was designed for a world in which a regional registry allocated, assigned and recorded number resources under community policy. Scarcity changes the organisational bargain. When the free pool shrinks, the institution's central economic role moves away from distribution and toward recognition: maintaining records, processing transfers, verifying claims, marking disputes and preserving security metadata around resources already in use.

That shift should narrow the mandate. A registry that no longer allocates from abundance should become more like a disciplined records utility and less like a planning authority. The temptation runs the other way. When allocation declines, relevance may be preserved through policy control: transfer restrictions, regional retention, needs reassessment, business-model scrutiny, membership leverage and status conversion. The institution can lose its original allocator role and then recreate importance by controlling movement.

This is the second institutional chain. IPv4 scarcity increases resource value. Higher resource value increases the stakes of registry recognition. Higher stakes increase litigation, lobbying and political attention. More conflict increases legal spending and reserve needs. Higher spending creates pressure for fees. Fee dependence creates incentives to defend the institution's discretionary reach. The institution then has reason to preserve the control points that make members care about it. Survival and stewardship begin to blur.

No one needs to assume bad faith for this to happen. Organisations protect their budgets, staff, reputation and authority. Boards fear being blamed for collapse. Staff fear that a narrow mandate will be treated as weakness. Supporters may genuinely believe that regional restrictions protect African networks. Critics may genuinely believe that liquidity and portability protect those same networks. The danger lies in incentives, not secret motive.

The Cloud Innovation dispute shows the downside of making discretionary review existential. IGP's 2021 analysis argued that AFRINIC overreached when it threatened resource reclamation over a contested theory of use, while also criticising Cloud Innovation's legal response as excessive because it endangered the registry's operations. That two-sided criticism is useful. It treats escalation as a product of asymmetric downside. AFRINIC feared loss of policy authority and regional legitimacy. Cloud Innovation feared destruction of a business built around millions of IPv4 addresses. Once both sides saw existential stakes, legal budgets and reserves became weapons.

Members outside the dispute still paid the externalities. A small ISP with no interest in Cloud Innovation's leasing model can still face delayed services, reduced training, slower transfers, uncertainty around records, higher future fees and a registry staff absorbed by legal and governance crises. A dispute between one registry and one large holder becomes a tax on the ecosystem.

The economics points toward proportional remedies. A registry needs tools for fraud, stolen records, duplicate claims, false contacts, abandoned resources and security-integrity failures. It should not reach first for remedies that threaten business destruction unless evidence and authority are correspondingly strong. When the remedy is existential, the target will litigate as if survival depends on it. The legal cost then becomes predictable, not surprising.

Reserves should support de-escalation rather than escalation. They should fund independent review, standstill arrangements, preservation of the last verified record state, dispute notation, escrow of critical data and continuity of essential services. They should not be treated as fuel for all-or-nothing fights over institutional prestige. If AFRINIC wants less litigation, it should make the official path safer than the court path.

The official story cannot be the valuation frame

Official materials are indispensable for facts. AFRINIC publishes fee categories, billing dates, service descriptions, exhaustion pages, policy text and governance notices. The NRO and ICANN publish statements that explain continuity concerns and systemic roles. Those materials matter. They do not decide how markets price registry risk.

An institution will describe its mandatory fees as operational support. It will describe reserves as prudence. It will describe policy restrictions as stewardship. It will describe litigation as defence. Those descriptions may be accurate within the institution's own frame. But the outside valuation frame is built from behaviour: whether blocks trade at a discount, whether counterparties demand warranties, whether leasing structures grow, whether buyers avoid a region, whether members treat billing and legal status as asset risk, and whether disputes make direct holding less attractive.

The official non-property position does not settle the matter either. ICANN's May 2026 intervention in the winding-up case, as reported by The Register, stressed that numbering resources allocated through AFRINIC are not AFRINIC assets available for distribution in a liquidation. That is an important continuity argument. It prevents the registry's corporate estate from being confused with the number resources it administers.

But the fact that resources are not AFRINIC's assets does not mean holders have no economic reliance. It means the registry should be careful not to behave as owner. The relationship is neither simple property nor valueless permission. Members pay fees to maintain recognition over operational identifiers that can carry substantial commercial value. That relationship needs process, accounting, liability and remedies proportionate to the dependence it creates.

The same is true of development claims. A regional registry may want to protect African connectivity. But a fee-funded institution should not treat every discretionary programme, meeting, campaign, legal theory or policy ambition as if it were identical to core record service. Training, outreach and technical assistance can be valuable, particularly where operator capacity and routing security still need support. Value does not automatically justify compulsion through resource recognition. If a programme is a public good, fund and measure it as one. If it is core registry service, make that clear. If it is discretionary politics, do not hide it inside mandatory fees.

