Kenya has officially become the first country in East Africa to offer rentable, high-performance GPU infrastructure for artificial intelligence development, following the launch of NVIDIA-powered AI servers by Cassava Technologies at its Nairobi data center.
The service, operated under Africa Data Centres (a Cassava subsidiary), allows startups, researchers, academic institutions, and enterprises to rent GPU clusters on-demand paying only for compute time used without the prohibitive upfront cost of building private AI infrastructure.
A Game-Changer for Local AI Innovation:
Until now, Kenyan and East African developers training large language models, computer vision systems, or generative AI applications have relied heavily on cloud providers in the U.S., Europe, or Asia. This introduced three major pain points:
High latency: – Critical for real-time AI applications in healthcare, agriculture, and logistics. Data sovereignty risks: – Sensitive local data leaving the continent.
Cost unpredictability: – Due to currency fluctuations and egress fees.
The new Nairobi-based AI cloud eliminates these barriers.
“This is not just about hardware. It’s about control, speed, and affordability,” said Hardy Pemhiwa, President and CEO of Cassava Technologies. “For the first time, a Kenyan agritech startup can train a crop disease detection model using local satellite and soil data on local infrastructure at a fraction of the cost.”
Technical Specs and Accessibility
The initial deployment includes:
NVIDIA H100 and A100 GPUs in scalable clusters
High-speed NVLink interconnects for multi-GPU training
Integration with popular ML frameworks (PyTorch, TensorFlow, JAX)
Secure, compliant environment certified for ISO 27001 and local data residency Pricing starts at $2.50 per GPU-hour, with volume discounts and academic grants available. A free tier offers 50 GPU-hours per month for verified Kenyan university researchers.
Early Adopters Speak:
iZola Health, a Nairobi-based AI mental health platform, is already using the cluster to fine-tune multilingual therapy chatbots in Swahili, Luo, and Kikuyu.
Twiga Foods is accelerating supply-chain forecasting models to reduce post-harvest losses.
Strathmore University’s @iLabAfrica has migrated its AI ethics and computer vision labs to the platform.
Why Nairobi? Why Now?
Kenya’s selection as the launch site reflects its growing role as East Africa’s tech hub: M-Pesa and fintech innovation have built a strong digital payments ecosystem. Konza Technopolis and Nairobi Garage foster startup density.
Government support via the Digital Economy Blueprint and National AI Strategy 2030. Cassava plans to expand the service to Rwanda, Uganda, and Tanzania within 18 months, with a long-term vision of a pan-African AI compute grid.
The Bigger Picture:
The launch comes amid a continental push for AI sovereignty. At the Transform Africa Summit in Conakry this week, leaders warned that without local infrastructure, Africa risks becoming a “data colony” extracting raw information but importing finished AI solutions. Kenya’s GPU cloud is a direct countermeasure.
As one X user (@moneyacademyKE) put it this morning:
“We’re not just consumers of AI anymore. We’re building the factory.”
Next Steps:
Developers can sign up at ai.africadatacentres.com. Early access includes a 2-hour live demo with NVIDIA engineers on November 20.
For Kenya and East Africa this isn’t just a launch.
It’s a declaration of computational independence.
