Three plausible conditions in 2031 and 2036, built from Angola's own recorded
history. Not a prediction — a stated set of assumptions you can check, disagree with, and recompute.
New here? This is one of 54 country briefs on Africa 2036 Intelligence, an
evidence-based foresight instrument. Values marked with a year are measurements; values attached to a
scenario are projections computed from this country's own history. Where a value is missing it is shown
as missing, never estimated.
SADCECCASEvidence: ConsideredInitial country review
How much weight this outlook bears
The evidence base is reasonably complete, current and stable enough to reason about. Still a scenario, not a prediction.
Core indicator series are close to complete.
Most recent observation is from 2025 — current.
Growth has been variable (sd 3.12); the band is wide for good reason.
Country-specific policy and project evidence has been reviewed and cited.
Why this matters
The same seven questions are asked of all 54 countries, and every answer below is
computed from Angola's own evidence and model state — nothing here is written by hand, so
nothing here can drift away from the data it rests on.
7 of 7 questions can be answered from the evidence held for Angola.
What could materially change by 2031
Conditional projection
Household electricity access moves from 55.5% of the population (2024) to between 63% and 72% by 2031, depending on the scenario. Simple continuation gives 67% — a rise of 11.4 percentage points.
In people rather than percentages: about 15.4 million would still be without it in 2031 under continuation — 2 million fewer than in 2024, because the population grows as the share improves.
Internet use moves from 40.7% of the population (2024) to between 48% and 56% by 2031, depending on the scenario. Simple continuation gives 51% — a rise of 10.6 percentage points.
In people rather than percentages: about 22.6 million would still be without it in 2031 under continuation — 527,736 fewer than in 2024, because the population grows as the share improves.
Output per person spans $2,428 to $3,354 by 2031. The $926 between them is a quarter to a half of today's $3,129.
That spread is the distance between scenarios, not a margin of error — and it is a national average, which does not identify whose income moved.
The population reaches about 46.4 million — roughly 7.4 million more people than in 2025, about 25.4 million of them of working age.
This figure is the same in all three scenarios. Most of the people who will be alive in 2031 have already been born.
What could materially change by 2036
Conditional projection
Household electricity access moves from 55.5% of the population (2024) to between 68% and 79% by 2036, depending on the scenario. Simple continuation gives 73% — a rise of 17.7 percentage points.
In people rather than percentages: about 14.3 million would still be without it in 2036 under continuation — 3.1 million fewer than in 2024, because the population grows as the share improves.
Internet use moves from 40.7% of the population (2024) to between 52% and 64% by 2036, depending on the scenario. Simple continuation gives 58% — a rise of 16.9 percentage points.
In people rather than percentages: about 22.6 million would still be without it in 2036 under continuation — 585,311 fewer than in 2024, because the population grows as the share improves.
Output per person spans $1,966 to $3,553 by 2036. The $1,587 between them is wider than half today's $3,129.
That spread is the distance between scenarios, not a margin of error — and it is a national average, which does not identify whose income moved.
The population reaches about 53.2 million — roughly 14.2 million more people than in 2025, about 29.8 million of them of working age.
This figure is the same in all three scenarios. Most of the people who will be alive in 2036 have already been born.
Who may benefit
Interpretation
Young people entering a labour market that has been prepared for them
Smallholder farmers, if input costs, storage and roads improve alongside prices
Households and small businesses reached by new connections
The public finances, when prices are high
People connecting for the first time, who move from no access to some — a larger step than any later improvement in speed or price
Construction, transport and services workers in growing cities
the 71% of people aged 15 and over with neither a bank nor a mobile-money account in 2014, if payment access widens before the projected connectivity does
rural households carrying a 54-point electricity gap, who have the most to gain from any of it closing
people coming online for the first time — internet use rises 23 percentage points by 2036 under the fastest band, which is roughly 18.2 million more people connected than in 2024
Who may be excluded
Interpretation
A cohort that reaches working age faster than jobs, training places or clinics are built
Rain-fed farming households, first and hardest, in a bad season
Educated young people whose qualifications do not convert into work
A state that must regulate technologies it has too few trained people to assess
Off-grid communities, whose wait lengthens if capital concentrates on urban and industrial load
Everyone dependent on public salaries and services when prices fall
Rural, older, poorer and less literate people, who are last in every connectivity sequence
Households already spending most of their income on food
New urban arrivals without tenure security, water or transport
Firms whose expansion depends on credit, in an economy holding little of it relative to output
rural households, 54 points behind on electricity in 2024 — a gap that closes last in every scenario here because access improves fastest where it is already highest
the 59% of people not online in 2024, for whom every digital projection on this page is a statement about somebody else
young people already outside work at 27.2% in 2025, in a labour market the projections grow but do not restructure
What must happen for this to be plausible
From the evidence
About US$753 million was committed for the Angolan section by development finance institutions in late 2025 (US DFC US$553m; DBSA…. Recorded as funded: it still has to actually reach recipients rather than sit in an envelope before it counts as money in motion.
For the Acceleration band to describe the decade, growth per person has to hold near 1.2% a year. This country has reached that rate before — its best year in the window was 4.6% — but has not sustained it across the 15-year record the band is drawn from.
What could invalidate this outlook
Interpretation
Commodity dependence — A sustained move in the country's principal export price of more than roughly a third. The band averages across a cycle and understates both tails.
Population pressure & the youth cohort — A significant revision to the UN World Population Prospects for this country. The population path here is carried through unchanged and is the least uncertain input on the page.
Education & workforce capability — A sustained collapse or expansion in enrolment, or large-scale emigration of trained professionals.
Climate exposure & disaster risk — A multi-season drought or a major flood event affecting the main producing regions. Discrete climate events are outside any band built from recent averages.
Energy availability — Either a step change in generation and transmission delivery, or a sustained supply failure. Both break a projection built on gradual access gains.
Urbanisation — A major change in land, housing or transport policy in the principal city.
Digital infrastructure — A step change in device or data affordability, or a major submarine cable landing.
Institutional capacity & regulatory reform — A change of government that materially alters delivery capacity, in either direction.
Food & water systems — A regional harvest failure, or a change in the affordability of food imports.
What to watch now
From the evidence
Real GDP per capita growth sustained above 1.16% for three consecutive years → The Acceleration band would be the better description of the decade.
Real GDP per capita growth below -4.14% for two consecutive years → The Disruption band would be the better description of the decade.
Electricity access above 71.5% by 2031 (momentum path reaches 66.9%) → Energy access is running ahead of the momentum path.
Commodity dependence, Population pressure & the youth cohort and Education & workforce capability are assessed as binding here. Movement on them would change the outlook faster than movement inside the growth band.
This country's review covers 2 of the five target evidence areas. Treat the outlook as a floor for what could be said about it, not a ceiling.
The three futures
The assumption, stated in full
Annual real GDP growth minus annual population growth, 2011–2025 (15 years). Mean -1.49%, standard deviation 3.12. Scenario rates are mean ± 0.85 sd, clamped to this country's own 5th–95th percentile and to [-6%, +9%].
Momentum — -1.49% a year
Present trajectories broadly continue.
Policy, investment, institutions and demographics carry on behaving as they have. No collapse, no breakthrough. This is not a forecast of what will happen — it is the shape of the recent past extended forward, which is the baseline every other scenario should be judged against.
The recent past, extended. The question it asks is whether that is enough.
For Angola this is -1.49% a year per person,
drawn from its own 15-year record (2011–2025),
not from an outside view of what this country might do.
What it is good at
Predictability. Institutions, tariffs, procurement and delivery capacity behave as they already do, so plans made today mostly survive contact with the decade. Existing programmes finish. Nothing has to be rebuilt.
What it costs
Everything that is currently too slow stays too slow. Where a gap is closing at two points a year and the population is growing at three percent, continuation means the absolute number of people excluded rises even as the percentage falls.
Who it reaches last
Whoever is already last. Continuation preserves the existing sequence of who gets connected, treated, schooled and paid — it does not reorder it.
What it quietly assumes
That the conditions of the last fifteen years hold for the next ten: no default, no major conflict, no discrete climate event, and no external shock large enough to break the trend the band is drawn from.
How it is usually misread
Momentum is routinely read as the safe or neutral case. It is neither. It is the case in which nothing is done differently, and for several countries on this platform that is the most consequential choice available.
Measure
2026
2031
2036
Population (same in all scenarios)
40.2m
46.4m
53.2m
Working-age population (15–64) (same in all scenarios)
21.4m
25.4m
29.8m
Real GDP per capita
$3,083
$2,860
$2,653
Electricity access
59.1%
66.9%
73.2%
Internet use
43.9%
51.3%
57.6%
Acceleration — 1.16% a year
The country sustains the pace of its own better years.
Reform, investment, regional trade, energy build-out, education and institutional capacity perform above the recent average — at a rate this country has actually reached before, held for a decade rather than a year or two.
Not a miracle — this country's own good years, held for a decade instead of a season.
For Angola this is 1.16% a year per person,
drawn from its own 15-year record (2011–2025),
not from an outside view of what this country might do.
What it is good at
Compounding. A rate held for ten years does something a rate held for two cannot: it changes the level, not just the direction. Access gaps close inside the horizon rather than beyond it, and the working-age bulge arrives into an economy that has grown to meet it.
What it costs
Speed is unevenly distributed by default. Growth concentrates where infrastructure, credit and skills already are, which in most of these economies means the largest city and the formal sector. Faster national numbers can coexist with a widening internal gap, and this platform cannot see that gap because the indicators behind it are national.
Who it reaches last
Rural, informal and non-connected populations, unless something specific is done to reach them. Acceleration reaches them faster in absolute terms and no sooner in sequence.
What it quietly assumes
Sustained implementation capacity — the same institutions delivering at their best, continuously, for a decade. It is the strongest assumption on this platform, and the historical record for holding a peak rate that long is thin everywhere, not only here.
How it is usually misread
Acceleration is routinely read as the target. It is a description of a pace, not of a distribution, and it says nothing about who the growth reaches.
Measure
2026
2031
2036
Population (same in all scenarios)
40.2m
46.4m
53.2m
Working-age population (15–64) (same in all scenarios)
21.4m
25.4m
29.8m
Real GDP per capita
$3,166
$3,354
$3,553
Electricity access
60.8%
71.5%
79.3%
Internet use
45.5%
55.8%
64.1%
Disruption — -4.14% a year
The country runs at the pace of its own worse years.
Debt service, weak implementation, instability, climate shocks, capital flight or external shocks hold performance at the low end of realised experience. Note the limit: this band is drawn from recent history, so it does not represent a war, a default or a catastrophic climate event. Those are listed as named risks instead of being given a false number.
