Allocating the continent's largest procurement lines — fuel, food, medicines, machines — without crowning new monopolies. A consultation paper for the member states of the African Union.
Colleagues,
Twelve categories carry nearly half of everything our continent imports. Fuel. Food. Medicines. The machines that scan our sick, power our grids, move our people and print our money.
These are the big-ticket items. They are where the Buy African Initiative will be won or lost — and they are precisely the categories we cannot simply hand out. Give any one of them to a single country and we create a monopoly with a twenty-five-year licence. A handful of strong economies would shoot the lights out while forty others watch. That is not integration; that is concentration with a new flag on it.
This paper does not pretend to solve that. It exists to put the problem in front of all of us honestly — every product broken down, every candidate supplier named, every allocation approach we are studying laid on the table — so that member states can see what we are thinking, test it, and improve it.
The countries named in these pages are those whose minerals and expertise position them today. The list is a starting point, not a door. If your state believes it belongs in a category, that conversation is exactly what this document is for.
Read it. Argue with it. Send us your thinking.
Twelve categories, USD 326.7 billion, 46 per cent of the continental import ledger — ordered by declared 2023 value, closing with the two sovereignty categories whose true size customs data only partly sees. Each carries a first-draft bench of candidate supplier states: real, and deliberately not closed.
Each category resolves into five sections — the product anatomy, the sovereign demand ledger, the current dependency, the continental bench, and the allocation architectures under study. Twelve categories, five sections each: the sixty-section skeleton of the full study to follow.
Diesel, petrol, jet fuel, LPG, kerosene, lubricants and bitumen — the full HS 2710–2715 anatomy of the continent's largest import line.
Fuel utilities, strategic reserves, power stations and government fleets: why this line anchors the Energy pillar's USD 71.65bn of measured sovereign purchases.
USD 120.81bn of refined product and gas imported in 2023 — much of it from refineries on other continents, some of them running African crude.
Nigeria's Dangote complex, Algeria, Egypt, Angola, Côte d'Ivoire, Senegal, Ghana, Libya and South Africa: nameplate capacity, actual runs, honest utilisation.
Why one refinery must never hold 54 states: regional supply zones, anchor-plus-satellite rights over storage, blending and distribution, and the caps between them.
The six staples inside the USD 46.93bn line — and which state grain boards and reserve agencies buy them by the shipload.
Food inside the USD 27.03bn agriculture pillar: subsidy programmes, strategic reserves and school feeding make the state the anchor buyer.
Black Sea wheat, Asian rice, South American sugar and oils: the dependency map — and what the supply shocks of this decade proved about it.
Ethiopia's wheat drive; Tanzania, Zambia, Zimbabwe and Sudan on grain; Egypt and Nigeria on rice; Eswatini and Mauritius on sugar; Côte d'Ivoire and Ghana on oils.
Rotating offtake rights, regional reserve mandates and laddered volumes that reward real yield gains — without handing any single breadbasket the continent.
Cars, buses, trucks, tractors and special-purpose vehicles: the HS 8701–8705 anatomy of a USD 37.05bn line.
Police, health, utilities, municipalities and ministries renew vehicles in their thousands — visible, cyclical, and squarely inside the match window.
Despite two genuine automotive powers on the continent, most member states' fleets still arrive fully built from Asia and Europe.
South Africa's full OEM base and Morocco's Tangier hub — plus Egypt, Ghana, Nigeria, Kenya, Ethiopia and Rwanda assembly plants hungry for fleet volume.
Lanes by vehicle class, regional assembly quotas and local-content ladders: how fleet demand seeds plants without creating a two-country duopoly.
Transformers, generators, switchgear, solar modules and cable: USD 22.50bn of hardware behind every electrification pledge on the continent.
State power companies and rural electrification agencies dominate demand — the second face of the Energy pillar's sovereign purchases.
The continent exports the copper and imports the cable — the clearest beneficiation failure in the entire database.
Egypt's cable and transformer champions, South Africa's transformer base, Zambia and DR Congo on metal, Tunisia and Morocco on components, Kenya on geothermal expertise.
Designated-input rights for copper states inside every cable and transformer contract; finished-goods lanes split by voltage class; caps that keep the category plural.
Iron, steel and cement inside a USD 21.71bn line that every public-works programme on the continent draws down.
