The Biggest AI Problem in Africa May Not Be AI

The Biggest AI Problem in Africa May Not Be AI Artificial intelligence (AI) is moving fast. Chatbots can write, software can generate code, computers can analyse medical information, and businesses are using AI to automate work that once required hours of human effort. Africa wants to be part of this transformation. And it should. But there is a problem we don’t talk about enough: You cannot build an AI-powered economy on digital infrastructure that is not strong enough to support it. The biggest AI problem in Africa may not be the AI itself. It may be electricity, internet access, computing power, skills and affordability. AI needs more than an app It is easy to think of AI as simply opening ChatGPT or another AI tool on your phone. But behind that simple screen is a huge technical system. AI models run on computers called servers. Servers are powerful computers that provide services and process information for other computers and users. Large AI systems require enormous amounts of compute, meaning computing power used to train and run AI models. These machines live in data centres — buildings filled with computers, networking equipment, cooling systems and power infrastructure. And all of this needs electricity. So when we ask whether Africa is ready for AI, we should not only ask: “How many people are using AI?” We should also ask: “Do we have the infrastructure to make AI widely useful and affordable?” 1. Electricity comes first This is the most basic problem. Computers need electricity. Data centres need a lot of electricity. AI systems need computing infrastructure, and computing infrastructure cannot operate without reliable power. The International Energy Agency (IEA) reported that electricity consumption by data centres increased by 17% globally in 2025, while electricity use by AI-focused data centres grew even faster. That matters for Africa because AI is not just about buying software. If a country wants to develop serious AI infrastructure, it needs reliable electricity to operate the physical machines behind it. This does not mean every African country needs to build giant AI data centres tomorrow. It means reliable and affordable electricity is part of the AI conversation. 2. The internet is still a barrier AI also needs connectivity. A person cannot use an online AI service effectively if their internet connection is unreliable, extremely slow or too expensive. The World Bank describes connectivity as the gateway to AI participation and highlights continuing differences in internet affordability, speed and usage between countries. The problem is therefore bigger than simply saying: “Africa needs more internet users.” People need internet connections that are good enough to actually use modern digital services. And they need devices capable of accessing those services. 3. Then there is compute This is one of the most important terms in the entire AI conversation. Compute simply means computing power. Think of it as the muscle that allows computers to perform difficult calculations. Modern AI systems can require specialised processors called GPUs, or Graphics Processing Units. GPUs are computer chips that can perform many calculations at the same time, making them particularly useful for AI workloads. The problem is that advanced computing capacity is heavily concentrated outside Africa. The United Nations Conference on Trade and Development (UNCTAD) says AI infrastructure requires computing power, servers and data centres, and notes that most supercomputers and data centres are located in developed countries. The World Bank also identifies compute as one of the major foundations needed for inclusive AI development. This creates an uncomfortable question: If Africa does not have enough affordable access to compute, how much AI can it build locally? Cloud computing — renting computing resources over the internet instead of owning the physical machines — can help. But relying heavily on computing infrastructure elsewhere also means Africa remains dependent on infrastructure it does not control. That is why compute is becoming a strategic issue. 4. Africa needs people who understand AI Hardware alone will not solve the problem. You can build a data centre and still have a shortage of people who know how to use, manage and develop the technology. AI needs different levels of skills. People need basic digital literacy, meaning the ability to confidently use digital devices and services. Businesses need people who understand how to apply AI to real problems. And advanced AI development requires specialists in areas such as machine learning (ML) — a branch of AI where computers learn patterns from data — as well as data science, software engineering and related fields. The African Development Bank’s recent AI roadmap identifies skills as one of the major foundations of Africa’s AI readiness and proposes a target of 3 million AI-capable professionals by 2030. This is why AI education cannot only be about teaching people how to write better prompts. People also need to understand how AI works, where it can be used, its limitations and how to build with it. 5. And then comes affordability This may be the issue ordinary people feel most directly. AI may exist, but can people afford the internet, electricity, smartphones, computers and services needed to use it? There is a difference between AI being available and AI being accessible. A powerful AI tool is not particularly useful to a student who cannot afford enough data to use it. It is not very useful to a small business whose electricity and internet costs are already difficult to manage. And it is difficult for a developer to build sophisticated AI products if access to powerful computing resources is prohibitively expensive. UNCTAD identifies affordable electricity, internet access, computing power, devices and digital skills as important parts of the infrastructure required for AI adoption. Africa does not have to copy Silicon Valley There is another important point. Africa does not need to reproduce the exact AI infrastructure of the United States or China to benefit from AI. There are opportunities to build smaller, practical systems around African problems. For example, AI can be applied to agriculture, education, healthcare, financial