…But 900 million Africans remain offline as the continent struggles with inadequate digital infrastructure and skills
Artificial intelligence (AI) could contribute as much as $1 trillion to Africa’s gross domestic product (GDP) by 2035, create between 35 million and 40 million digital jobs, and generate nearly $150 billion in annual tax revenue, the World Bank projects.
However, the continent’s ability to realise the enormous economic opportunity remains constrained by inadequate digital infrastructure, unreliable electricity, limited computing capacity, skills shortages and poor representation of African languages in AI training datasets.
The World Bank disclosed this in its October 2026 Africa Economic Update, titled Building AI Readiness, warning that Africa’s potential gains from artificial intelligence remain significantly higher than its current capacity to deploy the technology.
According to the report, approximately 900 million Africans remain offline, while the continent, which accounts for 18 per cent of the world’s population, possesses just 0.6 per cent of global data centre capacity. Only about 5 per cent of African data centres are currently equipped to support advanced AI workloads.
The report also estimates that only 2 per cent of global AI training data originates from Africa, highlighting the continent’s limited contribution to the datasets shaping the development of artificial intelligence.
The World Bank said that at current preparedness levels, AI integration in Sub-Saharan Africa could contribute only between 0.2 and 4 per cent of cumulative GDP over the next decade, with even this projection assuming adoption accelerates beyond the pace supported by existing infrastructure.
The difference between current preparedness and the ambitious full-activation scenario underscores the scale of investment and institutional reforms required to unlock the technology’s economic benefits.
The report identifies Nigeria, Kenya and South Africa as the major centres of AI activity in Sub-Saharan Africa. The three countries account for much of the region’s innovation, venture capital investment, research output and advanced AI applications.
Among larger firms, AI usage stands at 44 per cent across Kenya and Nigeria, compared with 61 per cent in the United States.
Nigeria has also introduced Service-Wise GPT as an official AI assistant for federal civil servants, providing access to information on public service rules and procedures.
The World Bank, however, noted that government adoption of AI across the region remains at an early stage, with applications concentrated largely in administrative and back-office functions rather than citizen-facing services.
Its analysis of government AI procurement records found that 90.8 per cent of identified AI systems supported internal government functions, while only 1.4 per cent employed generative AI. The report said the priority for African governments should be to move beyond basic experimentation and integrate AI into practical service delivery, economic planning and institutional operations.
The World Bank identified electricity, internet connectivity, affordable devices, data and computing infrastructure as essential foundations for AI adoption. It noted that although mobile network coverage has expanded across the continent, unreliable electricity, expensive internet access and the high cost of digital devices continue to exclude significant portions of the population.
The data centre gap is equally significant.
The report said investments in African data centres and cloud infrastructure have reached approximately $8 billion over the past decade, but capacity remains concentrated in a handful of countries.
South Africa, Nigeria and Kenya lead the continent in colocation data centre capacity. Nevertheless, their capacity relative to GDP per capita remains modest compared with global standards.
The World Bank also highlighted the underrepresentation of African languages in AI training datasets. For instance, Swahili has only 710 datasets on Hugging Face, compared with 88,844 for English.
It warned that AI models trained predominantly on data that do not reflect local languages and contexts may produce unreliable or exclusionary outcomes.
Addressing concerns about technological unemployment, the report suggested that AI is more likely to complement workers than replace them in the near term across Sub-Saharan Africa. It estimated that only 2.6 per cent of jobs in the region face near-term automation risk, compared with 14.2 per cent in high-income economies.
By contrast, 15.2 per cent of jobs have potential for AI-related augmentation, where technology could improve workers’ productivity and effectiveness. However, the report cautioned that the benefits would depend on workers having access to reliable connectivity and the skills required to use AI tools.
It cited Farmerline’s Darli AI, which reached 110,000 farmers in more than 27 languages within eight months, as an example of how locally adapted AI applications can address practical development challenges.
Kenya’s M-Shule, which has supported improvements in foundational literacy, was also highlighted.
The World Bank argued that Africa’s immediate opportunity lies less in competing to develop frontier AI models and more in adopting and adapting affordable applications to address challenges in agriculture, education, healthcare, finance, logistics and public administration.
Beyond infrastructure and skills, the report identified institutional capacity as a critical factor in determining whether AI adoption will deliver inclusive development.
It disclosed that 39 countries in Sub-Saharan Africa have data protection laws, while 45 have cybersecurity legislation.
However, the region recorded an average score of 41.6 out of 100 on the Oxford AI Governance Pillar, below the global average of 52.6.
According to the World Bank, the challenge is no longer simply developing additional AI strategies or regulations, but ensuring that existing frameworks are effectively implemented.
It called for stronger regulatory institutions, technical expertise, coordination across government agencies, responsible AI procurement and safeguards against discrimination, surveillance and misuse of automated decision-making.
The report also advocated regional cooperation to pool investments in computing infrastructure, data systems, specialised talent and research.
It said individual African economies are often too small to independently support the substantial fixed costs associated with advanced AI infrastructure.