Not collapse. The low end of what this country has already survived — which is the point.
For Angola this is -4.14% a year per person,
drawn from its own 15-year record (2011–2025),
not from an outside view of what this country might do.
What it is good at
Clarity about what is load-bearing. The commitments that survive a bad decade are the ones with financing already closed and construction already started; everything at announcement stage is what disappears first. Disruption is the scenario that separates the two.
What it costs
Time. Gaps that would close inside the horizon move outside it, and a cohort passes through school, into work and into household formation while the conditions do not improve. That cost is paid by specific ages of specific people and is not recoverable later.
Who it reaches last
Nobody new is reached. The people this scenario hits first are those with the least buffer — households already spending most of their income on food, workers in rain-fed agriculture, and anyone whose access depends on a service that gets cut before it gets extended.
What it quietly assumes
That the bad years look like the bad years already in the record. It does NOT model war, sovereign default or a catastrophic climate event — those are outside any band built from realised history, and this platform names them as risks instead of giving them a number it cannot support.
How it is usually misread
Disruption is routinely read as the collapse case. It is the opposite: it is bounded by what has already happened, which makes it the most conservative of the three about how bad things could get.
Measure
2026
2031
2036
Population (same in all scenarios)
40.2m
46.4m
53.2m
Working-age population (15–64) (same in all scenarios)
21.4m
25.4m
29.8m
Real GDP per capita
$3,000
$2,428
$1,966
Electricity access
57.9%
63.3%
68%
Internet use
42.8%
47.8%
52.3%
GDP per capita paths are expressed in constant present-day dollars — a real-output path, not a forecast
of prices or exchange rates. Population comes from the UN World Population Prospects and is carried
through unchanged, which is why it does not vary between scenarios: ten-year demographic momentum is
close to fixed.
Early signals
Thresholds derived from the model itself. You can check which band reality is tracking without waiting for us to tell you.
Real GDP per capita growth sustained above 1.16% for three consecutive years
The Acceleration band would be the better description of the decade.
Real GDP per capita growth below -4.14% for two consecutive years
The Disruption band would be the better description of the decade.
Electricity access above 71.5% by 2031 (momentum path reaches 66.9%)
Energy access is running ahead of the momentum path.
Internet use above 55.8% by 2031 (momentum path reaches 51.3%)
Digital foundations are running ahead of the momentum path.
Lived experience
What the projected numbers would mean for ten representative situations. Not predictions, and not stories about real people.
Momentum · 2036
What the projected numbers would mean, taken together, for ten representative situations under
Momentum in 2036.
These are not predictions and not stories about real people. Each one states the evidence it rests
on, the assumptions it makes, and who the improvement would miss.
A young person entering the workforce
Roughly 8.4 million more people reach working age between 2026 and 2036. They enter a labour market where 27.2% of 15–24s were already counted as unemployed in 2025, and where 52% of all work is in agriculture (2025). Under Momentum, internet use moves from 41% to 58% — a materially larger share of this cohort can reach work, training and customers beyond walking distance. With tertiary enrolment at 10%, most will not pass through formal higher education, so what they can do at 25 is set by what employers and informal apprenticeship actually teach — not by universities.
Better if
Technical training is aligned to what employers actually hire for, and the evidence on which courses lead to work is published.
Worse if
The cohort arrives faster than jobs, training places or electricity, in which case a demographic dividend becomes a grievance.
Who this misses
Young people in farming households, who typically leave school earliest and are furthest from the connectivity and training that the projection describes.
Assumptions and evidence
Working-age population follows the UN World Population Prospects projection, unchanged.
Youth unemployment is an ILO modelled estimate and understates underemployment in largely informal labour markets.
Connectivity gains are assumed to reach young people at least as fast as the population average — historically optimistic.
Population aged 0–14 44.1% · 2025Youth unemployment, 15–24 (ILO modelled) 27.2% · 2025School enrolment, tertiary (gross) 10% · 2024Employment in agriculture 51.9% · 2025Individuals using the internet 40.7% · 2024
A small-business owner
Electricity access moves from 56% of the population (2024) to 73% by 2036 under Momentum. That measures connections, not hours of supply: a workshop counted as served may still be waiting on power for part of the day, and this platform holds no outage or tariff data for it. Domestic credit to the private sector stood at 6.6% of GDP in 2025, which is thin by any comparison. Plausible implication, not measured: a credit-to-GDP ratio this low is consistent with borrowing being hard for small firms, but the ratio is an aggregate and cannot show which firms were refused or why. A firm-level credit survey would establish it. 29% of people aged 15 and over held an account at a bank or other financial institution, or personally used a mobile-money service, in 2014. Mobile money is counted. Plausible implication, not measured: account records are the kind of evidence a lender can price risk against, so widening ownership could open borrowing to traders without land title. Nothing in this platform measures whether that has happened here — lender practice, or credit-bureau coverage, would show it. Average output per person in the surrounding market moves from $3,129 to $2,653 a year. That is a national mean and does not identify whose income changed.
Better if
Collateral registries, insolvency procedure and payment rails are fixed before credit is subsidised.
Worse if
Credit is expanded into a weak enforcement system, producing bad loans rather than businesses.
Who this misses
Unregistered traders, who are the majority of businesses in most of these economies and are invisible to both the credit statistics and the policies built on them.
Assumptions and evidence
GDP per capita is expressed in constant present-day dollars and describes average output, not the income of any particular household.
Electricity access measures connection, not hours of supply or price. A connected business with eight-hour outages is counted as served.
Domestic credit to the private sector is a stock ratio. It does not distinguish lending to large firms from lending to small ones, and most of the businesses this situation describes are unregistered and outside it entirely.
Access to electricity 55.5% · 2024Domestic credit to private sector 6.6% · 2025Account at a bank or mobile-money provider, age 15+ 29.3% · 2014GDP per capita $3,129 · 2025
A farmer or food producer
52% of the country's workers were in agriculture in 2025. Cereal yields averaged 1,062 kg per hectare — well below what the same seed achieves with reliable inputs, water and storage, so the largest available productivity gain is not a new technology but the ordinary one already used elsewhere. By 2036, 73% of the population has electricity under Momentum — the precondition for irrigation pumps, cold storage and anything that stops a harvest rotting between field and market. 58% internet use puts price information, weather and pest identification within reach of a phone, for those who have one and can afford data. Undernourishment affected 22.5% of people in 2023: food exists in the region and cannot affordably be moved, which is a roads and storage problem more than a farming one.
Better if
Investment goes into what happens after harvest — storage, roads, cold chain, market information — not only into what happens before it.
Worse if
A multi-season drought arrives, which no band on this platform models, or input costs rise faster than farmgate prices.
Who this misses
Rain-fed smallholders without irrigation, who feel a bad season first and hardest, and women farmers, who in most of these countries hold weaker land tenure and less access to credit.
Assumptions and evidence
National electricity and connectivity projections are applied to a rural situation, though rural access consistently lags the national figure.
Yield gaps are treated as closeable with existing technology; that assumes input supply, credit and extension services that may not exist.
Employment in agriculture 51.9% · 2025Cereal yield 1,062 kg/ha · 2023Prevalence of undernourishment 22.5% · 2023Freshwater withdrawal as share of available resources 1.9% · 2022
A healthcare worker, and a patient
There were 2.4 physicians per 10,000 people in 2022. The population they serve grows by 13 million between 2026 and 2036, so holding the current ratio requires training and retaining staff at that rate merely to stand still. Electricity reaching 73% by 2036 under Momentum is what makes a vaccine cold chain, a functioning theatre and a night-time delivery possible outside the largest towns. Out-of-pocket payments were 33% of health spending in 2023. Under-five mortality stood at 49 per 1,000 live births (2024).
Better if
Health financing moves toward pooled, prepaid arrangements, and task-shifting is supported by law rather than tolerated informally.
Worse if
Trained clinicians emigrate faster than they are replaced, or external health financing contracts.
Who this misses
People in rural districts and informal settlements, where the physician ratio is a fraction of the national figure, and anyone whose care is rationed by ability to pay at the point of service.
Assumptions and evidence
Workforce density is a national average; it conceals extreme concentration in capital cities.
This platform holds no projection for health workforce or financing — only the current position against a projected population.
Physicians per 1,000 people 0.24 / 1,000 · 2022Out-of-pocket health expenditure 32.5% · 2023Under-5 mortality rate 49 / 1,000 · 2024
A teacher, and a student
The school-age population grows with a national population rising by 13 million to 2036. Gross secondary enrolment was 51.5% in 2023 and tertiary 10%. Government spent 2.5% of GDP on education in 2023. Under Momentum, 73% of households have electricity by 2036 — which is when homework after dark stops being a privilege. 58% internet use makes digital material reachable for that share of students, and no more; 42% of the population remains outside it. Adult literacy at 68% (2023) means many of these students cannot be helped with schoolwork at home, whatever the school does.
Better if
Electricity and connectivity reach schools ahead of the national average rather than behind it, and teacher supply grows with enrolment.
Worse if
Enrolment expands faster than teachers are trained, converting access gains into larger classes and weaker learning.
Who this misses
Girls in regions where secondary completion diverges sharply by gender, and children in households where the opportunity cost of school attendance is immediate income.
Assumptions and evidence
Gross enrolment ratios can exceed 100% and include students outside the standard age range; they measure participation, not learning.
This platform holds no measure of learning outcomes, teacher supply or classroom quality — a substantial gap in describing education.
By 2036 under Momentum, roughly 30.7 million people in this country are online — the domestic audience reachable without leaving it. That is up from about 41% of the population in 2024 to 58%. 29% of people aged 15 and over held a bank or mobile-money account in 2014, so being paid directly by that audience is mechanically possible for a minority of it. A population where 44% were under 15 in 2025 is, within this horizon, an unusually young consumer market for music, film, fashion and games. Production still depends on power: 73% household electricity access by 2036 sets how far studios, editing and streaming can spread beyond the main city.
Better if
Payment rails work across borders and royalties are collected and enforced domestically rather than captured offshore.
Worse if
Data costs stay high relative to income, or distribution platforms owned elsewhere capture the value created here.
Who this misses
Creators working in languages the major platforms do not support well, and anyone without a device capable of running current tools.
Assumptions and evidence
Domestic reach is calculated as projected internet use multiplied by projected population. It counts people who can technically be reached, not an audience.
This platform holds no reviewed creative-sector evidence for this country. Reach and payment access are stand-ins for a sector nobody has measured here.
