Author: African Development Bank
Site of publication : https://www.afdb.org/en
Type of publication : Report
Date of publication : June 2025
How big is Africa’s attainable AI productivity gain by 2035?
US $1 Trillion AI Productivity Opportunity
If the continent stays on its current trajectory, defined by steady improvements in infrastructure, trade, and services, Africa’s GDP is expected to reach US $4.23 trillion by 2035. AI offers an alternative path. With the right conditions for adoption, including digital infrastructure, a skilled workforce, interoperable data, trusted systems, and sufficient capital, Africa’s economy could grow to US $5.23 trillion by 2035.
Tier-1 Sectors Deliver $580 bn: Over Half of Africa’s AI Upside
While AI is a general-purpose technology, its real-world uptake is rarely uniform. In every country, certain sectors are more prepared to benefit. This analysis identified five sectors: agriculture, retail, manufacturing, finance, and health.
- Agriculture remains the backbone of most African economies. Yet the sector operates well below its productive potential. AI is uniquely positioned to solve information asymmetry, low productivity, and market inefficiencies, which are the sector’s core challenges.
- The retail and wholesale sector is a cornerstone of Africa’s economic activity and livelihoods. Informal outlets account for over 90% of total retail transactions. AI can help formalise and upgrade these trade systems without erasing their decentralised strengths.
- Manufacturing plays a pivotal role in Africa’s industrialisation agenda, supporting export diversification, job creation, and economic resilience. AI technologies offer tangible solutions for factory floors. In addition, AI helps orchestrate factory-wide operations.
- Africa’s financial sector has undergone a quiet revolution over the past decade, driven primarily by the rapid uptake of mobile money. AI technologies allow institutions to move beyond conventional credit assessment and fraud detection tools.
- Health systems across Africa face a complex web of challenges. Artificial intelligence offers a powerful set of tools to support health workers, improve diagnostics, and streamline resource use. Imaging algorithms can prioritise X-rays and MRIs, helping radiologists identify the most urgent cases more quickly.
Implications for Strategic Planning
Converting sector-level potential into tangible outcomes will require strategic interventions tailored to both cross-cutting and sector-specific readiness gaps :
- Invest in shared data infrastructure;
- Expand blended finance for informal sectors and SMEs;
- Accelerate sector-specific talent development;
- Establish testbeds and regulatory sandboxes;
- Create public–private labs and open platforms.
Readiness Gaps : What must be fixed to deliver the dividend
The AI Readiness Flywheel
AI productivity only happens when five critical levers move together : data, compute, skills, trust and capital, momentum builds and transformative gains become achievable We call this the AI Readiness Flywheel.
Early sparks of flywheel activation are visible across the continent:
- Data: Ghana’s health datasets and Rwanda’s agriculture portals are FAIR aligned
- Compute: South Africa and Nigeria have active GPU zones connected to international. fibre
- Skills: Moringa School (Kenya) and AI Fridays (Senegal) train ML engineers
- Trust: Mauritius and Egypt have published AI guidelines; Kenya’s Data Protection Act includes fairness provisions
- Capital: Over US $500 million has been raised by African AI-adjacent startups since 2021 (e.g. InstaDeep, Viebeg, Zindi).
Continental and sub-regional platforms are pivotal to this synchronisation. The African Union’s Continental AI Strategy (2024) calls for cross-border interoperability in infrastructure and governance. These coordination bodies help countries pool resources, de-risk innovation, and unlock economies of scale ensuring that national progress is not trapped within borders, but shared continentally.
To pinpoint where Africa must invest and how quickly progress is possible, we built a data-centric, fully reproducible index that scores every African country against two baselines: 18 sovereign G20 economies. The index mirrors the report’s AI-productivity flywheel, tracking five equal-weight pillars (Data Ecosystems, Compute & Digital Infrastructure, Skills & Human Capital, Trust & Governance, Innovation & Capital). Africa composite score: 32 /100 -36 points behind the G20 (68) and 14 points behind the rest-of-world benchmark (46). Africa’s AI-readiness is far from uniform. A small group of front-runners already demonstrates what’s achievable for example, South Africa (sovereign GPU “data-embassy” hubs), Rwanda..
If only 8 or 10 countries reach AI-readiness, the dividend shrinks. Network effects are lost. Regional infrastructure remains underused. And inclusive growth the real goal becomes harder to deliver. This is why AI must be approached as a continental development priority, not a national pilot.
Continental and sub-regional platforms are pivotal to this synchronisation. The African Union’s Continental AI Strategy (2024) calls for cross-border interoperability in infrastructure and governance. These coordination bodies help countries pool resources, de-risk innovation, and unlock economies of scale ensuring that national progress is not trapped within borders, but shared continentally
In-Depth Analysis of Readiness Enablers
Africa’s AI dividend depends not just on innovation, but on infrastructure.
- Data Ecosystem : Africa scores 45/100 on data ecosystems
- Compute & Digital Infrastructure : Africa scores 22/100 on compute
- Skills & Human Capital : Africa scores 29/100 on skills, a 32-point gap with the G20 average.
- Trust & Governance : Africa scores 36/100 on trust and governance, a 42-point gap with the G20
- Innovation & Capital : Africa scores 28/100 on capital and innovation.
Country Archetypes: Four Paths to Readiness
To move from one-size-fits-all advice to country-specific roadmaps, we grouped Africa-54 into four readiness archetypes using a two-axis, data-driven segmentation: Capability Score; Market Weight. A four-branch decision tree (Annex D) then classifies countries as Catalytic Agents, Innovation Hubs, Scale Accelerators, or Foundation Builders.
Readiness Scenarios: Drift, Fragmentation and Full Activation.
Africa’s AI dividend is not guaranteed. Scenario thinking is essential when capabilities are uneven.
