Abuja:A Nigerian technology company, Udu Technologies, has launched a cloud platform providing Graphics Processing Units (GPU) access to enhance Artificial Intelligence (AI) applications for developers, businesses, and governments across Africa.
According to News Agency of Nigeria, the platform, named Africa GPU Hub (AGHCloud.ai), offers on-demand access to GPUs, which are essential for training and running AI models and other computationally intensive tasks. The platform is set to provide GPU computing at a cost of less than one dollar per hour.
The Chief Executive Officer of Udu Technologies, Mr. Alexander Tsado, stated that the platform includes pre-configured AI development tools like PyTorch, TensorFlow, vLLM, and LoRA, allowing users to develop and deploy AI applications without the need to build their own GPU infrastructure. He emphasized that this initiative aims to lower the financial and technical barriers to accessing high-performance computing in Africa.
The company has deployed NVIDIA H100 and Blackwell Pro 6000 GPU clusters in Kenya, Malawi, Rwanda, South Africa, Togo, and Zambia. Tsado highlighted the importance of complementing access to computing infrastructure with skills development, which the company supports through collaboration with Alliance4AI, benefiting professionals in fields such as medicine, logistics, and finance.
The initiative has already reached approximately 2,000 developers and supported over 30 government AI use cases in sectors like agriculture, customs, digital public infrastructure, mining, and education. Tsado noted that expanding access to affordable computing infrastructure and AI skills is crucial for empowering African developers and organizations to create and deploy AI solutions within the continent.
Oluwafunmilayo Olumoko, Head of Customer Success at Udu Technologies, added that deploying infrastructure within African markets allows users access to computing resources closer to their data, supporting data residency and sovereign AI infrastructure needs. She also mentioned that the company’s GPU virtual machines and optimised open-source large language models could reduce token costs by up to 60 percent for organizations running AI applications at scale.