Individuals using the internet 40.7% · 2024Account at a bank or mobile-money provider, age 15+ 29.3% · 2014Population aged 0–14 44.1% · 2025
A trader using regional markets
Trade was 35% of GDP in 2025, so what happens at a neighbour's border arrives as a domestic price within months. This country sits inside SADC, ECCAS. AfCFTA has been signed by every African Union member state except Eritrea, but tariff schedules, rules of origin and customs practice — not the treaty — decide whether a small consignment moves more cheaply in 2036 than it does today. 29% of people aged 15 and over held a bank or mobile-money account in 2014 — the share able to settle across a border without carrying cash.
Better if
Rules of origin are agreed on contested product lines and applied at national customs, not just gazetted.
Worse if
Regional blocs fragment further, or preferential tariffs remain unimplemented at the border despite ratification.
Who this misses
Small cross-border traders, disproportionately women, who operate below the threshold at which formal preferential regimes are usable at all.
Assumptions and evidence
Trade openness is a national accounts ratio and says nothing about the cost or time of a specific border crossing.
This platform holds no country-level data on intra-African trade share, non-tariff barriers or border clearance times — the measures that would matter most here.
Trade (exports + imports) 34.6% · 2025Account at a bank or mobile-money provider, age 15+ 29.3% · 2014
A family in a growing city
Urban population was growing 4.1% a year in 2025, doubling roughly every 17 years. On the projected population, of the order of 9.3 million additional people are living in this country's towns and cities by 2036. 63% of urban residents were in informal settlements in 2022 — housing built faster than it was serviced. Under Momentum, 73% of the population has electricity by 2036. The decisions that determine whether this family's commute is one hour or three — where roads, water mains and drainage go, and which land is reserved before it is built on — are being made now, and are expensive to reverse afterwards.
Better if
Land for roads, water mains and drainage is bought and reserved before settlement arrives — the cheapest infrastructure decision available and the one most often skipped.
Worse if
Growth outpaces servicing, in which case today's new neighbourhoods become the next decade's retrofit problem at many times the cost.
Who this misses
Renters and residents without secure tenure, who bear eviction risk and are usually the first displaced when upgrading finally arrives.
Assumptions and evidence
Urban share is held constant when estimating additional urban residents; in practice it is rising, so this is a conservative figure.
This platform holds no city-level data. National averages conceal very large differences between a capital and a secondary town.
Urban population 71.5% · 2025Urban population growth 4.1% · 2025Population living in slums 62.7% · 2022
A citizen dealing with government
Government effectiveness scored -0.8 in 2024 on a scale running roughly −2.5 to +2.5. With 58% internet use by 2036 under Momentum, permits, payments, registrations and records can plausibly be handled without a physical queue for that share of the population — the most tangible improvement most citizens would actually notice. Voice and accountability scored -1.1 (2024). The same identity systems, registries and monitoring capability that shorten a queue are also what surveillance is built from, and the measures that would constrain misuse currently score in the lower band. Both outcomes are live and depend on choices not yet made.
Better if
Digital public infrastructure is built with data protection law, independent oversight and offline fallbacks from the start.
Worse if
Identity and payment systems become a precondition for services before coverage is universal, excluding people from entitlements they already hold.
Who this misses
People without formal identity documents, who are disproportionately rural, poor, migrant or stateless — and who are excluded twice when services move online.
Assumptions and evidence
Governance indicators are perception-based composites. They signal direction and relative position, not measured administrative performance.
Connectivity is assumed to be usable for government services; in practice affordability, literacy and identity documentation each gate access again.
Government effectiveness -0.79 · 2024Voice and accountability -1.06 · 2024Government revenue excl. grants 18.2% · 2024
A member of the diaspora weighing return or investment
Remittances were 0% of GDP in 2025, against foreign direct investment at 0.9% — meaning money sent home by citizens abroad is a significant and unusually reliable external flow. Under Momentum, real output per person moves from $3,129 (2025) to $2,653 by 2036, about -1.5% a year. By 2036, 73% electricity access and 58% internet use set what can actually be operated on the ground. Rule of law scored -1.2 (2024), which is the measure that most often decides whether a diaspora investor commits capital rather than only sending support.
Better if
Property registration, repatriation of profits and dispute resolution are made predictable — the three things that convert diaspora sentiment into diaspora capital.
Worse if
Currency controls or an unpredictable tax treatment of returning residents make commitment more expensive than continued remittance.
Who this misses
Diaspora members without capital to invest, whose contribution is already the largest single external flow and who are rarely the audience for investment policy.
Assumptions and evidence
GDP per capita paths are in constant present-day dollars and assume no exchange-rate or price forecast.
Remittance ratios are balance-of-payments estimates and exclude informal channels, which are substantial in most of these countries.
Personal remittances received 0% · 2025Foreign direct investment, net inflows 0.9% · 2025GDP per capita $3,129 · 2025Rule of law -1.17 · 2024
Acceleration · 2036
What the projected numbers would mean, taken together, for ten representative situations under
Acceleration in 2036.
These are not predictions and not stories about real people. Each one states the evidence it rests
on, the assumptions it makes, and who the improvement would miss.
A young person entering the workforce
Roughly 8.4 million more people reach working age between 2026 and 2036. They enter a labour market where 27.2% of 15–24s were already counted as unemployed in 2025, and where 52% of all work is in agriculture (2025). Under Acceleration, internet use moves from 41% to 64% — a materially larger share of this cohort can reach work, training and customers beyond walking distance. With tertiary enrolment at 10%, most will not pass through formal higher education, so what they can do at 25 is set by what employers and informal apprenticeship actually teach — not by universities.
Better if
Technical training is aligned to what employers actually hire for, and the evidence on which courses lead to work is published.
Worse if
The cohort arrives faster than jobs, training places or electricity, in which case a demographic dividend becomes a grievance.
Who this misses
Young people in farming households, who typically leave school earliest and are furthest from the connectivity and training that the projection describes.
Assumptions and evidence
Working-age population follows the UN World Population Prospects projection, unchanged.
Youth unemployment is an ILO modelled estimate and understates underemployment in largely informal labour markets.
Connectivity gains are assumed to reach young people at least as fast as the population average — historically optimistic.
Population aged 0–14 44.1% · 2025Youth unemployment, 15–24 (ILO modelled) 27.2% · 2025School enrolment, tertiary (gross) 10% · 2024Employment in agriculture 51.9% · 2025Individuals using the internet 40.7% · 2024
A small-business owner
Electricity access moves from 56% of the population (2024) to 79% by 2036 under Acceleration. That measures connections, not hours of supply: a workshop counted as served may still be waiting on power for part of the day, and this platform holds no outage or tariff data for it. Domestic credit to the private sector stood at 6.6% of GDP in 2025, which is thin by any comparison. Plausible implication, not measured: a credit-to-GDP ratio this low is consistent with borrowing being hard for small firms, but the ratio is an aggregate and cannot show which firms were refused or why. A firm-level credit survey would establish it. 29% of people aged 15 and over held an account at a bank or other financial institution, or personally used a mobile-money service, in 2014. Mobile money is counted. Plausible implication, not measured: account records are the kind of evidence a lender can price risk against, so widening ownership could open borrowing to traders without land title. Nothing in this platform measures whether that has happened here — lender practice, or credit-bureau coverage, would show it. Average output per person in the surrounding market moves from $3,129 to $3,553 a year. That is a national mean and does not identify whose income changed.
Better if
Collateral registries, insolvency procedure and payment rails are fixed before credit is subsidised.
Worse if
Credit is expanded into a weak enforcement system, producing bad loans rather than businesses.
Who this misses
Unregistered traders, who are the majority of businesses in most of these economies and are invisible to both the credit statistics and the policies built on them.
Assumptions and evidence
GDP per capita is expressed in constant present-day dollars and describes average output, not the income of any particular household.
Electricity access measures connection, not hours of supply or price. A connected business with eight-hour outages is counted as served.
Domestic credit to the private sector is a stock ratio. It does not distinguish lending to large firms from lending to small ones, and most of the businesses this situation describes are unregistered and outside it entirely.
Access to electricity 55.5% · 2024Domestic credit to private sector 6.6% · 2025Account at a bank or mobile-money provider, age 15+ 29.3% · 2014GDP per capita $3,129 · 2025
A farmer or food producer
52% of the country's workers were in agriculture in 2025. Cereal yields averaged 1,062 kg per hectare — well below what the same seed achieves with reliable inputs, water and storage, so the largest available productivity gain is not a new technology but the ordinary one already used elsewhere. By 2036, 79% of the population has electricity under Acceleration — the precondition for irrigation pumps, cold storage and anything that stops a harvest rotting between field and market. 64% internet use puts price information, weather and pest identification within reach of a phone, for those who have one and can afford data. Undernourishment affected 22.5% of people in 2023: food exists in the region and cannot affordably be moved, which is a roads and storage problem more than a farming one.
Better if
Investment goes into what happens after harvest — storage, roads, cold chain, market information — not only into what happens before it.
Worse if
A multi-season drought arrives, which no band on this platform models, or input costs rise faster than farmgate prices.
Who this misses
Rain-fed smallholders without irrigation, who feel a bad season first and hardest, and women farmers, who in most of these countries hold weaker land tenure and less access to credit.
Assumptions and evidence
National electricity and connectivity projections are applied to a rural situation, though rural access consistently lags the national figure.
Yield gaps are treated as closeable with existing technology; that assumes input supply, credit and extension services that may not exist.
Employment in agriculture 51.9% · 2025Cereal yield 1,062 kg/ha · 2023Prevalence of undernourishment 22.5% · 2023Freshwater withdrawal as share of available resources 1.9% · 2022
A healthcare worker, and a patient
There were 2.4 physicians per 10,000 people in 2022. The population they serve grows by 13 million between 2026 and 2036, so holding the current ratio requires training and retaining staff at that rate merely to stand still. Electricity reaching 79% by 2036 under Acceleration is what makes a vaccine cold chain, a functioning theatre and a night-time delivery possible outside the largest towns. Out-of-pocket payments were 33% of health spending in 2023. Under-five mortality stood at 49 per 1,000 live births (2024).
Better if
Health financing moves toward pooled, prepaid arrangements, and task-shifting is supported by law rather than tolerated informally.
Worse if
Trained clinicians emigrate faster than they are replaced, or external health financing contracts.
Who this misses
People in rural districts and informal settlements, where the physician ratio is a fraction of the national figure, and anyone whose care is rationed by ability to pay at the point of service.
Assumptions and evidence
Workforce density is a national average; it conceals extreme concentration in capital cities.
