Summary

  • AFRINIC’s unaudited report for the six months ended 30 June 2026 records USD 20,790 of actual MyAFRINIC v2 capital expenditure against USD 650,000 approved. The ratio is about 3.198%, but it is a spending ratio, not a completion percentage.
  • The AFRINIC-37 minutes of 24 June say human resources had been channelled to MyAFRINIC v2, which was expected to go live by year-end while scope was limited. The record does not quantify the people, time or cost behind that statement.
  • The approved budget calls the old portal end-of-life, a single point of failure and a bottleneck for policy, RPKI and RDAP improvements. This makes resource attribution important, but it does not prove an outage, delay or failed project.
  • AFRINIC can make the two reports comparable through a project-resource crosswalk separating staff effort, operating cost, commitments, capital invoices, cash paid, accepted milestones and residual service risk.

Two public observations sit six days apart.

On 24 June 2026, the AFRINIC-37 meeting minutes recorded that AFRINIC’s human resources had been channelled to the MyAFRINIC v2 project. The same discussion said the long-awaited replacement was expected to go live by the end of the year and that its scope was being limited to avoid scope creep.

At 30 June, the unaudited half-year financial report recorded USD 20,790 of actual MyAFRINIC v2 CAPEX against USD 650,000 approved for the year.

The tempting reaction is to place the sentences on one scale: all hands, little money. That reaction would be wrong. One statement describes an operational priority. The other is a line in a capital-expenditure table. A developer’s time, a vendor commitment, an invoice, a payment and an accepted software increment can enter an organisation’s records at different moments and under different classifications. The documents do not say which of those paths AFRINIC uses for this project.

The real finding is therefore not that the numbers conflict. It is that AFRINIC has published no crosswalk by which a member can reconcile them.

The percentage that must not become progress

The arithmetic is simple. USD 20,790 divided by USD 650,000 is approximately 3.198%. Across the full capital table, USD 42,053 of actual CAPEX divided by USD 1,265,875 approved is approximately 3.322%.

Both ratios describe reported capital expenditure through the end of June against annual approval. Neither says that MyAFRINIC v2 was 3.198% complete or that AFRINIC’s capital programme was 3.322% delivered.

Capital work need not spend evenly through a year. A contract may be signed before an invoice falls due. Hardware may be delivered before final acceptance. Software can be developed internally while little project-specific capital cost appears in a table. A milestone payment can concentrate expenditure after testing. Conversely, a high payment percentage can coexist with unfinished work. Cash timing, accounting recognition and technical completion are different clocks.

The Q2 report is useful precisely because it gives an exact observation. It identifies the MyAFRINIC and infrastructure lines and shows how much actual CAPEX had been recorded by the cut-off. It does not label the figures as commitments, total project cost, earned value, work completed or service acceptance. Adding one of those meanings would not strengthen the report; it would replace its accounting label with an inference.

This boundary also protects AFRINIC. A low first-half number is not evidence that procurement failed, that staff were idle or that the year-end expectation had been missed. The evidence cut-off precedes the end of the year. No checked source reports a failed user test, a cancelled contract or an abandoned release. Scrutiny should begin where the record ends, not where suspicion wants it to end.

What “human resources” can and cannot mean

The meeting language is broad. Read literally, “all the AFRINIC human resources” could sound as if every employee had stopped every other task. The minutes themselves show that AFRINIC was still running a policy meeting, member services and registry operations. The safer reading is that staff described an organisation-wide concentration of scarce implementation capacity on MyAFRINIC v2. The public text does not provide a roster, percentage allocation or time period.

The financial report does contain an HR line. It shows USD 822,142 actual against an annual budget of USD 2,901,811, or 28%. But this is an organisational aggregate. The table does not allocate payroll or contractor time to MyAFRINIC v2, infrastructure, policy support or ordinary registry work. It cannot be used to estimate project labour by subtraction.

That leaves several plausible treatments. Internal software work may remain an operating staff cost. Some development may be supplied under a service contract. Some eligible work may be capitalised. Different phases may use different combinations. Those are accounting possibilities, not facts established by the sources. The public problem is that the project line and the staff-priority statement cannot be joined without knowing the boundary AFRINIC applied.

The distinction matters beyond accounting. If staff work sits outside project CAPEX, the USD 20,790 line understates total resources consumed even while accurately reporting capital expenditure. If committed work had not yet been invoiced, the line understates contractual exposure at the cut-off. If a delivered increment had not passed acceptance, payment would overstate usable capability. Each possibility leads to a different governance question.

