AI holds immense potential to transform economies and improve livelihoods in Sub-Saharan Africa yet large infrastructure gaps in the region are threatening to create a new digital divide. This blog explores current obstacles to creating resilient computational power, or compute, ecosystems in Africa and highlights pathways for the continent to unlock the transformative power of AI.
Limited data centre capacity and high compute costs are holding Africa back
Deploying AI requires access to infrastructure and computing power to train and run AI models. One major issue in Africa is the shortage of data centres across the continent, which provide essential storage and processing power. Africa has fewer data centres than any other major region, meaning much of the continent’s data must travel to servers in Europe or North America for processing. Submarine cables are the backbone of the digital economy, carrying almost 99% of the world’s data traffic. Yet, Africa is connected to just a handful of these cables. This creates high latency and makes Africa’s digital infrastructure more vulnerable to service disruptions, highlighting the need for local data centres. Estimates suggest that to meet growing demands, African countries will need to more than double their data centre hosting capacity by 2030.
Map of global data centers
Source: Data Center Map
The cost of compute is another key challenge. The high price of graphic processing units (GPUs), essential for high-performance computing, is a major hurdle for African innovators. Although the absolute costs of computing are comparable across regions, they become disproportionately high for those in Africa when factoring in local income levels. This is likely to impact the development of Generative AI in Africa, in particular. In Kenya and Senegal, the price of a GPU represents 75% and 69% of GDP per capita respectively . Rising global demand and limited supply continue to drive up prices, with Nvidia’s most powerful chips selling for as much as $40,000. In countries such as Ethiopia, high import taxes on hardware and devices add to the cost burden, while access to cloud services is constrained by foreign exchange restrictions. In Nigeria, the issue has been further exacerbated by the devaluation of the local currency.
New opportunities for African compute are emerging
As local compute ecosystems continue to develop, countries in Sub-Saharan Africa have an opportunity to tap into their mobile-first markets and leverage edge computing. Edge computing processes data closer to where it’s generated – right on users’ devices. The steady rise of smartphone penetration in Africa, expected to reach 88% by 2030, is creating opportunities for mobile-based edge computing. Smartphones are already facilitating computer vision tasks like crop disease detection in agriculture and digital record-keeping in healthcare. Ensuring this technology is successful and inclusive requires making it work on low-end devices. An example is Kenya-based deep tech startup Fastagger, which develops ML models capable of running on lower-end smartphones with inexpensive chips and low connectivity.
Sub-Saharan Africa: smartphone adoption. Percentage of connections (excluding licensed cellular IoT)
Source: GSMA Intelligence
Another opportunity area is in state-of-the-art data centers. Despite current limited capacity, Sub-Saharan Africa has shown a rapid rise in server investments, outpacing the global average by 13%. Countries like Nigeria, Rwanda, and Senegal are attracting significant investments. With less dependency on fossil fuels than many other regions, Sub-Saharan Africa countries are well-positioned to deploy clean data center infrastructure. In Zambia, Ethiopia, Kenya and the DRC for example, over 80% of energy is generated from renewables. The Olkaria Ecocloud Data Centre in Kenya, powered entirely by geothermal energy, is a prime example of innovation in green data center technology.
How can Africa strengthen its compute capacity?
1. Strengthen baseline infrastructure and promote renewable energy
Strategic partnerships and investments can help countries in Sub-Saharan Africa build the necessary infrastructure and develop sustainable compute technologies. For example, Microsoft, AI company G42 and the Kenya government announced a $1 billion investment, with part of the funding towards building a green data centre.
2. Provide credits and donate hardware
To overcome affordability challenges, governments and development partners could promote public-private partnerships to unlock access to cloud credits, GPUs and laptops to AI startups and researchers. Incentives like tax benefits could encourage participation.
3. Increase opportunities for funding and resource sharing
Governments can facilitate collaboration by establishing resource-sharing consortiums and hosting investor events to attract local and foreign investment. Supportive regulations, tax incentives, and subsidies can further enhance data center capacity and cloud service access.
4. Enhance edge computing capabilities
Making mobile devices more affordable and supporting device financing are critical steps to equitable access and edge computing growth. Optimising devices for low-resource settings will also boost usage in rural areas with limited connectivity.
“If we equip people with better phones and provide adequate financing, we could shift the focus and invest in better computing at the edge rather than computing in the cloud.” – Startup in Nigeria
5. Build institutional capacity
Countries should assess their compute needs to create investment roadmaps that balance short-term and sustainable computing assets. Dedicated agencies could help coordinate efforts across stakeholders and promote regional and international collaboration.
The Central Insights Unit is currently funded by the UK Foreign, Commonwealth & Development Office and supported by the GSMA and its members.
The views expressed do not necessarily reflect the UK government’s official policies.