Infrastructure budgets make the state the dominant end-buyer even where contractors sign the import forms — the tender-level truth of the honesty seam.
Mauritania, Guinea and others ship the ore; the sections, sheet and coil sail back with the margin attached.
Egypt, Algeria, Nigeria, South Africa and Morocco's steel base; Zimbabwe's new capacity; cement at continental scale already crossing borders.
Product-class lanes, regional cement zones and ore-state input rights that stop one furnace complex owning the continent's rebar.
Medicaments, vaccines, blood products and dressings: what actually sits inside USD 18.35bn of Chapter 30 imports.
Essential-medicine lists, national tenders and donor co-financed programmes — the core of the USD 14.02bn health pillar.
Continental manufacturing covers only a fraction of the continent's own essential-medicines demand; vaccines sit nearer total dependency.
Egypt, Morocco, South Africa, Kenya and Nigeria's plants; Senegal's Institut Pasteur; Rwanda's new vaccine capacity; Algeria, Ghana, Ethiopia, Tunisia.
Therapeutic-class lanes, fill-finish rights first, an API beneficiation ladder behind them — and the match window sitting at every national tender.
Computing and network hardware inside a USD 15.97bn line — the machinery of every e-government promise.
National data centres, school-device programmes, digital-ID and connectivity roll-outs: the USD 8.53bn digital pillar in hardware form.
The physical layer of African data sovereignty is currently manufactured almost entirely on other continents.
Egypt and South Africa's electronics base; Kenya, Rwanda, Nigeria and Senegal's device-assembly and data-centre plays; Morocco's component corridor.
Split by device class, assemble first, deepen later — with security-critical lanes flagged for consortium treatment from day one.
Urea, phosphates, potash and blends: the USD 8.07bn line that decides next season's yields.
National fertiliser subsidy schemes make governments the anchor buyer across much of the continent.
World-scale African production already operates — yet member states still import blends from outside the continent.
Morocco's phosphate giant at world scale, Nigeria's urea complexes, Egypt and Algeria's nitrogen base, Togo, Senegal and Tunisia on phosphate.
Why fertiliser may pilot first: capability exists today, the fairness question is real, and the Cell Phone Test could run tomorrow morning.
Rolling stock, port cranes and earthmoving plant inside a USD 7.02bn line every corridor project depends on.
Railways, port authorities and public-works agencies: purchases so large that each one is a national event.
The trade area's own logistics backbone is currently bought almost entirely from outside it.
South Africa's rolling-stock and yellow-metal base, Egypt's wagon works, Morocco, Nigeria's new rail plant, Tunisia's components.
Lumpy demand needs consortium vehicles and order-book rotation rather than annual lanes — the hardest sequencing problem in the twelve.
Imaging, radiotherapy, dialysis, surgical and laboratory instruments: the USD 4.58bn precision tier of the health system.
Teaching-hospital builds and national equipment programmes concentrate demand into large, infrequent sovereign purchases.
No member state today manufactures an MRI machine — and service contracts deepen the dependency after every purchase.
South Africa's imaging innovators, Egypt and Tunisia's device plants, Morocco's electronics corridor — pieces of a machine no single state can build.
Why an African Medical Imaging Consortium — one designated entity, many flags — is this paper's flagship proposal for study.
Declared arms imports of USD 0.63bn against USD 3.40bn of measured defence procurement: the gap is itself the finding.
Defence buying is government by definition — and the least transparent line in the sovereign ledger.
African-led missions still source the bulk of their equipment off-continent, despite proven African platforms.
South Africa's defence industry, Egypt's industrial complex, Nigeria, Algeria and Kenya's growing capability — sensitive, real, already exporting.
Allocating the least transparent category: peer-state review, AU-mission procurement first, and disclosure standards the framework can live with.
The paper and polymer of statehood itself: a declared USD 2.57bn banknote line inside a USD 5.54bn sovereign-print pillar.
Every central bank, electoral commission and home-affairs ministry on the continent is a repeat customer — almost always abroad.
Most African currencies are printed outside Africa. The symbolism needs no elaboration.
South Africa's mint and note printers, Nigeria's security printing and minting company, Kenya, Egypt and Morocco's facilities — capacity with headroom between national runs.
Security printing turns on bilateral trust more than lowest price: regional print partnerships, peer audit — and why this small line might change the most minds.