Public LARUS and NRS materials also require caution. LARUS is commercial. NRS is activist. Heng Lu is connected to both LARUS and Cloud Innovation. Their claims about registry risk, ownership, fees and member funds are interested. Yet interested actors reveal market pressure. The existence of products and campaigns built around registry exposure shows that counterparties are already treating the registry layer as a material economic risk. Official language cannot make that risk disappear.

Incentives should be judged by institutional behaviour. If fees fund narrow service, if reserves preserve records and continuity, if legal budgets are disclosed and proportionate, and if policy avoids discretionary control over capital movement, AFRINIC's nonprofit language becomes more credible. If fees finance opaque legal strategy, if reserves become a conflict fund, and if standing rules affect high-value transactions beyond reasonable service needs, the language becomes decoration.

Compulsory money should not fund every ambition

AFRINIC's public materials describe a broad institutional life: registry services, number-resource management, WHOIS and RDAP, reverse DNS, IRR, DNSSEC, RPKI, support, training, meetings, fellowships, outreach, development programmes and participation channels. Some of this is core. Some is useful but discretionary. Some is part of the organisation's identity. The fee question is how much of each should be funded through mandatory resource recognition.

The distinction matters more in Africa than it might in a mature, well-capitalised market. Training and technical assistance can be valuable where routing security, IPv6 deployment, local peering, abuse handling and registry procedure still need capacity. A mechanical campaign to strip every non-record activity from AFRINIC's budget would miss genuine regional needs. A registry that helps operators understand RPKI, maintain contacts, correct reverse DNS and participate in policy can create real value.

Yet value is not the same as compulsion. A small access network may need stable records and predictable support more than it needs a conference programme. A university may value training but not want legal disputes funded through the same mandatory line. A large carrier may support regional development spending while resisting transfer controls. A holder focused on leasing may care mainly about clean record changes and dispute notation. Treating all of these preferences as one member interest hides the political allocation inside the fee.

The most dangerous word in such budgets is "mission". Once every activity is described as part of the mission, every spending line becomes essential. Training is mission. Travel is mission. Litigation is mission. Communications are mission. Governance reconstruction is mission. Reserve rebuilding is mission. The phrase may be sincere, but it removes the accounting discipline members need. A compulsory fee should not become an all-purpose contribution to institutional self-definition.

A better model would split the budget into visible layers. The first layer is mandatory record service: uniqueness, accuracy, publication, support, transfers, reverse DNS, routing-security functions, audit and resilience. The second layer is member governance: elections, meetings, accounts, board support and minimum participation machinery. The third layer is development and capacity-building: training, fellowships, outreach and research. The fourth layer is extraordinary conflict: litigation, outside interventions and crisis communications. Each layer may be justified. They should not be blurred.

Such separation would not force AFRINIC to abandon regional development. It would make the case for it stronger. Members and sponsors are more likely to fund programmes that disclose purpose, cost and results than programmes hidden inside a general levy. A transparent training budget can be defended on outcomes: operators trained, routing-security adoption, contact accuracy, ticket resolution, local resilience. A hidden training budget becomes hostage to every fight over legal spending.

The same separation would lower suspicion in disputes. If members can see that essential registry service is protected, a litigant has less ability to claim that the whole institution is financially opaque. If members can see that legal spending is exceptional and authorised, management has a stronger case when it asks for support. If discretionary programmes have their own funding logic, they are less likely to be sacrificed whenever litigation consumes the general fund.

This is the practical financial meaning of narrowing the registry. It is not an argument for a smaller Africa internet. It is an argument for making the mandatory part of the institution legible. Compulsory money should follow compulsory functions. Optional ambitions should stand on their own budgets, evidence and support. The more scarce IPv4 becomes, the more important that distinction is, because the member fee is no longer just a subscription. It is attached to the market's confidence in the recognition layer.

What disciplined cost recovery would require

A disciplined AFRINIC fee model would begin with function. Mandatory fees should pay for the lowest-cost reliable version of necessary registry services: accurate records, uniqueness, WHOIS and RDAP, reverse DNS, IRR, RPKI, member support, transfer recording, abuse-contact publication, security operations, audit, backups, data escrow and disaster recovery. These are not optional ornaments. They are the reason the registry exists.

The second principle is separation. Capacity-building, research, fellowships, meetings, outreach and development programmes may deserve funding. Some may be very useful. But they should be separately budgeted and justified. Some can attract sponsors, grants or voluntary contributions. Some can be scaled to measurable outcomes. Bundling all of them into resource recognition should require a clear explanation of why compulsion is necessary.

The third principle is legal-spend transparency. Members should see annual legal expenditure by category: core service continuity, corporate governance, elections, resource disputes, employment matters, external interventions and strategic litigation. They do not need privileged advice. They do need enough information to understand whether their fees protect the registry function or finance a contested control strategy.

