The integration of Artificial Intelligence (AI) into various sectors is transforming the global economy, and Africa is no exception. However, the continent’s growth potential continues to be constrained by a significant power gap, limiting its ability to fully benefit from the opportunities AI offers.
The International Monetary Fund (IMF) projects that artificial intelligence could increase Sub-Saharan Africa’s economic output by approximately 4% over the next decade, providing a significant boost to the region’s GDP. This growth depends on AI’s ability to improve productivity, efficiency, and competitiveness across industries. The GSMA further estimates that AI could contribute $2.9 trillion to Africa’s economy by 2030, largely through advances in agriculture, energy, and healthcare. However, this represents a best-case scenario that depends on overcoming several fundamental barriers. Unlocking even a portion of this potential will require closing major infrastructure and skills gaps, often referred to as the “power gap.” Without addressing these challenges, much of AI’s projected economic value may remain unrealised.
READ ALSO: Building Africa’s Digital Workforce: Preparing Tomorrow’s Tech Leaders
The most significant obstacle is Africa’s persistent electricity shortage. AI systems require reliable, high-capacity power to process data and operate cooling systems. According to the International Energy Agency (IEA), around 600 million Africans still lack access to electricity, making it difficult to deploy even basic digital technologies. World Bank data also shows that the average data centre in Africa is considerably smaller than those in Europe, largely because of unreliable electricity supply. Many facilities rely heavily on expensive and polluting diesel generators for backup power, driving up operating costs and making it more expensive to run AI infrastructure in many African cities than in other parts of the world. This discourages local innovation and limits the continent’s ability to retain and manage its own data.
Another major challenge is low internet penetration and the high cost of mobile data, both of which prevent AI-powered services from reaching large numbers of people. While the International Telecommunication Union (ITU) previously estimated internet penetration at around 43%, more recent figures from GSMA Intelligence show that mobile internet penetration in Sub-Saharan Africa stood at just 27% in 2023, leaving a usage gap of 60%. This means hundreds of millions of people live within network coverage but remain offline, largely because internet access is too expensive. In some countries, 1GB of mobile data costs nearly 6% of the average monthly income, far above the UN Broadband Commission’s affordability target of 2%. As a result, cloud-based AI applications remain inaccessible to much of the population, limiting their adoption and reducing the amount of locally generated data available for innovation.
The shortage of digital skills adds another layer to the challenge, affecting both the development and maintenance of AI systems. The World Economic Forum has warned of a global shortage of skilled technology professionals, and the problem is especially evident across Africa. The International Finance Corporation (IFC) notes that only a small proportion of graduates specialise in STEM disciplines. A 2023 study by AI firm Amyoli Labs and the University of Pretoria found that the limited number of experienced machine learning engineers has contributed to an ongoing brain drain, with many skilled professionals working remotely for overseas companies. As a result, local startups, universities, and governments often lack the expertise needed to build AI systems suited to African realities, increasing dependence on foreign-developed models that may not reflect local languages, cultures, or environmental conditions.
At the same time, AI offers practical solutions that can help address some of these very challenges. Google DeepMind’s collaboration with Kenya’s national grid operator, Ketraco, demonstrates how AI can predict and optimise geothermal energy integration, reducing dependence on fossil fuel backup systems. The UN Food and Agriculture Organisation also highlights the AI-powered Nuru app, which is being used across East Africa to help farmers diagnose crop diseases using offline-compatible smartphones. Researchers at the University of Oxford’s Smith School have found that precision agriculture tools such as these can increase crop yields by as much as 20% while reducing water and energy consumption. These examples show how AI can improve infrastructure efficiency while helping address the resource constraints that have slowed its wider adoption.
Healthcare provides another compelling example of AI’s potential, particularly in areas where infrastructure remains limited. A landmark study published in The Lancet Digital Health in 2022 validated an AI system in Zambia capable of interpreting chest X-rays for tuberculosis with expert-level accuracy using portable, battery-powered equipment. Similarly, the Philips Foundation’s AI-enabled ultrasound projects in Kenya use intelligent software to guide midwives through obstetric scans, helping to address the shortage of radiologists. These edge-computing solutions operate on solar-powered devices rather than relying on a stable national electricity grid, demonstrating how AI can be adapted to Africa’s infrastructure realities instead of waiting for those conditions to improve.
The financial services sector provides one of the strongest examples of AI delivering measurable economic benefits. Mobile money platforms have successfully combined AI with USSD and SMS technologies, allowing users to access financial services without requiring high-speed internet. M-Pesa, which serves more than 51 million customers, uses AI-driven credit scoring to support microloan products such as M-Shwari. Safaricom’s financial reports show that these services have disbursed billions of dollars to people who previously had little or no access to formal credit. Research from the Brookings Institution also shows that AI models built on mobile transaction histories, rather than traditional credit bureau records, have helped bring more than 30 million previously “invisible” borrowers into the formal financial system. This demonstrates that meaningful financial inclusion is possible even where electricity infrastructure remains limited, provided reliable mobile networks are available.
Failing to close the power and skills gap could significantly widen regional inequality and deepen what many describe as the “compute divide.” The IMF suggests that without substantial investment, Africa’s gains from AI could remain minimal, leaving the continent largely dependent on foreign technology while continuing to export raw data. Research from the Oxford Internet Institute indicates that African languages and contexts account for less than 0.1% of the training data used in large language models such as GPT-4. If AI development continues to take place primarily in high-income, energy-rich countries, many AI systems will struggle to serve African users effectively, potentially reinforcing bias in areas such as lending, healthcare, and law enforcement. This risks creating a new form of digital dependency in which Africa supplies data, but imports finished AI products, weakening local competitiveness and technological sovereignty.
Addressing these challenges will require coordinated investment in renewable energy, digital infrastructure, and skills development. The World Bank’s Digital Economy for Africa initiative advocates large-scale investment in solar-powered edge computing facilities and mini-grids designed to support data centres. At the same time, the African Institute for Mathematical Sciences (AIMS) estimates that the continent needs to train tens of thousands more data scientists using locally relevant curricula. Google’s recent investment in fellowships and research laboratories focused on African language AI also points in the right direction. Together, these initiatives highlight the need to develop reliable electricity, affordable digital devices, and local technical expertise simultaneously, recognising electricity not simply as a public utility but as an essential input for the digital economy.
In conclusion, Africa’s projected $2.9 trillion AI opportunity is both real and achievable, but only if the continent addresses the underlying infrastructure and skills challenges that stand in its way. The idea that AI can simply overcome weak infrastructure on its own is misleading. Instead, AI increases the cost of inadequate infrastructure by making access to reliable electricity and connectivity even more important. Moving forward, Africa must prioritise AI solutions designed for low-power and low-connectivity environments, using efficient machine learning models and solar-powered technologies where possible. By investing in local talent, expanding reliable energy access, and developing AI that reflects African realities, the continent can turn today’s power gap into the foundation of a resilient, competitive, and self-reliant digital economy.