The thinking behind the approach — specific where we can be, open where we must be. Circulated to accelerate a shared understanding of the problem and to invite member-state feedback before anything is fixed.
In 2023 the fifty-four member states of the African Union imported USD 709 billion of goods. Roughly USD 89 billion of that — about 12.5 per cent — came from other African countries. The remaining USD 620 billion left the continent entirely. That USD 620 billion is the prize the Buy African Initiative exists to contest, and this paper concerns the hardest, heaviest part of it.
Twelve categories — refined fuels, staple foods, vehicles, grid equipment, steel and cement, pharmaceuticals, ICT hardware, fertilisers, rail and heavy plant, advanced medical equipment, defence materiel, and the sovereign print of currency, identity and elections — account for USD 326.7 billion of the 2023 continental import bill: 46 per cent of everything the continent buys from anywhere. These are also, overwhelmingly, the categories where the state is the buyer. Of the USD 136.75 billion in measured government procurement across forty-eight member states in 2024, the seven sovereign pillars are led by exactly these lines: energy at USD 71.65 billion, agriculture and food at USD 27.03 billion, health at USD 14.02 billion, digital at USD 8.53 billion, infrastructure at USD 6.60 billion, the currency–government–elections complex at USD 5.54 billion, and defence at USD 3.40 billion.
In other words: the big twelve are where the money is, where the state signs, and where the Right of Supply either becomes an industrial policy of continental consequence — or a decorative preference scheme that changes nothing.
If these twelve categories could simply be handed out — copper cables to Zambia, refined fuel to Nigeria, pharmaceuticals to Egypt, vehicles to South Africa and Morocco — this would be a short document. It cannot be done that way, for one reason every finance minister will recognise the moment it is said aloud.
Capability in the big twelve is concentrated. Five or six economies could, between them, plausibly serve almost every line on the list. An allocation exercise run purely on present capability would award the majority of a USD 326.7 billion demand base — locked for twenty-five years — to a handful of member states. Those states would, in the phrase that started this paper, shoot the lights out. The other forty-plus would watch a framework built in the name of shared industrialisation function, in practice, as a preferential market for their largest neighbours. They would be right to refuse it. The framework would arrive at the STC already dead.
The opposite error is just as fatal. Scatter the big twelve evenly across fifty-four states regardless of capability, and the match window fails everywhere at once: quotes go unmatched, buyers release, contracts flow offshore exactly as they do today, and the framework is exposed within its first procurement cycle as a map of wishes. Allocation without capability is theatre. Capability without fairness is monopoly. The entire design question of the big twelve sits between those two failures — and the honest work of this paper is to hold both in view at once.
Allocation without capability is theatre. Capability without fairness is monopoly. The big twelve must be designed between those two failures.
The Right of Supply stands on two layers, and it is worth being precise about what each one carries. The allocation layer — the AU's pre-assignment of product categories to member states for the twenty-five-year framework term — is the fairness instrument. The match-or-release layer — the Cell Phone Test, in which the designated supplier must match a live open-market quote on price, quality, warranty and service, or release the buyer instantly and without penalty — is the discipline instrument.
Match-or-release already answers the most common fear about the big twelve: that preference means paying more. It cannot. No minister ever pays above the open-market quote, waits longer, or accepts less, because the moment the designated supplier misses on any of the four dimensions, the purchase is released — no penalty, no delay, no questions. What match-or-release does not do — cannot do — is prevent excellent suppliers from concentrating. If one country held the fuel category for the whole continent and matched every quote, it would win every contract, legitimately, for twenty-five years. Discipline is not distribution. For the ordinary run of product categories that is tolerable; for USD 326.7 billion it is not. The big twelve therefore require allocation engineering that the rest of the basket does not — and that engineering, not the principle, is what this paper puts before member states.
We are not proposing a single mechanism. We are proposing a toolkit, from which each of the twelve categories will take a different combination. Five instruments are under active study.
First: split the category. The category is not the unit of allocation; the product is. “Medical equipment” is not one right — imaging, dialysis, surgical instruments and laboratory diagnostics are separate lanes with different benches. “Fuel” is not one right — diesel, petrol, jet fuel, LPG and lubricants can carry different holders. Splitting the twelve categories into sixty-plus product lanes multiplies the number of member states that can hold something real, and it does so without inventing capability that does not exist. The sixty sub-sections of this paper's expanded table are the first draft of that lane map.