Why this portal carries an opportunity cost

The 2026 Approved Budget explains why the project received priority. It says MyAFRINIC v1 had been declared end-of-life five years earlier. It calls the portal a single point of failure for the registry and a bottleneck preventing improvements to policies, RPKI and RDAP. Replacing it was presented as the biggest 2026 expense.

These are AFRINIC’s risk statements. They are not evidence that the portal suffered an outage during the period or that a specific route, ROA or RDAP response failed. They do establish the purpose AFRINIC attached to the replacement: reduce a concentrated operational dependency and unlock work that touches number-resource administration.

The AFRINIC-37 minutes make the cost of that priority visible in another way. Staff said a hierarchical AS-SET policy could be implemented in under six months after Last Call, but resources were concentrated on MyAFRINIC v2 and the policy would be prioritised after that work. The same meeting described other ratified-policy paths whose full implementation depended on or followed the portal.

This does not make the portal the wrong priority. Limiting scope may be the most responsible way to move an old, consequential system toward a testable release. But prioritisation consumes more than money. It consumes engineering attention, product decisions, testing capacity and time in which other authorised work waits. A financial table that counts only project CAPEX cannot show that opportunity cost.

The strongest case for the existing disclosure

AFRINIC has done more than announce a transformation. It published an approved annual budget, a Q2 actual table and minutes describing operational constraints. Readers can compare the USD 650,000 plan with the USD 20,790 actual and can see that staff regarded the project as a shared dependency. The report is labelled unaudited, so its assurance level is not disguised.

Nor should a public project ledger become a surveillance system for employees. Individual timesheets, staff performance, vendor identities, source code, architecture and security topology may be private or operationally sensitive. Publishing them could make the registry less secure and turn ordinary project management into theatre.

The useful disclosure sits between the current aggregate and raw internal records. AFRINIC can publish resource classes, amount bands, milestone states and variances without naming an engineer or exposing a system diagram. It can show how categories reconcile without claiming that financial disclosure alone proves technical quality.

A project-resource crosswalk

The smallest useful record would give each reporting period a stable project baseline and keep the measures in separate columns.

Measure What it would answer What it must not imply
Approved project envelope What annual authority was granted That the amount was committed or spent
Contracted or committed value What obligation existed at the cut-off That an invoice was due or work was accepted
Invoiced value What a supplier billed That the invoice was paid or correct
Cash paid What left the organisation That the corresponding capability was live
Capital expenditure recognised What entered the CAPEX account under the applied policy Total effort, completion or service quality
Operating project cost What project work remained in OPEX That all OPEX was avoidable or project-specific
Internal effort band Which teams and aggregate person-month band contributed Individual productivity or a payroll disclosure
Accepted milestone Which bounded deliverable passed named criteria General production readiness
Release state Design, build, internal test, member UAT, beta or production Completion of every dependent policy
Residual dependency Which old-system risk still remained That the risk had caused an incident
Deferred work Which authorised work waited and under whose decision That deferral was wrongful
Correction event What prior entry changed and why Permission to overwrite the historical record

The project baseline matters. “MyAFRINIC v2” is a programme name, not a unit of measurement. A crosswalk should name the scope version, reporting cut-off and acceptance authority. If scope is reduced, the record should preserve both the removed work and the reason rather than making the smaller release look identical to the original plan.

Internal effort can be reported in bands. For example, AFRINIC could identify product, engineering, member-services and security participation and disclose an aggregate range of person-months for the period. That would make the phrase about human resources testable without publishing salaries, individual names or daily logs.

Accepted milestones should describe outcomes that a representative user or accountable owner can test: an identity path migrated, a defined resource-management workflow completed, a rollback exercised, a data reconciliation signed off. A login page is not acceptance of every portal function. A policy feature is not complete merely because the common platform is live.

The narrow conclusion

The Q2 table cannot be made to say more by calculating a percentage to additional decimal places. USD 20,790 is a financial observation under a label. The staff statement is an operational observation under a different label. Both add information; neither explains the other.

AFRINIC may have substantial internal work, late-year invoices, committed contracts or accepted increments behind the two public records. The sources reviewed here do not establish any of those states. They also do not establish failure. That uncertainty is not a licence to guess. It defines the next record the institution should publish.

If AFRINIC joins money, people and milestones in a bounded crosswalk, members will be able to see what the capital line excludes and what the staff-priority statement includes. They will also be able to distinguish a cash lag from a delivery lag and a portal release from the retirement of the old dependency. Until then, the USD 20,790 line counts capital expenditure. It cannot count the work AFRINIC says its people put into MyAFRINIC v2.

Sources