This platform holds no projection for health workforce or financing — only the current position against a projected population.
Physicians per 1,000 people 0.24 / 1,000 · 2022Out-of-pocket health expenditure 32.5% · 2023Under-5 mortality rate 49 / 1,000 · 2024
A teacher, and a student
The school-age population grows with a national population rising by 13 million to 2036. Gross secondary enrolment was 51.5% in 2023 and tertiary 10%. Government spent 2.5% of GDP on education in 2023. Under Acceleration, 79% of households have electricity by 2036 — which is when homework after dark stops being a privilege. 64% internet use makes digital material reachable for that share of students, and no more; 36% of the population remains outside it. Adult literacy at 68% (2023) means many of these students cannot be helped with schoolwork at home, whatever the school does.
Better if
Electricity and connectivity reach schools ahead of the national average rather than behind it, and teacher supply grows with enrolment.
Worse if
Enrolment expands faster than teachers are trained, converting access gains into larger classes and weaker learning.
Who this misses
Girls in regions where secondary completion diverges sharply by gender, and children in households where the opportunity cost of school attendance is immediate income.
Assumptions and evidence
Gross enrolment ratios can exceed 100% and include students outside the standard age range; they measure participation, not learning.
This platform holds no measure of learning outcomes, teacher supply or classroom quality — a substantial gap in describing education.
By 2036 under Acceleration, roughly 34.1 million people in this country are online — the domestic audience reachable without leaving it. That is up from about 41% of the population in 2024 to 64%. 29% of people aged 15 and over held a bank or mobile-money account in 2014, so being paid directly by that audience is mechanically possible for a minority of it. A population where 44% were under 15 in 2025 is, within this horizon, an unusually young consumer market for music, film, fashion and games. Production still depends on power: 79% household electricity access by 2036 sets how far studios, editing and streaming can spread beyond the main city.
Better if
Payment rails work across borders and royalties are collected and enforced domestically rather than captured offshore.
Worse if
Data costs stay high relative to income, or distribution platforms owned elsewhere capture the value created here.
Who this misses
Creators working in languages the major platforms do not support well, and anyone without a device capable of running current tools.
Assumptions and evidence
Domestic reach is calculated as projected internet use multiplied by projected population. It counts people who can technically be reached, not an audience.
This platform holds no reviewed creative-sector evidence for this country. Reach and payment access are stand-ins for a sector nobody has measured here.
Individuals using the internet 40.7% · 2024Account at a bank or mobile-money provider, age 15+ 29.3% · 2014Population aged 0–14 44.1% · 2025
A trader using regional markets
Trade was 35% of GDP in 2025, so what happens at a neighbour's border arrives as a domestic price within months. This country sits inside SADC, ECCAS. AfCFTA has been signed by every African Union member state except Eritrea, but tariff schedules, rules of origin and customs practice — not the treaty — decide whether a small consignment moves more cheaply in 2036 than it does today. 29% of people aged 15 and over held a bank or mobile-money account in 2014 — the share able to settle across a border without carrying cash.
Better if
Rules of origin are agreed on contested product lines and applied at national customs, not just gazetted.
Worse if
Regional blocs fragment further, or preferential tariffs remain unimplemented at the border despite ratification.
Who this misses
Small cross-border traders, disproportionately women, who operate below the threshold at which formal preferential regimes are usable at all.
Assumptions and evidence
Trade openness is a national accounts ratio and says nothing about the cost or time of a specific border crossing.
This platform holds no country-level data on intra-African trade share, non-tariff barriers or border clearance times — the measures that would matter most here.
Trade (exports + imports) 34.6% · 2025Account at a bank or mobile-money provider, age 15+ 29.3% · 2014
A family in a growing city
Urban population was growing 4.1% a year in 2025, doubling roughly every 17 years. On the projected population, of the order of 9.3 million additional people are living in this country's towns and cities by 2036. 63% of urban residents were in informal settlements in 2022 — housing built faster than it was serviced. Under Acceleration, 79% of the population has electricity by 2036. The decisions that determine whether this family's commute is one hour or three — where roads, water mains and drainage go, and which land is reserved before it is built on — are being made now, and are expensive to reverse afterwards.
Better if
Land for roads, water mains and drainage is bought and reserved before settlement arrives — the cheapest infrastructure decision available and the one most often skipped.
Worse if
Growth outpaces servicing, in which case today's new neighbourhoods become the next decade's retrofit problem at many times the cost.
Who this misses
Renters and residents without secure tenure, who bear eviction risk and are usually the first displaced when upgrading finally arrives.
Assumptions and evidence
Urban share is held constant when estimating additional urban residents; in practice it is rising, so this is a conservative figure.
This platform holds no city-level data. National averages conceal very large differences between a capital and a secondary town.
Urban population 71.5% · 2025Urban population growth 4.1% · 2025Population living in slums 62.7% · 2022
A citizen dealing with government
Government effectiveness scored -0.8 in 2024 on a scale running roughly −2.5 to +2.5. With 64% internet use by 2036 under Acceleration, permits, payments, registrations and records can plausibly be handled without a physical queue for that share of the population — the most tangible improvement most citizens would actually notice. Voice and accountability scored -1.1 (2024). The same identity systems, registries and monitoring capability that shorten a queue are also what surveillance is built from, and the measures that would constrain misuse currently score in the lower band. Both outcomes are live and depend on choices not yet made.
Better if
Digital public infrastructure is built with data protection law, independent oversight and offline fallbacks from the start.
Worse if
Identity and payment systems become a precondition for services before coverage is universal, excluding people from entitlements they already hold.
Who this misses
People without formal identity documents, who are disproportionately rural, poor, migrant or stateless — and who are excluded twice when services move online.
Assumptions and evidence
Governance indicators are perception-based composites. They signal direction and relative position, not measured administrative performance.
Connectivity is assumed to be usable for government services; in practice affordability, literacy and identity documentation each gate access again.
Government effectiveness -0.79 · 2024Voice and accountability -1.06 · 2024Government revenue excl. grants 18.2% · 2024
A member of the diaspora weighing return or investment
Remittances were 0% of GDP in 2025, against foreign direct investment at 0.9% — meaning money sent home by citizens abroad is a significant and unusually reliable external flow. Under Acceleration, real output per person moves from $3,129 (2025) to $3,553 by 2036, about 1.2% a year. By 2036, 79% electricity access and 64% internet use set what can actually be operated on the ground. Rule of law scored -1.2 (2024), which is the measure that most often decides whether a diaspora investor commits capital rather than only sending support.
Better if
Property registration, repatriation of profits and dispute resolution are made predictable — the three things that convert diaspora sentiment into diaspora capital.
Worse if
Currency controls or an unpredictable tax treatment of returning residents make commitment more expensive than continued remittance.
Who this misses
Diaspora members without capital to invest, whose contribution is already the largest single external flow and who are rarely the audience for investment policy.
Assumptions and evidence
GDP per capita paths are in constant present-day dollars and assume no exchange-rate or price forecast.
Remittance ratios are balance-of-payments estimates and exclude informal channels, which are substantial in most of these countries.
Personal remittances received 0% · 2025Foreign direct investment, net inflows 0.9% · 2025GDP per capita $3,129 · 2025Rule of law -1.17 · 2024
Disruption · 2036
What the projected numbers would mean, taken together, for ten representative situations under
Disruption in 2036.
These are not predictions and not stories about real people. Each one states the evidence it rests
on, the assumptions it makes, and who the improvement would miss.
A young person entering the workforce
Roughly 8.4 million more people reach working age between 2026 and 2036. They enter a labour market where 27.2% of 15–24s were already counted as unemployed in 2025, and where 52% of all work is in agriculture (2025). Under Disruption, internet use moves from 41% to 52% — a modest widening, meaning connectivity remains a sorting mechanism rather than a leveller. With tertiary enrolment at 10%, most will not pass through formal higher education, so what they can do at 25 is set by what employers and informal apprenticeship actually teach — not by universities.
Better if
Technical training is aligned to what employers actually hire for, and the evidence on which courses lead to work is published.
Worse if
The cohort arrives faster than jobs, training places or electricity, in which case a demographic dividend becomes a grievance.
Who this misses
Young people in farming households, who typically leave school earliest and are furthest from the connectivity and training that the projection describes.
Assumptions and evidence
Working-age population follows the UN World Population Prospects projection, unchanged.
Youth unemployment is an ILO modelled estimate and understates underemployment in largely informal labour markets.
Connectivity gains are assumed to reach young people at least as fast as the population average — historically optimistic.
Population aged 0–14 44.1% · 2025Youth unemployment, 15–24 (ILO modelled) 27.2% · 2025School enrolment, tertiary (gross) 10% · 2024Employment in agriculture 51.9% · 2025Individuals using the internet 40.7% · 2024
A small-business owner
Electricity access moves from 56% of the population (2024) to 68% by 2036 under Disruption. That measures connections, not hours of supply: a workshop counted as served may still be waiting on power for part of the day, and this platform holds no outage or tariff data for it. Domestic credit to the private sector stood at 6.6% of GDP in 2025, which is thin by any comparison. Plausible implication, not measured: a credit-to-GDP ratio this low is consistent with borrowing being hard for small firms, but the ratio is an aggregate and cannot show which firms were refused or why. A firm-level credit survey would establish it. 29% of people aged 15 and over held an account at a bank or other financial institution, or personally used a mobile-money service, in 2014. Mobile money is counted. Plausible implication, not measured: account records are the kind of evidence a lender can price risk against, so widening ownership could open borrowing to traders without land title. Nothing in this platform measures whether that has happened here — lender practice, or credit-bureau coverage, would show it. Average output per person in the surrounding market moves from $3,129 to $1,966 a year. That is a national mean and does not identify whose income changed.
Better if
Collateral registries, insolvency procedure and payment rails are fixed before credit is subsidised.
Worse if
Credit is expanded into a weak enforcement system, producing bad loans rather than businesses.
Who this misses
Unregistered traders, who are the majority of businesses in most of these economies and are invisible to both the credit statistics and the policies built on them.
Assumptions and evidence
GDP per capita is expressed in constant present-day dollars and describes average output, not the income of any particular household.
Electricity access measures connection, not hours of supply or price. A connected business with eight-hour outages is counted as served.
Domestic credit to the private sector is a stock ratio. It does not distinguish lending to large firms from lending to small ones, and most of the businesses this situation describes are unregistered and outside it entirely.
