The global conversation around Generative AI often feels like it’s happening in a vacuum of high-end Silicon Valley labs. We talk about the magic of Large Language Models (LLMs) and the creative prowess of image generators, but we rarely talk about the “bricks and mortar” of the digital age: the physical infrastructure that makes this magic possible. For Nigeria, a nation brimming with a young, tech-savvy population and a burgeoning startup scene, the promise of AI is immense. However, to move from being consumers of AI to creators and masters of it, Nigeria must urgently address a foundational triad: Power, Data, and Compute.
The energy paradox: Powering the intelligence
Generative AI is hungry—not just for data, but for electricity. Training a single large-scale model can consume as much energy as hundreds of homes use in a year. In Nigeria, where the national grid frequently teeters between 4,000 and 6,000 MW for a population of over 200 million, the energy gap is the first and most formidable hurdle.
To unlock AI, Nigeria cannot rely solely on a fragile central grid. The solution lies in decentralised, renewable energy clusters. Imagine “AI Special Economic Zones” powered by dedicated solar farms and natural gas turbines. By co-locating data centres with independent power providers (IPPs), the country can bypass the instability of the national grid, ensuring the 99.9 percent uptime required for high-performance computing. Without stable power, the “intelligence” we seek to build will remain perpetually in the dark.
The sovereign data mine
If power is the blood, data is the soul of Generative AI. Currently, most AI models are trained on Western datasets, which often leads to a “cultural hallucination” where the AI fails to understand Nigerian nuances, languages, or socio-economic contexts.
For Nigeria to build AI that truly serves its people —whether in diagnosing tropical diseases or automating credit scoring for market traders— it needs sovereign data infrastructure. This means:
- Digitising Local Knowledge: Converting oral histories, local languages (Yoruba, Igbo, Hausa, and more), and legal archives into machine-readable formats.
- Localised Hosting: Encouraging the growth of domestic Tier III and Tier IV data centres so that Nigerian data stays within Nigerian borders, reducing latency and ensuring data privacy.
- Open Data Initiatives: Government-led frameworks that make non-sensitive public data available to local researchers and startups.
When we host our data locally, we don’t just improve speed; we claim ownership of our digital narrative.
Compute: The engines of innovation
The most visceral bottleneck for any Nigerian AI developer today is “compute.” To train or even fine-tune a Generative AI model, you need specialised hardware — specifically Graphics Processing Units (GPUs) like the NVIDIA H100s. Currently, a single high-end GPU can cost more than a luxury SUV, placing them out of reach for most local innovators.
Nigeria’s path to compute parity isn’t necessarily about every startup owning its own “supercomputer.” Instead, the focus should be on Shared Compute Resources. The government’s National AI Strategy, which includes plans for a “National AI Compute Centre,” is a step in the right direction. By providing “Compute-as-a-Service,” the state can lower the barrier to entry, allowing a student in Kano or a developer in Enugu to access the raw processing power needed to build world-class models without the prohibitive upfront costs.
Nigeria is not starting from zero. With over 120 active AI startups and a government that has recently inaugurated a National AI Trust, the momentum is palpable. However, the transition from “AI-ready” in policy to “AI-active” in practice requires a massive shift in infrastructure investment.
The “Leapfrog” theory suggests that Africa can skip older technologies and jump straight to the latest. But you cannot leapfrog physics. Generative AI requires physical cooling, physical wires, and physical electrons.
business a.m. commits to publishing a diversity of views, opinions and comments. It, therefore, welcomes your reaction to this and any of our articles via email: comment@businessamlive.com
Olusegun Afolabi has a first degree in biochemistry from the University of Ilorin, Nigeria, and a master’s in computer science from Hertfordshire University in the United Kingdom. He is an AWS solutions architect professional, a Microsoft certified Azure solutions architect expert, co-founder and chief innovations architect of Face Technologies UK Limited. He can be reached at … and on Linkedin: https://www.linkedin.com/in/olusegun-afolabi-307931184/
When women own the room — and the system