South Africa

Scenarios and horizons

What could materially change by 2031 and 2036?

Scenarios are arithmetic on observed history under a published assumption. They are not predictions and carry no probability. 2026, 2031, 2036 are the only formal horizons — no value is projected for any year between them.

Youth labour-market entry

How many young people may enter working age, and what does entering the labour market look like?

• Approved

  South Africa’s 2026 Budget Review projects gross government debt stabilising at 78.9% of GDP in 2025/26 and declining to 76.5% over the medium term, with the consolidated deficit narrowing from 4.5% of GDP in 2025/26 to 3.1% by 2028/29 and a main-budget primary surplus rising from 1.6% to 2.3% of GDP.

  Budget Review 2026 — Chapter 7: Government Debt and Contingent Liabilities National Treasury, Republic of South Africa, 2026-02 · Primary

What kinds of work dominate

Where do people actually work, and how might that change?

Female participation rate — Labor force participation rate, female (% of female population ages 15+) (modeled ILO estimate). Percentage of the FEMALE population aged 15 and over who are in the labour force, employed or unemployed.

No reviewed document in this platform's research record bears on this question for this country. That is a gap in the evidence, not a finding about the country.

Infrastructure constraints

What are the physical limits — connection, power, and where they reach?

• Operational

  South Africa holds the largest share of African data centre capacity, within a continent that holds about 0.6% of the global total.

  Africa holds just 0.6% of global data centre capacity as $60bn AI drive spurs 1.2GW expansion by 2030 TechEconomy, 2026 · Secondary

• Approved

  Cape Town approved two hyperscale data centres drawing around 174 MW combined once fully operational.

  Are Africa's power grids ready for the AI data centre boom? World Economic Forum, 2026 · Secondary

Education and practical AI skills

What is the education base, and what is being built on top of it?

• Under construction Continental scope

  A partnership between Nvidia and Cassava Technologies valued at roughly US$700 million is deploying GPUs in South Africa.

  Africa's AI sovereignty problem is really an energy problem TechCabal, 2026-05 · Secondary

• Announced Continental scope

  The Nvidia–Cassava GPU programme has announced planned expansion beyond South Africa to Egypt, Nigeria, Kenya and Morocco.

  Africa's AI sovereignty problem is really an energy problem TechCabal, 2026-05 · Secondary

• Proposed

  South Africa has been working on a draft National Artificial Intelligence Policy rather than an adopted strategy.

  Strengthening AI governance in Africa OECD, 2026-04 · Primary

Informal enterprise and MSMEs

How may informal enterprises and small firms be affected, and what would let one grow?

Account ownership (incl. mobile money) — Account ownership at a financial institution or with a mobile-money-service provider (% of population ages 15+). Percentage of the population aged 15 and over. Counts an account at a bank or other financial institution, OR personal use of a mobile-money service in the past 12 months. Mobile money IS included.

No reviewed document in this platform's research record bears on this question for this country. That is a gap in the evidence, not a finding about the country.

Financial inclusion

Who can be paid, save or borrow — and on what definition?

Account ownership (incl. mobile money) — Account ownership at a financial institution or with a mobile-money-service provider (% of population ages 15+). Percentage of the population aged 15 and over. Counts an account at a bank or other financial institution, OR personal use of a mobile-money service in the past 12 months. Mobile money IS included.

No reviewed document in this platform's research record bears on this question for this country. That is a gap in the evidence, not a finding about the country.

Digital public services

What can a citizen do with the state without queueing?

Account ownership (incl. mobile money) — Account ownership at a financial institution or with a mobile-money-service provider (% of population ages 15+). Percentage of the population aged 15 and over. Counts an account at a bank or other financial institution, OR personal use of a mobile-money service in the past 12 months. Mobile money IS included.

• Suspended

  Implementation of South Africa’s National Health Insurance Act was suspended in February 2026 pending Constitutional Court proceedings, with the first challenge — brought by the Board of Healthcare Funders and the Western Cape government on public-participation grounds — heard from 5 to 7 May 2026 and judgment reserved.

  South Africa’s National Health Insurance faces constitutional challenge IOL / Business Report, 2026-05 · Secondary

  Corroborated: South Africa: Cyril Ramaphosa postpones implementation of National Health Insurance law Financial Afrik, 2026-03 · Secondary — Independently dates the presidential decision to postpone proclamation to a statement of 24 February 2026, and attributes it to consultation between the President and the Health Minister — establishing that the suspension is an executive act, not merely a consequence of the litigation.

Institutional implementation capability

Can the state actually deliver what it commits to — fiscally, administratively, and through a political cycle?

• Announced

  Local government elections were expected during 2026.