Access to electricity 55.5% · 2024Domestic credit to private sector 6.6% · 2025Account at a bank or mobile-money provider, age 15+ 29.3% · 2014GDP per capita $3,129 · 2025
A farmer or food producer
52% of the country's workers were in agriculture in 2025. Cereal yields averaged 1,062 kg per hectare — well below what the same seed achieves with reliable inputs, water and storage, so the largest available productivity gain is not a new technology but the ordinary one already used elsewhere. By 2036, 68% of the population has electricity under Disruption — the precondition for irrigation pumps, cold storage and anything that stops a harvest rotting between field and market. 52% internet use puts price information, weather and pest identification within reach of a phone, for those who have one and can afford data. Undernourishment affected 22.5% of people in 2023: food exists in the region and cannot affordably be moved, which is a roads and storage problem more than a farming one.
Better if
Investment goes into what happens after harvest — storage, roads, cold chain, market information — not only into what happens before it.
Worse if
A multi-season drought arrives, which no band on this platform models, or input costs rise faster than farmgate prices.
Who this misses
Rain-fed smallholders without irrigation, who feel a bad season first and hardest, and women farmers, who in most of these countries hold weaker land tenure and less access to credit.
Assumptions and evidence
National electricity and connectivity projections are applied to a rural situation, though rural access consistently lags the national figure.
Yield gaps are treated as closeable with existing technology; that assumes input supply, credit and extension services that may not exist.
Employment in agriculture 51.9% · 2025Cereal yield 1,062 kg/ha · 2023Prevalence of undernourishment 22.5% · 2023Freshwater withdrawal as share of available resources 1.9% · 2022
A healthcare worker, and a patient
There were 2.4 physicians per 10,000 people in 2022. The population they serve grows by 13 million between 2026 and 2036, so holding the current ratio requires training and retaining staff at that rate merely to stand still. Electricity reaching 68% by 2036 under Disruption is what makes a vaccine cold chain, a functioning theatre and a night-time delivery possible outside the largest towns. Out-of-pocket payments were 33% of health spending in 2023. Under-five mortality stood at 49 per 1,000 live births (2024).
Better if
Health financing moves toward pooled, prepaid arrangements, and task-shifting is supported by law rather than tolerated informally.
Worse if
Trained clinicians emigrate faster than they are replaced, or external health financing contracts.
Who this misses
People in rural districts and informal settlements, where the physician ratio is a fraction of the national figure, and anyone whose care is rationed by ability to pay at the point of service.
Assumptions and evidence
Workforce density is a national average; it conceals extreme concentration in capital cities.
This platform holds no projection for health workforce or financing — only the current position against a projected population.
Physicians per 1,000 people 0.24 / 1,000 · 2022Out-of-pocket health expenditure 32.5% · 2023Under-5 mortality rate 49 / 1,000 · 2024
A teacher, and a student
The school-age population grows with a national population rising by 13 million to 2036. Gross secondary enrolment was 51.5% in 2023 and tertiary 10%. Government spent 2.5% of GDP on education in 2023. Under Disruption, 68% of households have electricity by 2036 — which is when homework after dark stops being a privilege. 52% internet use makes digital material reachable for that share of students, and no more; 48% of the population remains outside it. Adult literacy at 68% (2023) means many of these students cannot be helped with schoolwork at home, whatever the school does.
Better if
Electricity and connectivity reach schools ahead of the national average rather than behind it, and teacher supply grows with enrolment.
Worse if
Enrolment expands faster than teachers are trained, converting access gains into larger classes and weaker learning.
Who this misses
Girls in regions where secondary completion diverges sharply by gender, and children in households where the opportunity cost of school attendance is immediate income.
Assumptions and evidence
Gross enrolment ratios can exceed 100% and include students outside the standard age range; they measure participation, not learning.
This platform holds no measure of learning outcomes, teacher supply or classroom quality — a substantial gap in describing education.
By 2036 under Disruption, roughly 27.8 million people in this country are online — the domestic audience reachable without leaving it. That is up from about 41% of the population in 2024 to 52%. 29% of people aged 15 and over held a bank or mobile-money account in 2014, so being paid directly by that audience is mechanically possible for a minority of it. A population where 44% were under 15 in 2025 is, within this horizon, an unusually young consumer market for music, film, fashion and games. Production still depends on power: 68% household electricity access by 2036 sets how far studios, editing and streaming can spread beyond the main city.
Better if
Payment rails work across borders and royalties are collected and enforced domestically rather than captured offshore.
Worse if
Data costs stay high relative to income, or distribution platforms owned elsewhere capture the value created here.
Who this misses
Creators working in languages the major platforms do not support well, and anyone without a device capable of running current tools.
Assumptions and evidence
Domestic reach is calculated as projected internet use multiplied by projected population. It counts people who can technically be reached, not an audience.
This platform holds no reviewed creative-sector evidence for this country. Reach and payment access are stand-ins for a sector nobody has measured here.
Individuals using the internet 40.7% · 2024Account at a bank or mobile-money provider, age 15+ 29.3% · 2014Population aged 0–14 44.1% · 2025
A trader using regional markets
Trade was 35% of GDP in 2025, so what happens at a neighbour's border arrives as a domestic price within months. This country sits inside SADC, ECCAS. AfCFTA has been signed by every African Union member state except Eritrea, but tariff schedules, rules of origin and customs practice — not the treaty — decide whether a small consignment moves more cheaply in 2036 than it does today. 29% of people aged 15 and over held a bank or mobile-money account in 2014 — the share able to settle across a border without carrying cash.
Better if
Rules of origin are agreed on contested product lines and applied at national customs, not just gazetted.
Worse if
Regional blocs fragment further, or preferential tariffs remain unimplemented at the border despite ratification.
Who this misses
Small cross-border traders, disproportionately women, who operate below the threshold at which formal preferential regimes are usable at all.
Assumptions and evidence
Trade openness is a national accounts ratio and says nothing about the cost or time of a specific border crossing.
This platform holds no country-level data on intra-African trade share, non-tariff barriers or border clearance times — the measures that would matter most here.
Trade (exports + imports) 34.6% · 2025Account at a bank or mobile-money provider, age 15+ 29.3% · 2014
A family in a growing city
Urban population was growing 4.1% a year in 2025, doubling roughly every 17 years. On the projected population, of the order of 9.3 million additional people are living in this country's towns and cities by 2036. 63% of urban residents were in informal settlements in 2022 — housing built faster than it was serviced. Under Disruption, 68% of the population has electricity by 2036. The decisions that determine whether this family's commute is one hour or three — where roads, water mains and drainage go, and which land is reserved before it is built on — are being made now, and are expensive to reverse afterwards.
Better if
Land for roads, water mains and drainage is bought and reserved before settlement arrives — the cheapest infrastructure decision available and the one most often skipped.
Worse if
Growth outpaces servicing, in which case today's new neighbourhoods become the next decade's retrofit problem at many times the cost.
Who this misses
Renters and residents without secure tenure, who bear eviction risk and are usually the first displaced when upgrading finally arrives.
Assumptions and evidence
Urban share is held constant when estimating additional urban residents; in practice it is rising, so this is a conservative figure.
This platform holds no city-level data. National averages conceal very large differences between a capital and a secondary town.
Urban population 71.5% · 2025Urban population growth 4.1% · 2025Population living in slums 62.7% · 2022
A citizen dealing with government
Government effectiveness scored -0.8 in 2024 on a scale running roughly −2.5 to +2.5. With 52% internet use by 2036 under Disruption, permits, payments, registrations and records can plausibly be handled without a physical queue for that share of the population — the most tangible improvement most citizens would actually notice. Voice and accountability scored -1.1 (2024). The same identity systems, registries and monitoring capability that shorten a queue are also what surveillance is built from, and the measures that would constrain misuse currently score in the lower band. Both outcomes are live and depend on choices not yet made.
Better if
Digital public infrastructure is built with data protection law, independent oversight and offline fallbacks from the start.
Worse if
Identity and payment systems become a precondition for services before coverage is universal, excluding people from entitlements they already hold.
Who this misses
People without formal identity documents, who are disproportionately rural, poor, migrant or stateless — and who are excluded twice when services move online.
Assumptions and evidence
Governance indicators are perception-based composites. They signal direction and relative position, not measured administrative performance.
Connectivity is assumed to be usable for government services; in practice affordability, literacy and identity documentation each gate access again.
Government effectiveness -0.79 · 2024Voice and accountability -1.06 · 2024Government revenue excl. grants 18.2% · 2024
A member of the diaspora weighing return or investment
Remittances were 0% of GDP in 2025, against foreign direct investment at 0.9% — meaning money sent home by citizens abroad is a significant and unusually reliable external flow. Under Disruption, real output per person moves from $3,129 (2025) to $1,966 by 2036, about -4.1% a year. By 2036, 68% electricity access and 52% internet use set what can actually be operated on the ground. Rule of law scored -1.2 (2024), which is the measure that most often decides whether a diaspora investor commits capital rather than only sending support.
Better if
Property registration, repatriation of profits and dispute resolution are made predictable — the three things that convert diaspora sentiment into diaspora capital.
Worse if
Currency controls or an unpredictable tax treatment of returning residents make commitment more expensive than continued remittance.
Who this misses
Diaspora members without capital to invest, whose contribution is already the largest single external flow and who are rarely the audience for investment policy.
Assumptions and evidence
GDP per capita paths are in constant present-day dollars and assume no exchange-rate or price forecast.
Remittance ratios are balance-of-payments estimates and exclude informal channels, which are substantial in most of these countries.
Personal remittances received 0% · 2025Foreign direct investment, net inflows 0.9% · 2025GDP per capita $3,129 · 2025Rule of law -1.17 · 2024
Structural forces the growth model cannot see
The quantitative band on this page extrapolates fifteen years of this country's own growth. It is
defensible because it is narrow — and what it cannot see is anything those fifteen years did not
contain. These are those things, assessed from evidence and deliberately not converted into numbers.
Commodity dependence
Binding constraint
Natural resource rents were 30% of GDP in 2021. At this concentration the budget — and therefore public salaries and services — tracks prices set on other continents. The scenario question is not whether the resources exist but whether any value is captured, processed and taxed before they leave.
Why the growth model cannot see this
The growth band averages across a commodity cycle, which flatters producers in a downswing and understates their exposure in an upswing. Price is set elsewhere and transmits to the budget within a quarter.
Total natural resources rents 30% · 2021Exports of goods and services 23% · 2025
Population pressure & the youth cohort
Binding constraint
Population is projected to rise 32.3% by 2036, from 40.2 million to 53.2 million — 13 million additional people, on a base where 44% were already under 15 in 2025. Holding living standards constant requires building at that rate merely to stand still. This is the most fixed variable on the page and the least optional.