Second: zone the continent. Some products carry scale economics that no fairness formula can wish away — refining is the honest example. Where a single plant most efficiently serves a region rather than a continent, rights should be zoned: anchor suppliers designated per region, with defined satellite rights — storage, blending, distribution, maintenance — held by neighbouring states inside each zone. Nobody supplies fifty-four states. Anchors serve zones; satellites hold real, revenue-bearing rights inside them. Zoning converts an unavoidable concentration of plant into a deliberate distribution of participation.
Third: widen the entity. The framework's rule of one designated entity per country per category is deliberate, and it stands. But nothing prevents that one entity from being jointly owned. Where capability is genuinely continental rather than national — advanced medical imaging is the flagship case — the designated entity can be a purpose-built African consortium: assembly in one member state, components, software, service hubs and training academies in four or five others, with equity and board seats spread accordingly. One entity, many flags. The African Medical Imaging Consortium sketched in Section 10 is offered precisely as the case study for member states to react to.
Fourth: rights along the chain. A finished-goods allocation should not capture the whole value chain. Where a member state holds the mineral, it can hold a designated-input right: Zambian and Congolese copper written into every continental cable and transformer contract; Moroccan, Togolese and Senegalese phosphate inside the fertiliser lanes; Guinean and Mauritanian ore inside the steel lanes. The state that assembles does not swallow the states that supply. Chain rights turn resource endowments into standing commercial positions rather than talking points — and they are how the beneficiation ladder is paid for.
Fifth: stage the right and bound the share. Two guards run across every category. The ladder: rights can begin at assembly or fill-finish and mature to full manufacture as capability is proven, with aspirational allocations — the South Sudan Principle — carrying dated review points rather than empty promises. And the cap: consistent with the framework's Balance screen, no member state's designated rights across the big twelve should vastly exceed its own import bill, and no single state should hold more than a defined ceiling of continental demand in any one of the twelve. The cap is the sentence that keeps forty finance ministers reading.
Consultation is honest only if the current leaning is stated plainly, so here it is. We are leaning towards a standing rule that none of the twelve categories is ever allocated whole to a single member state. We are leaning towards lane-level allocation as the default; zoning for scale-bound products; consortium entities where capability exists only in combination; chain rights wherever the continent holds the mineral; ladders for every aspirational allocation; and hard share caps across the board. We are leaning towards fertiliser and pharmaceuticals as the first pilots — the first because world-scale capability already operates, the second because health sovereignty is the argument no minister needs persuading of — with refined fuels studied in parallel as the largest and hardest case. And we are leaning towards publishing the AI-driven matching methodology openly: the initial lane map and country benches will be produced by the declared AI matching exercise, put to member states as a starting frame, and refined country by country in consultation — never imposed.
None of this is fixed. All of it is the point of the exercise.
Each category section names a bench — the member states whose minerals, plant and expertise position them today. Nigeria, Algeria, Egypt and Angola on fuels; Ethiopia, Tanzania, Zambia and Sudan among the grain bench; South Africa and Morocco anchoring vehicles; Egypt and South Africa, with Zambia and DR Congo on metal, across the grid; Morocco's world-scale phosphates; Senegal and Rwanda in vaccines; and so on across the twelve. Two things must be said about these benches at once. They are real: the framework will not pretend capability exists where it does not, because the match window would expose the pretence within a quarter. And they are not a door: no bench in this paper is a closed list. A member state that believes it belongs on a bench — now, or by year five via the ladder — is invited to say so, with evidence, as part of this consultation. The bench is where the analysis starts. It is explicitly not where the allocation ends.
This paper succeeds if it comes back covered in ink. Specifically:
One. Which lanes, in which of the twelve categories, does your state hold capability for today — and which could it hold by year five under a laddered right? What evidence should the matching exercise weigh?
Two. Would your state participate in a consortium entity — as assembler, component supplier, service hub or equity partner — and what governance would make a multi-flag entity trustworthy enough to hold a designated right?
Three. What ceiling on any single state's share of a category — a quarter, a third — would your government regard as fair, and what floor of aspirational allocation should fragile member states carry under the South Sudan Principle?
Four. Which two categories should pilot first, and what would your ministry need to see from a pilot before extending the model?
