Africa’s AI moment: How coordinated investment in 'green' computing can unlock $1.5 trillion

Africa’s talent, resources and vision for innovation in artificial intelligence (AI) are clear. Across the continent, innovators are already building solutions – from crop disease detection to medical imaging – that demonstrate AI’s transformative potential.

Yet without affordable graphics processing units (GPUs), relevant datasets, aligned demand and coordinated investment, new data centres risk sitting empty. GPUs are specialized computer chips originally designed to render images quickly. Today, they are essential for AI because their ability to process many tasks in parallel makes them ideal for training and running complex AI models.

The stakes are immense: beyond the potential $1.5 trillion projected in economic value by 2030, AI offers tools to address development challenges while strengthening technological sovereignty. However, there is a fundamental constraint.

Despite housing 18% of the world’s people, according to the Africa Data Centres Association and the World Economic Forum, Africa accounts for less than 1% of global data centre capacity – and far less of the GPU infrastructure that powers AI.

Without accessible and affordable computing infrastructure, Africa risks being locked out of the AI economy, becoming a consumer of imported solutions rather than a producer of homegrown innovation. The question is no longer whether Africa has the talent or demand, but whether it can meet both – and build the sustainable AI infrastructure to unlock them.

The challenge is one of simultaneous abundance and scarcity. On one side, demand for compute is rising fast. Researchers, startups and small- and medium-sized enterprises (SMEs) already face 7 million GPU hours of unmet demand for model training over the next three years, while only 5% of AI innovators have reliable access to advanced compute – a sign of latent demand held back by the absence of infrastructure rather than a lack of interest.

On the supply side, capacity is expanding.

Nairobi’s IXAfrica campus is operational and in 2025, entered a strategic partnership with Safaricom to deliver AI-ready infrastructure. Microsoft and Abu Dhabi-based G42 announced a commitment with the Kenyan government, with timelines still evolving.

Senegal’s Diamniadio National Datacenter, launched in 2021, and Cassava Technologies’ partnership with NVIDIA to deploy 12,000 GPUs through its AI Factory both signal increasing national and private sector investment in advanced compute.

While many of these projects are still scaling up, they illustrate growing momentum towards a more connected and capable African compute ecosystem.

Alongside these infrastructure developments, African-led organizations are helping to translate emerging capacity into accessible, AI-ready compute. Companies such as Udutech are pioneering GPU-as-a-Service models that allow impact-first researchers, startups and enterprises to access high-performance computing without the prohibitive cost of owning hardware.

Others, including Amini.ai, are building deep-tech infrastructure across Africa and the Global South, designed to close data and compute gaps. This infrastructure processes fragmented datasets, develops foundational models, and deploys insights that integrate seamlessly into business workflows, making data accessible and actionable.

Together, these initiatives complement the physical data centre investments by making advanced compute more usable, affordable and locally relevant – a critical bridge between hardware availability and real-world innovation.

Yet, many data centre operators still report under-utilized racks and slower-than-expected returns. The constraint is shifting from physical capacity to uncertain workload demand and the high cost of GPU access. Global supply dynamics continue to favour hyperscalers, leaving smaller African buyers facing steep prices and long delivery times.

In some markets, the cost of a single high-end GPU rivals the annual income of an average citizen, effectively pricing out local innovators. While many remain in the early stage of development, emerging GPU-as-a-Service platforms aim to lower these barriers by pooling demand and procurement.

In short, Africa’s compute gap may be less about new data centres and more about aligning GPU economics, GPU economics, data optimization, power costs and demand orchestration.

The deficit persists because of self-reinforcing market failures:

These failures keep utilization low, prices high and innovation suppressed – a cycle of underinvestment that risks leaving only the best-capitalized actors served.

Breaking this cycle could trigger the opposite dynamic: early access enables innovation; proven use cases reduce risk perceptions; higher utilization cuts costs; and renewable energy stabilizes supply. Early examples illustrate this potential.

Tunisia’s InstaDeep, later acquired by BioNTech, invested in an NVIDIA DGX system in 2018, catalyzing its growth to a $680 million valuation. The country’s leading incubator then shared a similar system among eight startups, which in turn attracted new capital. In Kenya, geothermal-powered facilities near Olkaria, while under development, aim to offer low-carbon, reliable compute for AI and cloud workloads.

These cases show that GPU infrastructure delivers a transformative impact when paired with talent development and demand aggregation. A call to coordinated action, Africa’s compute challenge is a systems problem requiring collective intervention.

Africa’s compute challenge is a systems problem requiring collective intervention. Strategic steps include:

Commercial investors recognize the opportunity, but face distorted risk signals and nascent demand.

Philanthropy and impact capital can:

Market creation in infrastructure typically requires five to seven years. Catalytic capital is necessary to bridge this gap and prepare a viable market for commercial sustainability and scale.

The route to Africa’s AI sovereignty runs through green compute infrastructure but progress depends on systemic investment. Africa has the renewable energy advantage, the talent and emerging solutions for data and demand. What is needed is coordinated investment that turns these assets into the backbone of Africa’s AI future.

The question is not whether Africa needs its own compute, but whether interests can be aligned to build the systems that unlock prosperity, sovereignty and innovation, or remain dependent on others.

This is the vision of the Africa Green Compute Coalition (AGCC) – a flagship of the AI Hub for Sustainable Development that is catalyzing research and development, as well as building new coalitions. AGCC aims to demonstrate how green, accessible compute infrastructure can unlock innovation and shared prosperity across Africa.

As a systems orchestrator, the coalition connects policy, finance, energy and technical capacity to ensure that infrastructure investments translate into real-world use cases, measurable inclusion and economic impact.

The authors would like to thank Tonee Ndungu, one of the founding members of the AGCC, for his contributions to this piece.