Why the growth model cannot see this
Per-capita growth already nets out population, which hides the absolute scale of what must be built. A constant GDP per capita with a doubling population means twice the schools for the same living standard.
Population aged 0–14 44.1% · 2025
Education & workforce capability
Binding constraint
Gross tertiary enrolment was 10% in 2024, with adult literacy at 68% (2023). A state cannot regulate, procure, audit or localise technologies it has too few people trained to understand. This constrains AI governance, energy planning and public administration simultaneously, and it takes a decade to change — which is exactly this platform's horizon.
Why the growth model cannot see this
Human capital compounds on a decade lag. The tertiary enrolment rate today sets the ceiling on what the state can regulate and what firms can build in 2036, and no growth average encodes that.
Water withdrawals are 2% of available resources (2022). 52% of workers are in agriculture (2025). Exposure is significant enough that a severe season would show up in national accounts.
Why the growth model cannot see this
Climate impact is non-linear and arrives as discrete events. A decade of trailing growth contains the droughts that happened, not the ones now becoming more likely.
Freshwater withdrawal as share of available resources 1.9% · 2022Employment in agriculture 51.9% · 2025New displacement associated with disasters 79,000 · 2023Prevalence of undernourishment 22.5% · 2023
Energy availability
Material
56% of the population had electricity in 2024, and 1% in rural areas. Access is advancing but remains a real ceiling on industrial and digital ambition.
Why the growth model cannot see this
Electricity is a precondition, not an output. A growth average cannot express that a factory, clinic or data centre simply cannot be operated at all.
Access to electricity 55.5% · 2024Access to electricity, rural 1.3% · 2024
Urbanisation
Material
Urban population was growing 4.1% a year in 2025 — a doubling roughly every 17 years, with 63% of urban residents in informal settlements (2022). The land reserved (or not reserved) for roads, water and drainage in this decade determines what these cities cost to run for the next fifty years.
Why the growth model cannot see this
Where people live determines what infrastructure is worth building. The decisions being made now about land, water mains and drainage are expensive to reverse and invisible in GDP.
Urban population growth 4.1% · 2025Urban population 71.5% · 2025Population living in slums 62.7% · 2022
Digital infrastructure
Material
41% of people used the internet in 2024, and 29% of adults held an account (2014). Enough reach for real services, not enough for universal ones.
Why the growth model cannot see this
Connectivity is a platform for other sectors rather than a sector itself. Its absence caps what health, education, finance and government can each become, in ways no aggregate growth figure separates out.
Individuals using the internet 40.7% · 2024Account at a bank or mobile-money provider, age 15+ 29.3% · 2014Fixed broadband subscriptions 0.4 / 100 · 2024
Institutional capacity & regulatory reform
Material
Government effectiveness scored -0.8 in 2024, regulatory quality -0.8. Read as a signal about implementation capacity rather than a verdict — but implementation capacity is precisely what decides whether announcements become assets.
Why the growth model cannot see this
Implementation capacity is what separates an announced plan from a delivered one. It is the single largest reason infrastructure pipelines across this continent under-deliver, and it is invisible in output data.
Government effectiveness -0.79 · 2024Regulatory quality -0.77 · 2024Rule of law -1.17 · 2024
Food & water systems
Material
Undernourishment affected 22.5% of the population in 2023, with food 16% of merchandise imports (2024).
Why the growth model cannot see this
Food security is a function of income and logistics as much as of harvests. Food usually exists in the region and cannot affordably be moved — a distribution failure that output data does not describe.
Political stability scored -0.6 in 2024. Neither a binding constraint nor a settled question.
Why the growth model cannot see this
Conflict does not appear in a trailing growth average until after it has already happened. A country at peace for fifteen years has no war in its variance.
Political stability and absence of violence -0.57 · 2024Refugees originating from this country (UNHCR mandate) 13,433 · 2025
Health-system resilience
Watch
2.4 physicians per 10,000 (2022). Out-of-pocket share 33% (2023).
Why the growth model cannot see this
A health shock removes labour and imposes costs simultaneously. Systems that were already thin absorb nothing, and the growth record of a decade without a pandemic says nothing about the next one.
Physicians per 1,000 people 0.24 / 1,000 · 2022Out-of-pocket health expenditure 32.5% · 2023Under-5 mortality rate 49 / 1,000 · 2024
Regional integration
Watch
Member of 2 regional bodies; trade is 35% of GDP (2025).
Why the growth model cannot see this
Market access changes the ceiling on what is worth producing. A trailing average is drawn from the market a country had, not the one a treaty may give it — or take away.
Trade (exports + imports) 34.6% · 2025
AI, automation & compute ownership
Watch
Africa holds about 0.6% of global data centre capacity, almost all of it in five other countries. For this country the practical questions are about terms rather than facilities: where public data sits, under whose jurisdiction, and whether any value from data generated here is captured here.
Why the growth model cannot see this
The economic value of AI accrues to whoever owns the compute, the models and the data. A national growth figure records output, not who captured it or where it was booked.
Individuals using the internet 40.7% · 2024Access to electricity 55.5% · 2024
External & geopolitical influence
Watch
FDI was 0.9% of GDP, remittances 0%.
Why the growth model cannot see this
Financing terms, security partnerships and market access are set in other capitals. They change faster than a fifteen-year average can register, and they change most sharply for the countries with least leverage.
Foreign direct investment, net inflows 0.9% · 2025Personal remittances received 0% · 2025
Major infrastructure completion
Supportive
1 reviewed project has been commissioned and is in service — fixed points in any 2031 or 2036 scenario, not promises. 0 further are funded or under construction.
Why the growth model cannot see this
A dam, port or corridor that has never existed contributes nothing to a trailing mean, then changes the economics of a whole region the year it is commissioned.
1 further forces assessed as not evident or unknown
Sovereign debt & fiscal distress
Unknown
Debt is not reported; revenue alone cannot establish fiscal position.
Why the growth model cannot see this
A default is a discontinuity. Trailing variance cannot contain an event that has not yet occurred, and debt service crowds out the spending that produces future growth long before any default.
Government revenue excl. grants 18.2% · 2024
What would invalidate the modelled band
Commodity dependence. A sustained move in the country's principal export price of more than roughly a third. The band averages across a cycle and understates both tails.
Population pressure & the youth cohort. A significant revision to the UN World Population Prospects for this country. The population path here is carried through unchanged and is the least uncertain input on the page.
Education & workforce capability. A sustained collapse or expansion in enrolment, or large-scale emigration of trained professionals.
Climate exposure & disaster risk. A multi-season drought or a major flood event affecting the main producing regions. Discrete climate events are outside any band built from recent averages.
Energy availability. Either a step change in generation and transmission delivery, or a sustained supply failure. Both break a projection built on gradual access gains.
Urbanisation. A major change in land, housing or transport policy in the principal city.
Digital infrastructure. A step change in device or data affordability, or a major submarine cable landing.
Institutional capacity & regulatory reform. A change of government that materially alters delivery capacity, in either direction.
Food & water systems. A regional harvest failure, or a change in the affordability of food imports.
Research depth and transparency
Initial country reviewLast reviewed 2026-07-25
Some country-specific evidence reviewed, but narrow — fewer than three of the five target areas covered.
Reviewed items
4
Categories covered
4 of 24
Target areas covered
2 of 5
Primary or official sources
3
Secondary or tertiary sources
1
Verified items
4
Unverified items
0
Oldest source
2026
Newest source
2026
Target areas not yet covered: Macroeconomic & fiscal, Human development, Technology & AI.
Reviewed country evidence
4 items across 4 categories,
reviewed 2026-07-25. Every item carries a publisher, a date, a link and a delivery status.
An announcement is not an asset — status is the field that matters most here.
Angola's decade turns substantially on whether the Lobito Corridor becomes a trade route or remains an export channel with a railway attached.
Operational3Funded1
Transport, logistics & corridors
OperationalAlready in effectCross-border corridorprimaryVerified delivery
The 1,289 km Benguela line from the Atlantic port of Lobito to the Angola–DRC border operates under a 30-year freight concession held by Lobito Atlantic Railway.
Financing is real and freight is moving, which puts this ahead of most announced African corridors. The open questions are distributional: whether processing happens in-country, whether the line carries regional trade as well as minerals, and whether the April 2026 flooding is treated as an anomaly or as an operating cost of a changed climate.
Creative and cultural intelligence
No creative-economy evidence reviewed for this country.
That is a statement about this platform's reading, not about the country. Creative activity is
poorly captured by official statistics almost everywhere on the continent: it is largely informal,
it straddles national-accounts categories, and few African statistics agencies publish a creative
satellite account. Absence here should be read as an unmeasured sector, never as an empty one.
Readings
The platform's own interpretation of the evidence, set in serif and marked so it never reads as
measurement. Each cites the values it fired on.
DriverInterpretation
Population is projected to rise from 40.2 million in 2026 to 53.2 million by 2036 — an additional 13 million people, 32.3% more than today. This is the most fixed variable in every scenario on this page: the people who will be adults in 2036 have already been born. What is genuinely uncertain is not how many there will be, but what they will have to work with.
Population (estimates & projections) 40215179 · 2026
DriverInterpretation
44% of the population was under 15 in 2025. That cohort reaches working age inside this platform's horizon. Whether it arrives as a dividend or a grievance is decided now, by school places, electricity, clinics and whether there is anything to do at eighteen — not by anything that happens in 2036.
Population aged 0–14 44.098 · 2025
DriverInterpretation
52% of employment was in agriculture in 2025, producing 25% of GDP (2025). The distance between those two numbers is the country's productivity problem stated in one line: most people work where least value is created. It also means climate is not an environmental topic here — it is the household income of the majority, and a bad season is a wage cut for most of the workforce.
Employment in agriculture 51.949 · 2025Agriculture, forestry & fishing value added 25.241 · 2025
RiskInterpretation
Youth unemployment stood at 27.2% in 2025, against 14.1% overall on ILO modelled estimates. Measured unemployment understates the problem in economies where most work is informal — the sharper question is not whether young people are working but whether the work pays, accumulates skill, or leads anywhere.
Gross tertiary enrolment was 10% in 2024. A country cannot regulate, procure, audit or localise technologies it has too few people trained to understand — this constrains AI adoption, energy planning and public administration simultaneously, and it takes a decade to fix, which is exactly this platform's horizon.