Five. What, concretely, would cause your state to opt out — and what would cause it to champion this at the STC? We would rather hear the first answer now than discover it in the plenary.
Precedent is not in question. The United States operates Buy American across federal procurement; the European Union advances its own preference frameworks; India ties public demand to domestic plants through production-linked incentives. The world's largest industrial economies all use sovereign purchasing as industrial policy, daily and without apology. Africa is asking for the same right its trading partners already exercise. The difference is that Africa must exercise it as fifty-four states acting as one market — which is precisely why the allocation question this paper opens is harder here than anywhere else, and precisely why it is worth solving properly.
The big twelve are where the plants are biggest, the jobs deepest and the sovereignty most literal. A continent that refines its own fuel, stocks its own grain reserves, fills its own vaccine vials, scans its patients on machines assembled by an African consortium and pays for all of it in currency printed on African presses is not a slogan — it is USD 326.7 billion a year of demand, already being spent, waiting to be pointed home. Handled naively, that sum entrenches a monopoly of the few. Handled as this paper proposes to study it, it becomes the largest shared industrial programme this continent has ever attempted. The difference between those two outcomes is the allocation architecture — and the allocation architecture is now, formally, open for your feedback.
| Category | HS mapping | USD bn, 2023 | Coverage note |
|---|---|---|---|
| 01 Refined fuels | HS Chapter 27 | 141.32 | Refined product & gas (2710+2711) alone: 120.81 |
| 02 Staple foods | 1001 · 1005 · 1006 · 1101 · 1511 · 1701 | 46.93 | Wheat, maize, rice, flour, palm oil, sugar |
| 03 Vehicles & fleets | 8701–8705 | 37.05 | Tractors, buses, cars, trucks, special-purpose |
| 04 Power & grid | 8501–8504 · 8535–8537 · 8541 · 8544 | 22.50 | Motors, gensets, transformers, switchgear, PV, cable |
| 05 Steel & cement | Chapter 72 · 2523 | 21.71 | Iron & steel primary/semi-finished; cement |
| 06 Pharmaceuticals | Chapter 30 | 18.35 | Medicaments, vaccines, blood products, dressings |
| 07 Digital & ICT | 8471 · 8517 | 15.97 | Computing hardware; telephony & network equipment |
| 08 Fertilisers | Chapter 31 | 8.07 | Nitrogenous, phosphatic, potassic, blends |
| 09 Rail & heavy | Chapter 86 · 8426 · 8429 | 7.02 | Rolling stock; cranes; earthmoving plant |
| 10 Medical equipment | 9018–9022 | 4.58 | Imaging, radiotherapy, dialysis, surgical, laboratory |
| 11 Defence | Chapter 93 (declared) | 0.63 | Declared arms only; measured defence procurement pillar: 3.40 |
| 12 Sovereign print | 4907 (declared) | 2.57 | Banknotes & instruments; sovereign-print pillar (measured): 5.54 |
All category values are computed from the Africa Trade Intelligence Master Database v3 — HS6-level import data for all 54 AU member states, 2020–2024, compiled from UN Comtrade primary reporting with ITC Trade Map partner-mirror extension — using 2023 values. The continental total reconciles to USD 709.0 billion as the integrity check. Macro figures (USD 709bn total; USD 620bn off-continent; ~USD 89bn intra-African; USD 136.75bn measured government procurement across 48 states, 2024; the seven sovereign pillars) are the project's locked, sourced figures. Two categories are only partly visible to customs data: defence (declared Chapter 93 imports of USD 0.63bn against a measured procurement pillar of USD 3.40bn) and the sovereign print (declared banknote line USD 2.57bn against a measured pillar of USD 5.54bn); both are presented with the measurement basis stated. Category HS mappings are deliberately conservative; the full study will publish lane-level definitions. Country benches are analytical starting positions drawn from the productive-capacity audit series, not allocations.
This consultation paper was prepared by a combined research team of human analysts and AI research agents working under the direction of the Chair’s office. The AI-driven matching methodology referenced in this paper is declared by design. Figures trace to the sources stated in Section 04.
Prepared on behalf of the Hon. Neal Rijkenberg, Minister of Finance, Kingdom of Eswatini, in his capacity as Chair of the AU STC-FMAEPI. Circulated for member-state consultation — comment invited through your country’s draft mandate document or the response instrument on the main page.