  Africa's 2026 Elections: Navigating Complexity to Deliver for Citizens Africa Center for Strategic Studies, 2026 · Secondary

• Approved

  South Africa’s 2026 Budget Review projects gross government debt stabilising at 78.9% of GDP in 2025/26 and declining to 76.5% over the medium term, with the consolidated deficit narrowing from 4.5% of GDP in 2025/26 to 3.1% by 2028/29 and a main-budget primary surplus rising from 1.6% to 2.3% of GDP.

  Budget Review 2026 — Chapter 7: Government Debt and Contingent Liabilities National Treasury, Republic of South Africa, 2026-02 · Primary

Announced versus delivered

How much of what has been announced has actually arrived?

Based on 11 dated records. A larger number means more evidence has been reviewed, not that more has happened.

Milestone timeline

What is dated, what state is it in, and what does it not establish?

Lived experience

What does the evidence imply for a specific person, and who does it miss?

• Observed Plausible implication

  A young person entering work

  25.7% of the population was under 15 in 2025; gross secondary enrolment was 104.1% in 2023; recorded youth unemployment was 59.9% in 2025.

  Plausible implication, not measured: A cohort this size arriving against an enrolment rate below universal means the schooling a young person brings to the labour market varies widely, and a programme pitched above that level selects a minority of them.

  What would settle it. Completion rates by sex and location, and destination data six and twelve months after leaving school. Neither is held here.

  Who this misses. Young people already out of school, who are outside an enrolment measure by construction.

  Evidence used: youth, secondary, youthunemp.

• Observed Plausible implication

  An informal trader or micro-enterprise owner

  81.1% of people aged 15 and over held an account at a financial institution or personally used mobile money in 2024; domestic credit to the private sector was 89.4% of GDP in 2024; 73.8% of employment was in services in 2025.

  Plausible implication, not measured: Being payable is not being financeable. An account is a payment rail, and nothing in this platform establishes that account records are used in any lending decision.

  What would settle it. Firm-level credit surveys recording refusals and their reasons, and credit-bureau coverage figures.

  Who this misses. Unregistered traders, who are outside most support instruments by design rather than by accident.

  Evidence used: account, credit, servemp.

• Observed Plausible implication

  A farmer

  5.4% of employment was in agriculture in 2025, producing 2.8% of GDP in 2025; rural household electricity access was 86.3% in 2024.

  Plausible implication, not measured: Where the employment share far exceeds the output share, the productivity problem and the employment problem are the same problem. Rural power is the constraint on anything requiring storage, processing or irrigation.

  What would settle it. Yield and farm-income panels linked to adoption of a specific tool or service.

  Who this misses. Farmers without a connection at all, and anyone whose land tenure prevents investment regardless of technology.

  Evidence used: agriemp, agrigdp, electricityrural.

• Observed Plausible implication

  A teacher or learner

  Gross secondary enrolment 104.1% (2023); adult literacy 91.2% (2024); household electricity access 90.2% (2024).

  Plausible implication, not measured: Evening study, device charging and any screen-based material all assume power at home. Where access is short of universal, a digital learning programme reaches the connected fraction of an already-enrolled fraction.

  What would settle it. School-level electrification and connectivity audits, and learning outcomes disaggregated by both.

  Who this misses. Non-literate adults, for whom a text-based service is unavailable regardless of coverage.

  Evidence used: secondary, literacy, electricity.

• Observed Plausible implication

  A citizen using public services

  78.4% of people used the internet in 2024; 81.1% of people aged 15 and over held an account or used mobile money in 2024; the government effectiveness estimate was -0.23 in 2024.

  Plausible implication, not measured: Digitising a service moves it towards the people who can already reach it. Retaining an assisted or offline channel is what decides whether the rest keep access.

  What would settle it. Service-level uptake by channel, and the share of transactions still completed in person.

  Who this misses. People whose identity documents do not match registry records — unmeasured everywhere in this instrument.

  Evidence used: internet, account, goveff.

• Observed Plausible implication

  A rural resident

  Rural household electricity access 86.3% against 90.2% nationally (2024); basic drinking water 68% (2024).

  Plausible implication, not measured: The 3.9-point gap between the national and rural figures is the distance between a national average and this person's actual position. An intervention designed against the average is designed for somebody else.

  What would settle it. Sub-national service data. This platform holds a national and a rural figure and nothing between them.

  Who this misses. Everyone in the unserved rural remainder, whose position no series here describes individually.

  Evidence used: electricityrural, electricity, water.

Each situation is built only where the indicators it names are held for this country. Where they are not, it says so rather than substituting general prose — a situation the evidence cannot support would read as a finding. The observed line is measured; the line marked as a plausible implication is an argument, and it names what would settle it.

Evidence quality and gaps

How much has been reviewed here, how good is it, and what is ageing?

Evidence depth differs between countries and this instrument renders the difference rather than smoothing it. A thin record is a statement about what has been reviewed. It is never a statement about the country, and it is never comparable to another country as a measure of performance.