School enrolment, tertiary (gross) 9.955 · 2024
DriverInterpretation
56% of the population had access to electricity in 2024, falling to 1% in rural areas — roughly 17.9 million people without. Electricity is the constraint that sits underneath the others: without it, there is no cold chain for clinics, no irrigation pump, no evening study, no small workshop, and certainly no data centre. Every optimistic scenario for this country runs through this number.
Access to electricity 55.5 · 2024Access to electricity, rural 1.3 · 2024
RiskInterpretation
Natural resource rents were 30% of GDP in 2021. Revenue at this concentration ties the budget — and therefore schools, clinics and salaries — to prices set on other continents. The scenario question is not whether the resources exist but whether any of the value is captured, processed and taxed domestically before it leaves.
Total natural resources rents 29.967 · 2021
DriverInterpretation
41% of people used the internet in 2024, against 70 mobile subscriptions per 100 people (2024). Owning a phone and being genuinely online are different things. The distance between them is affordable data, a device that runs modern software, power to charge it, and content in a language you read. That distance decides who any AI deployment reaches.
Individuals using the internet 40.705 · 2024Mobile cellular subscriptions 69.704 · 2024
RiskInterpretation
Research and development spending was 0% of GDP in 2016. At this level a country is overwhelmingly a consumer of technology designed elsewhere for conditions elsewhere — which shapes not only who profits, but which problems get solved at all.
Research & development expenditure 0.032 · 2016
RiskInterpretation
There were 2.4 physicians per 10,000 people in 2022. Workforce scarcity at this level is what makes task-shifting, community health workers and decision-support tools consequential rather than merely fashionable — and it is also why emigration of trained clinicians is a first-order economic issue, not a footnote.
Physicians per 1,000 people 0.244 · 2022
SignalInterpretation
Under-five mortality fell from 64.5 to 49 per 1,000 live births between 2014 and 2024. Sustained declines like this are among the most reliable evidence that basic public systems are functioning, because they cannot be produced by a single project or a good year.
Under-5 mortality rate 49 · 2024
RiskInterpretation
Undernourishment affected 22.5% of the population in 2023 — on current population, in the order of 9 million people. Food security here is a function of income and logistics at least as much as of harvests: food usually exists somewhere in the region, and cannot affordably be moved to where it is needed.
Prevalence of undernourishment 22.5 · 2023
OpportunityInterpretation
Cereal yields averaged 1,062 kg per hectare in 2023, well below what the same seed achieves under reliable inputs and water. The yield gap is the largest single productivity opportunity in the economy — and closing it is an input, credit, storage and road problem more than a technology problem.
Cereal yield 1061.8 · 2023
DriverInterpretation
Urban population was growing 4.1% a year in 2025, with 71% of people already living in towns and cities. At that pace urban population doubles in roughly 17 years. 63% of urban residents were living in informal settlements (2022). The decisions that matter — where roads, water mains and land titles go — are being made now, and are expensive to reverse once built.
Urban population growth 4.136 · 2025Urban population 71.462 · 2025Population living in slums 62.7 · 2022
RiskInterpretation
Domestic credit to the private sector was 6.6% of GDP in 2025. That is a stock ratio for the whole economy: it does not separate lending to large firms from lending to small ones, and it cannot show which firms sought credit or were refused. What it does establish is how little total credit the private sector holds relative to output, which bounds how much of any expansion can be debt-financed.
Domestic credit to private sector 6.581 · 2025
DriverInterpretation
Government effectiveness scored -0.8 in 2024 on the Worldwide Governance Indicators scale (roughly −2.5 to +2.5), well below the global mid-point, with regulatory quality at -0.8. These are perception-based composite measures and should be read as a signal about implementation capacity rather than as a verdict — but implementation capacity is precisely what separates an announced plan from a delivered one.
Government effectiveness -0.785 · 2024Regulatory quality -0.773 · 2024
Who this lands on
Likely to gain
Young people entering a labour market that has been prepared for them
Smallholder farmers, if input costs, storage and roads improve alongside prices
Households and small businesses reached by new connections
The public finances, when prices are high
People connecting for the first time, who move from no access to some — a larger step than any later improvement in speed or price
Construction, transport and services workers in growing cities
At risk of being left behind
A cohort that reaches working age faster than jobs, training places or clinics are built
Rain-fed farming households, first and hardest, in a bad season
Educated young people whose qualifications do not convert into work
A state that must regulate technologies it has too few trained people to assess
Off-grid communities, whose wait lengthens if capital concentrates on urban and industrial load
Everyone dependent on public salaries and services when prices fall
Rural, older, poorer and less literate people, who are last in every connectivity sequence
Households already spending most of their income on food
New urban arrivals without tenure security, water or transport
Firms whose expansion depends on credit, in an economy holding little of it relative to output
Decisions available now
Align technical training with what employers actually hire for, and make the evidence public, so young people can see which courses lead to work.
Put money into what happens after harvest — storage, roads, cold chain, market information. Yield gains that rot on the way to market do not raise incomes.
Build state capacity to evaluate and procure technology, not only to use it. A regulator that cannot read a model card cannot govern a deployment.
Treat connection targets and generation headroom as one plan, not two — and publish the delivery data monthly, so slippage is visible while it can still be corrected.
Bank the windfalls in a rules-based fund and publish the balance. The commodity cycle is the one thing about which every producer has perfect foresight and imperfect discipline.
Attack device and data cost directly, and require local-language performance in public digital procurement.
Fund a small number of research groups properly and for a decade, rather than many for three years. Capability compounds; pilots do not.
Buy and reserve the land for roads, water mains and drainage before the settlement arrives. It is the cheapest infrastructure decision available and the one most often skipped.
Fix collateral registries, insolvency procedure and payment rails before subsidising credit. Cheap credit into a broken enforcement system creates bad loans, not businesses.
Artificial intelligence — one layer, not the whole story
FoundationsInterpretation
With 41% internet use (2024) and 56% electricity access (2024), AI-mediated services can reach a substantial minority — largely urban, largely younger, largely already connected. Deployments built on these foundations will tend to widen the gap with everyone else unless they are explicitly designed not to.
Individuals using the internet 40.705 · 2024Access to electricity 55.5 · 2024
Work and productivityInterpretation
With 52% of employment in agriculture (2025), the dominant AI question here is not white-collar displacement — the exposed occupations barely exist at scale. It is whether AI reaches farming and informal trade at all: pest and disease identification, weather and price information, and credit assessment for people with no formal credit history. The risk is exclusion, not redundancy.
Employment in agriculture 51.949 · 2025Employment in services 40.851 · 2025
Compute and ownershipInterpretation
Africa holds about 0.6% of global data centre capacity, and almost all of it sits in South Africa, Egypt, Kenya, Morocco and Nigeria. For this country, the practical questions in this decade are therefore about terms rather than facilities: where public data is hosted and under whose jurisdiction, what a government pays for inference, whether procurement requires local language performance, and whether any value from data generated here is captured here.
Language and inclusionInterpretation
Adult literacy was 68% in 2023. Systems that only work in a former colonial language, in text, for confident readers, will reach the people who already had access to services. Voice interfaces and genuinely capable local-language models are the difference between AI as a broadening technology and AI as another sorting mechanism — and that capability depends on datasets that mostly do not exist yet, which is itself an opportunity for whoever builds them.
Adult literacy rate 68.18 · 2023
Government and civil libertiesInterpretation
Government effectiveness scored -0.8 (2024) and voice and accountability -1.1 on the Worldwide Governance Indicators. AI in public administration can compress waiting times for permits, payments and records — the most tangible improvement most citizens would notice. The same infrastructure — identity systems, biometric registries, communications monitoring — is also what surveillance is built from, and the accountability measures that would constrain its misuse currently score in the lower band. Both possibilities are live and depend on choices not yet made.
Government effectiveness -0.785 · 2024Voice and accountability -1.059 · 2024
Information integrityInterpretation
With 41% of people online (2024), synthetic audio and video are cheap enough to matter in elections, communal disputes and markets. Detection is not a solved problem anywhere, and it is hardest for exactly the languages and dialects with the least training data — which describes most of the continent's information environment.
Individuals using the internet 40.705 · 2024
The twenty-dimension profile
20 dimensions. 6 rest on a document or a
measurement, 5 are inferred from adjacent
indicators, 6 were searched without
result, and 3 have not been examined at all.
That last number is a statement about this platform, not about Angola — and keeping the
two apart is why there is no score here.
Can this country power a data centre without taking supply from households?
Household electricity access is 55.5% (2024), rural access 1.3%. No document on generation capacity for compute has been reviewed, so this is inferred from household access alone — which is a poor proxy for industrial supply.
Why it matters. Compute is a load before it is an industry. A country adding megawatts of IT load onto a grid that already fails is choosing between two customers, and the choice is political, not technical.
What would change this. A grid operator statement, a power-purchase agreement for a named facility, or a published reserve-margin figure.
MeasuredConnectivity foundation
How much of the population can reach a network at all?
40.7% of the population used the internet in 2024, with 70 mobile subscriptions per 100 people and 0.4 fixed broadband subscriptions per 100.
Why it matters. Every AI application that reaches a person reaches them over a connection. The connectivity number sets the ceiling on adoption regardless of what is built.
What would change this. Nothing — this dimension is measured. It ages rather than changes state.
Not evidentData-centre capacity
Is there compute inside this country, or only compute it rents elsewhere?
No data-centre document found for this country. Africa holds well under 1% of global capacity and it is concentrated in a handful of markets, so absence here is the expected case rather than a surprising one.
Why it matters. Where the compute sits determines who sets the terms, where the jobs are, and whose law applies to the data. It is the difference between a market and a customer.
What would change this. An operator announcement with a named site, a commissioning notice, or an installed-IT-load figure.
Not evidentInternational bandwidth
How does traffic leave the country, and how many ways are there out?
No verified cable landing recorded for this coastal country in the systems layer.
Why it matters. A landlocked country, or a coastal one served by a single cable, pays more for bandwidth and loses it entirely when the cable breaks. Route diversity is a resilience property, not a luxury.
What would change this. A cable consortium landing announcement, or an IXP traffic figure showing how much traffic stays domestic.
Governance
Not evidentNational AI policy or strategy
Has the state written down what it intends to do about AI?
No national AI policy or strategy document found for this country. Roughly a third of African states have adopted one, so absence is common and not disqualifying.
Why it matters. A strategy is not capability. But its absence means every decision about compute, data and procurement is being taken case by case, usually by whoever arrives with a proposal.
What would change this. A published national strategy, a cabinet approval, or an AU-level commitment with a national instrument behind it.
UnexaminedData protection and privacy law
Is there a law governing what may be done with data about people here?