The fourth principle is reserve ring-fencing. AFRINIC should define a target continuity reserve in months of core registry operating cost, a separate governance reserve if necessary, and rules for extraordinary drawdowns. If litigation consumes reserves, members should know which bucket was consumed and what replenishment will cost. A reserve without purpose categories is too easy to politicise.

The fifth principle is proportional enforcement. Fee arrears may justify late penalties, collection steps and limits on non-essential discretionary services. They should not automatically threaten live network continuity, necessary record correction or transfer dealings unrelated to the arrears without notice, cure periods and independent review. Billing discipline is legitimate. Turning billing into existential leverage is not.

The sixth principle is transaction neutrality. Charges for recording transfers should recover reasonable processing cost, not extract rent from capital movement or subsidise unrelated activity. AFRINIC's role should be to verify control, prevent duplicate claims, maintain accurate records and preserve security continuity. The more it asks about business model, customer geography or speculative motive, the more cost recovery turns into economic permission.

The seventh principle is distributional clarity. Small operators, end sites, critical infrastructure and large holders have different exposure. Discounts can be justified when they reduce barriers to essential connectivity. But discounts should not hide cross-subsidies for discretionary spending. A cheap end-site fee is helpful only if the broader registry environment does not impose larger hidden costs through uncertainty.

These principles would not end AFRINIC's disputes overnight. They would improve incentives. Members could see what they are paying for. Management would know which funds are protected. Courts and counterparties would see that essential registry service is separable from institutional conflict. Markets would price AFRINIC resources with less uncertainty.

The point is not to impoverish AFRINIC. A weak registry is not in the region's interest. The point is to make mandatory money follow mandatory functions, and make optional ambitions stand on disclosed budgets. That is the financial meaning of a thinner, more trusted registry.

Watchpoints for fees, reserves and incentives

The first watchpoint is the next published budget and accounts. The issue is not merely whether AFRINIC can present a budget after years of dysfunction. The Register reported in February 2026 that the organisation expected to deliver a budget and action plan as part of its recovery. The test is whether that budget separates core registry service, development programmes, governance reconstruction, legal spending and reserve rebuilding.

The second watchpoint is the legal-cost ledger. NRS's public claim about $3,289,408 in disclosed 2022-2025 legal expenditure is not enough by itself. Members should seek supporting accounts, authority records and spending categories. AFRINIC should want the same clarity if it believes the spending was necessary. Secrecy turns every invoice into a political weapon.

The third watchpoint is reserve policy. AFRINIC should state how many months of core operation it intends to hold, what counts as core operation, when reserves may be used for litigation, and what member approval is required for extraordinary drawdowns. A reserve target without use restrictions is incomplete.

The fourth watchpoint is billing enforcement. Late-payment penalties are ordinary. Closure language is more serious. Members should examine what services remain available during cure periods, whether essential record correction is preserved, whether transfer processing can be blocked for unrelated arrears, and whether live network continuity is protected against administrative default.

The fifth watchpoint is fee-tier politics. Any change to LIR, end-site, ASN, IPv6, academic or critical-infrastructure fees should disclose not only total revenue effect but distributional effect. Which members pay more? Which activities are funded? Does the change lower barriers to connectivity, or does it finance institutional conflict?

The sixth watchpoint is the interaction between transfer fees and transfer policy. A transfer charge may be reasonable if it reflects processing cost. It becomes economically significant when combined with needs assessment, membership requirements, regional restrictions, status conversion or uncertain appeal. Members should judge the total transaction cost, not the line item.

The seventh watchpoint is litigation strategy after board reconstitution. The September 2025 election gave AFRINIC a route back to corporate governance, but The Register reported continuing legal uncertainty, potential court challenges and investigation threads. A restored board can either disclose and discipline legal strategy or inherit the crisis incentives of the receivership period.

The eighth watchpoint is whether AFRINIC treats the end of the free pool as discipline or permission. When unallocated IPv4 approaches zero, the registry's economic role shifts further from allocator to recogniser. That should narrow its mandate. If instead it expands transfer control to preserve relevance, members will pay through liquidity discounts.

The final watchpoint is market behaviour. If AFRINIC-recognised resources trade at a discount, if inbound sellers avoid the region, if leasing grows because direct holding feels risky, if buyers demand warranties about registry action, or if members treat legal and billing status as asset risk, the verdict will be visible. Fee schedules claim one economics. Markets reveal another.

AFRINIC's financial problem is therefore not simply whether fees are high or low. It is whether mandatory member money funds a narrow, trusted record layer or a broader institution with incentives to preserve discretion over scarce capital. Cheap fees attached to unpredictable control are not cheap. Large reserves without purpose discipline are not safety. Legal budgets without member-visible authority are not accountability. A registry that understands this can become boring again. A registry that does not will discover that its most important price is not in the fee table, but in the discount markets attach to trust.