Nvidia Partners with Cassava to Build AI-Ready Data Centres Across Africa in US$700 Million Deal

Nvidia has entered into a landmark US$700 million deal with Cassava Technologies (Cassava) to establish AI-ready data centres across several key African markets. This partnership, first announced in March 2025, represents Nvidia’s first direct infrastructure deployment on the continent. The goal is to provide access to Nvidia’s advanced graphics processing units (GPUs) in regions that have historically lacked the necessary computing capacity for AI development and data-driven industries.

This collaboration stands in stark contrast to China’s state-led investments in Africa’s digital infrastructure. Instead, Nvidia’s approach prioritizes private-sector execution, with Cassava taking the lead in designing and delivering data centres specifically engineered to support AI development in diverse sectors like healthcare, agriculture, and fintech.

Ziaad Suleman, CEO of Cassava Technologies for South Africa and Botswana, emphasizes the significance of this partnership for African enterprises, stating: “Africa has historically contended with receiving second-rate technology.” He further explains that Nvidia’s support will play a pivotal role in altering this narrative and will bring high-performance computing capabilities to the continent’s AI developers and data-driven industries.

The project’s first phase was launched in June 2025, with the delivery of 3,000 Nvidia GPUs to a Cassava-built data centre in South Africa. Over the next three to four years, a series of AI factories will be developed, deploying an additional 12,000 GPUs across new facilities in Egypt, Nigeria, Kenya, and Morocco.

Eric Omorogieva, an analyst at the New Lines Institute, notes the potential transformative impact of this partnership on local AI capabilities. “Few companies match Nvidia’s dominance in the AI sector due to its advancements in GPUs,” he says. “With a 93% share in the global GPU market, Nvidia is the driving force behind the advanced chips powering top systems around the world. For African start-ups, access to these chips has been limited due to high costs and lack of availability. By bringing them to Africa, Nvidia will help bridge the computing gap, providing much-needed access to hardware for key sectors.”

In addition to the GPU deployment, Cassava has also signed a memorandum of understanding with the South African AI Association, offering over 3,000 AI professionals direct access to the GPUs. This move is designed to develop human capital alongside the growing technical infrastructure.

The Nvidia–Cassava collaboration specifically targets the continent’s AI infrastructure deficit. According to the UN Development Programme, a mere 5% of African AI practitioners currently have access to the computing power necessary for advanced work. Of that small percentage, only a fifth have on-premises access to GPUs, while the majority are reliant on limited cloud resources.

Eric Omorogieva elaborates, explaining that addressing this computational gap requires specialized infrastructure for AI development. The introduction of AI factories will provide African organizations with better system performance, improved scalability, and the ability to turn raw data into actionable insights. These new data centres will be optimized for AI workloads, offering local developers and institutions low-latency, high-bandwidth access to GPU clusters for model training and inference—capabilities that have been largely unavailable within African borders until now.

This partnership also comes at a time of changing global dynamics surrounding African infrastructure. While China’s Belt & Road and Digital Silk Road programs have historically funded state-led initiatives across the continent, the United States is now promoting a strategy known as “commercial diplomacy”—focusing on private investment and expanded trade to build long-term digital partnerships.

Nvidia’s involvement in Africa’s data centre space signals a broader effort by U.S. companies to counter China’s increasing influence in AI systems. While Chinese platforms like DeepSeek R1 and Qwen offer more affordable models, U.S. companies maintain an edge in hardware performance, particularly in GPUs and supercomputing. This performance advantage, especially in GPUs, could play a decisive role in shaping the future of AI infrastructure worldwide.

Eric Omorogieva concludes that Nvidia’s involvement will enable African nations to move beyond being passive consumers of AI technology. With access to cutting-edge chips and specialized infrastructure, African nations will be empowered to develop indigenous AI systems that can compete on a global scale.