- Scenario 1. Status Quo Drift, GDP impact: Africa captures ~1% of the global AI dividend (~US $250B). Under this scenario, pilot projects proliferate but do not scale. Sandboxes remain isolated.
- Scenario 3. Full Flywheel Activation, GDP impact: Africa captures 4% of global AI dividend (US $1 trillion). In this scenario, the flywheel turns at scale. Regional GPU hubs and compute credits make infrastructure accessible.
From Inputs to Impact: An AI-Productivity Theory of Change
- Inputs: The Five Enablers, Activated Together.
- Outputs (2025–2030): Tangible Readiness Milestones : 60 FAIR-compliant national and regional data catalogues published 4 Data Embassies fully operational (one per high-capacity REC), with two edge-corridor nodes financed and under construction; keeping the network on track to reach six embassies by 2035 under the same AU charter 30,000 mid-career ML engineers and 5,000 monthly hours of MLOps training delivered 20 countries adopt audit-ready AI laws and algorithm registers US $5 billion in growth-stage capital deployed across five priority sectors.
- Outcomes (2028–2032): Systems Begin to Shift : 40 percent reduction in training and inference costs across compute-active RECs, 30 percentage-point jump in AI adoption by SMEs and public service platforms, 30 percent female representation in AI leadership and technical roles. Private AI capital expenditure at or above 1 percent of GDP in Scale Accelerator countries
- Impact (2035 and beyond): The Dividend Realised : US $1 trillion in incremental GDP by 2035, 35–40 million net new digital and digitally enabled jobs, 10 percent of global training tokens are African and multilingual, embedding the continent into frontier AI models
Delivering the roadmap, how Africa captures the AI dividend.
A Three-Phase Path to Scale (2025–2035).
Risk Management and Governance
The roadmap recognises five systemic risks and ring-fences them from day one, assigning owners and mitigation instruments.
- Coordination & trust gaps : REC-level incentive compacts, public progress dashboards, and an AU, Capacity-Building Facility for regulators.
- Policy drift Annual REC-level scorecards and model law adoption tracker
- Funding shortfall : US $4B first close with 20% first-loss guarantees
- Cybersecurity : failure Shared cyber response frameworks and mandatory breach insurance
- Compute downtime : Dual-feed power for GPU hubs; PPAs with renewable providers
- Talent flight : Bonded fellowships and return-path sabbaticals
A call to action : delivering the AI dividend together
Priority Deliverables by December 2026
- AU Commission & RECs : Adopt continent-wide FAIR-data and AI-trust frameworks. Publish the first REC scorecards covering data, compute, skills, trust and capital. Break ground on two GPU “Data Embassies” and launch the pooled voucher pilot.
- National Governments : Close the largest gap for each archetype: Foundation Builders: digital-ID plus baseline cyber-hygiene active in ≥ 50 % of districts. Scale Accelerators: at least one interoperable AI sandbox outside FinTech. Catalytic Agents: publish sovereign compute audits and AI-risk registers. Adopt an AfDB benchmark: earmark 5% of ICT procurement for open- source public-good solutions.
- DFIs & Donors : Close a US $4 billion first-close ofthe Africa AI Growth & Innovation Fund (with 20 % first-loss guarantees). Approve a concessional-finance envelope that subsidises voucher access for Foundation-Builder countries until Data Embassies reach scale. Endow an AU-level facility to train 1,000 AI-risk supervisors.
- Private-Sector Innovators Provide a compute-voucher pool sized jointly with the AU (stretch target ≈ 50 million GPU-hours). Publish bias-audited model cards in ≥ 3 AU languages. Co-finance two edge-compute corridors in Scale-Accelerator RECs.
- Universities & Technical Institutes : Issue a continent-wide micro- credential standard for AI skills. Launch the pipeline toward 15 000 certified mid-career ML engineers by 2028 and certify the first 5 000 by 2026. Release an open 5 bn-token multilingual corpus (stretch target). Open four MLOps labs, one per high-readiness REC.
The G20 Opportunity: A Strategic Win-Win
Africa’s AI dividend is Africa’s to capture, but the G20 has a direct stake in its success. G20 engagement is not charity. It is a strategic, economic and governance investment; one that expands global markets, improves model robustness, strengthens multilateral governance, and advances the G20’s own commitments to fairness and inclusion.
By backing Africa’s AI readiness roadmap, the G20 can:
- Help shape shared norms through the AU Trust Network Strengthen global datasets through Africa’s multilingual, real-world contexts
- Reduce risk by supporting regulatory sandboxes and open infrastructure
- Accelerate global digital SDG progress
- This is a practical opportunity and a geopolitical signal. The world’s fastest-growing continent is not asking to be uplifted. It is offering a partnership. The G20 should meet it.
Les Wathinotes sont soit des résumés de publications sélectionnées par WATHI, conformes aux résumés originaux, soit des versions modifiées des résumés originaux, soit des extraits choisis par WATHI compte tenu de leur pertinence par rapport au thème du Débat. Lorsque les publications et leurs résumés ne sont disponibles qu’en français ou en anglais, WATHI se charge de la traduction des extraits choisis dans l’autre langue. Toutes les Wathinotes renvoient aux publications originales et intégrales qui ne sont pas hébergées par le site de WATHI, et sont destinées à promouvoir la lecture de ces documents, fruit du travail de recherche d’universitaires et d’experts.
The Wathinotes are either original abstracts of publications selected by WATHI, modified original summaries or publication quotes selected for their relevance for the theme of the Debate. When publications and abstracts are only available either in French or in English, the translation is done by WATHI. All the Wathinotes link to the original and integral publications that are not hosted on the WATHI website. WATHI participates to the promotion of these documents that have been written by university professors and experts.