This platform has not yet reviewed data-protection legislation for this country. The African Union Malabo Convention and a substantial number of national laws exist; none has been read and cited here. This is a research gap, not a finding.
Why it matters. Training data comes from somewhere. Without an enforceable data-protection regime, the terms on which a country's population becomes training data are set entirely by whoever is collecting.
What would change this. The national data protection act, the establishment of a supervisory authority, or a first enforcement action.
Not evidentGovernment AI adoption
Is the state itself using these systems, and on whom?
No document found on government use of AI systems in this country. Deployments in eligibility, policing and border control are rarely announced, so absence here is weak evidence.
Why it matters. Government is usually the largest single deployer in a low-income economy, and the deployments that matter most — benefits eligibility, policing, border control — are the ones least likely to be evaluated in public.
What would change this. A procurement notice, an audit, a parliamentary answer, or a civil-society investigation.
InferredDigital public infrastructure
Is there national identity, payment and data-exchange infrastructure for AI systems to attach to?
No DPI document reviewed. Account ownership of 29.3% of adults (2014) is used as a partial stand-in for payment rails, and stands in for nothing else.
Why it matters. DPI decides whether an AI service can identify a person, be paid, and read a record. It is the rail that determines whether an application reaches anyone.
What would change this. An identity system enrolment figure, an instant-payment system launch, or a data-exchange layer going live.
Capability
MeasuredResearch capacity
Does the country fund people to work on this?
Research and development spending was 0.03% of GDP in 2016. For reference, sustained capability-building elsewhere has generally required above 1%.
Why it matters. Capability compounds and cannot be imported at short notice. R&D spending in 2026 sets what is possible domestically in 2036, not in 2028.
What would change this. A national R&D survey, or a named research institute with published funding.
InferredSkills and tertiary education
How many people are being educated to the level this work requires?
Tertiary gross enrolment was 10.0% in 2024. No education-sector document has been reviewed, and enrolment says nothing about field of study — the technical share is not captured here.
Why it matters. Tertiary enrolment today is the technical workforce of 2036. It is one of the few variables on this platform whose effect on the far horizon is close to arithmetic.
What would change this. Graduate numbers by field, or a national skills programme with enrolment figures.
DocumentedStartup and venture ecosystem
Is there private capital forming around this domestically?
1 document reviewed on the startup or investment ecosystem.
Why it matters. Where the capital comes from decides where the equity ends up. An ecosystem funded entirely offshore builds capability locally and captures returns elsewhere.
What would change this. A funding total from a named tracker, a domestic venture fund launch, or a public co-investment vehicle.
UnexaminedLocal-language technology
Do these systems work in the languages people here actually speak?
This platform has not reviewed local-language technology evidence for this country. A language appearing on a vendor's supported-languages list is marketing, not a measurement, and is not accepted here as evidence of capability. Benchmarked performance is what would count.
Why it matters. A model that performs well in English and badly in the working language of a market excludes most of that market from anything built on it — and does so invisibly, because the failure looks like a user problem.
What would change this. A published benchmark on a named language, a corpus release, or an evaluation by a local research group.
MeasuredData about the country itself
Does the data needed to build anything useful here exist?
This country has values for 96% of the indicator series carried on this platform. Thin coverage is concentrated in the countries with the weakest statistical systems, so it compounds: the places hardest to model are the places least modelled.
Why it matters. Models are built on records. A country whose own statistical system is thin cannot be modelled well by anyone, including by itself, and thin coverage is not randomly distributed.
What would change this. A census, a household survey release, or a national statistics office publishing an open data portal.
Consequence
MeasuredWork and exposure
What kind of work does the economy actually consist of?
51.9% of employment was in agriculture (2025); 40.9% in services (2025). Youth unemployment was 27.2%. Exposure to automation follows this composition, and no country-specific study of it has been reviewed here.
Why it matters. The automation debate imported from high-income countries assumes an office-heavy workforce. Where most employment is agricultural or informal, the near-term exposure is different in kind, not just in degree.
What would change this. A national labour-force survey with occupational detail, or a country-specific exposure study.
Not evidentApplication: agriculture
Is anything being applied to the sector most people work in?
No agricultural-technology document found, in an economy where 51.9% of employment is agricultural (2025). The gap between where the workforce is and where the documented activity is, is the finding.
Why it matters. Advisory, weather and pest-detection systems are the applications with the widest possible reach in most of these economies, and the ones least represented in the funding.
What would change this. An extension programme with enrolment figures, or an evaluation of an advisory service.
InferredApplication: health systems
Is any of this reaching clinical care or health financing?
0.24 physicians per 1,000 people (2022). No health-system document reviewed; clinician scarcity is recorded as context, not as evidence of any deployment.
Why it matters. Diagnostic support matters most where clinicians are scarcest, which is exactly where the infrastructure to run it is thinnest. The two constraints are inversely correlated.
What would change this. A ministry deployment, a regulatory approval for a diagnostic device, or a published clinical evaluation.
Not evidentApplication: creative production
What do generative tools do to the sectors that make things people watch, wear and hear?
No creative-economy document found for this country, so no basis exists for saying anything about AI's effect on its creative sector.
Why it matters. Screen and music production are among the most directly exposed activities anywhere, and in several of these economies they are among the fastest-growing measured sectors. Nobody has published what happens when the two meet.
What would change this. A sector survey, a union or guild position with membership data, or a rights-body statement on training data.
InferredInformation integrity
How exposed is public life here to synthetic media?
Voice and accountability at -1.06 (2024). No document on information integrity or platform governance has been reviewed for this country; this is inference from a governance indicator, which is a weak basis.
Why it matters. Synthetic audio and video are cheapest to produce and hardest to verify exactly where fact-checking capacity is thinnest and elections are closest.
What would change this. An electoral commission position on synthetic media, a platform transparency report covering the country, or a documented incident.
UnexaminedCompute ownership and value capture
If compute is built here, who owns it and where do the returns go?
No compute document, so no ownership question arises yet.
Why it matters. A data centre built, owned and operated by a foreign firm, serving foreign customers, on subsidised local power, is an export of electricity dressed as an investment. Whether that is what is happening is an ownership question, not a capacity question.
What would change this. A shareholding disclosure, a tariff or tax-incentive instrument, or a power-purchase agreement made public.
DocumentedRegional and continental position
Does anything about this country's AI position depend on its neighbours?
Member of 2 regional blocs recorded in the continental systems layer, with 1 trade or AfCFTA document reviewed. Digital-trade provisions specifically have not been reviewed for this country.
Why it matters. Bandwidth transits borders, power pools cross them, and AfCFTA governs whether a service sold in one market can be sold in the next. None of this is decided nationally.
What would change this. The AfCFTA Digital Trade Protocol's ratification status for this country, or a regional data-flow agreement.
Current position
Every measure the platform holds for Angola, with the year of observation. Gaps are shown as gaps.
Economy & growth
Measure
Value
Year
GDP per capita
$3,129
2025
GDP
$122.2bn
2025
GDP growth, annual
3.1%
2025
Inflation, consumer prices
20.2%
2025
Manufacturing value added
6.8%
2025
Government revenue excl. grants
18.2%
2024
Central government debt
not available
Population & demography
Measure
Value
Year
Population
39m
2025
Population growth, annual
3%
2025
Population aged 0–14
44.1%
2025
Population aged 15–64
53%
2025
Fertility rate, total
5.05
2024
Life expectancy at birth
64.8 yrs
2024
Cities & urbanisation
Measure
Value
Year
Urban population
71.5%
2025
Urban population growth
4.1%
2025
Urban population, total
27.9m
2025
Population living in slums
62.7%
2022
Work, skills & youth
Measure
Value
Year
Unemployment, total (ILO modelled)
14.1%
2025
Youth unemployment, 15–24 (ILO modelled)
27.2%
2025
Employment in agriculture
51.9%
2025
Employment in services
40.9%
2025
Female labour force participation rate, 15+
72.2%
2025
School enrolment, secondary (gross)
51.5%
2023
School enrolment, tertiary (gross)
10%
2024
Adult literacy rate
68.2%
2023
Government expenditure on education
2.5%
2023
Health & public services
Measure
Value
Year
Current health expenditure
2.5%
2023
Out-of-pocket health expenditure
32.5%
2023
Under-5 mortality rate
49 / 1,000
2024
Maternal mortality ratio
183 / 100k
2023
Physicians per 1,000 people
0.24 / 1,000
2022
Safely managed drinking water
not available
Agriculture & food systems
Measure
Value
Year
Agriculture, forestry & fishing value added
25.2%
2025
Cereal yield
1,062 kg/ha
2023
Prevalence of undernourishment
22.5%
2023
Food imports
15.7%
2024
Arable land
4.3%
2023
Energy & resources
Measure
Value
Year
Access to electricity
55.5%
2024
Access to electricity, rural
1.3%
2024
Access to clean cooking fuels
50.2%
2023
Renewable energy consumption
52.9%
2021
Total natural resources rents
30%
2021
Digital & AI foundations
Measure
Value
Year
Individuals using the internet
40.7%
2024
Mobile cellular subscriptions
69.7 / 100
2024
Fixed broadband subscriptions
0.4 / 100
2024
Account at a bank or mobile-money provider, age 15+
29.3%
2014
Research & development expenditure
0%
2016
High-technology exports
16.7%
2024
Trade, investment & enterprise
Measure
Value
Year
Trade (exports + imports)
34.6%
2025
Exports of goods and services
23%
2025
Foreign direct investment, net inflows
0.9%
2025
Personal remittances received
0%
2025
Domestic credit to private sector
6.6%
2025
Governance & institutions
Measure
Value
Year
Government effectiveness
-0.79
2024
Regulatory quality
-0.77
2024
Rule of law
-1.17
2024
Control of corruption
-0.62
2024
Voice and accountability
-1.06
2024
Political stability and absence of violence
-0.57
2024
Climate & environment
Measure
Value
Year
CO₂ emissions per capita
0.74 t
2024
Forest area
52.1%
2023
Freshwater withdrawal as share of available resources
1.9%
2022
Agricultural land
36.9%
2023
Security, migration & displacement
Measure
Value
Year
Net migration
-3,524
2025
Refugees hosted (UNHCR mandate, by country of asylum)
25,396
2025
Refugees originating from this country (UNHCR mandate)
13,433
2025
Military expenditure
1%
2024
Internally displaced people
not available
What we do not know
3 of 74 tracked measures have no value for Angola. They are never estimated.
Central government debtSafely managed drinking waterInternally displaced people
Sources
Every value above comes from one of these series, republished by the World Bank Indicators API from the primary compilations named.