Deploying 100 MW
Scaling 1 GW of compute to define the next era of AI.
The Africa Compute Fund has initiated the deployment of a 100 MW AI data center, beginning in 2026 with the Nairobi Supercluster — the first node in a broader, multi-region infrastructure rollout designed to support large-scale artificial intelligence workloads across the continent.
This will form the foundation of a distributed, GPU-optimized network built to accommodate more than 100,000 NVIDIA H100-class systems, delivering petascale compute performance to research institutions, enterprises, and emerging AI labs. The project represents the first phase in our plan to deploy 1 GW of compute capacity by 2030, establishing a scalable foundation for the region’s digital and AI infrastructure.
The Partnership
Each facility is architected as a Tier III/IV-compliant, modular data center, integrating liquid-cooled, high-density GPU architectures and renewable or hybrid power systems for optimal PUE and minimal thermal overhead. Each module functions as a prefabricated, scalable unit — a self-contained compute pod that can be deployed independently or interconnected to form megawatt-scale clusters.
The infrastructure meets hyperscale performance and redundancy standards while preserving rapid hardware refresh capability, enabling continuous adaptation to next-generation AI, HPC, and quantum workloads across distributed sovereign regions.
The broader network will extend beyond Nairobi into additional high-density zones across East, West, and Southern Africa, forming an interconnected grid of sovereign AI data centers. Each site will act as both a physical and logical node within our Monarch compute utility, enabling efficient scheduling, allocation, and monetization of GPU resources at continental scale.
This announcement follows months of collaboration with global hardware suppliers and regional partners, aligning with our broader mission to consolidate Africa’s fragmented compute landscape into a single coordinated infrastructure layer.
Initial site engineering and power allocation design have commenced, and we’re engaged in energy integration, facility construction, and advanced cooling system design. Procurement planning is aligned with global supply chains to ensure compatibility with next-generation GPU clusters and future proofing our rack configurations.
Reserve Compute
Organizations and researchers seeking access to compute can now reserve capacity through our Reservation Program, which allocates GPUs on a first-come basis ahead of the full capacity launch. A global pool of H100-class systems is already online and available for early access, providing developers, enterprises, and institutions the opportunity to begin training, fine-tuning, and deploying models within our network prior to full grid activation.
Together, Africa Compute Fund and its global partners are laying the groundwork for a new continental infrastructure paradigm, one where compute becomes a public utility powering scientific, economic, and cultural transformation.
This initiative marks a decisive shift from dependency to sovereignty: the capacity to train, host, and deploy frontier AI models from within Africa itself.
