It is easy to take pictures of Nigeria’s AI story: a new chatbot, a startup getting funding, a training programme claiming it will produce thousands of AI workers. These things are important, but the less visible parts of this story may be even more important for deciding who profits most from this boom, because AI doesn’t just appear on screen; it has to run on something.
For every Nigerian company designing an AI product, building a brand or owning the relationship with its customers, there’s a good chance that much of the machinery underneath that business is being rented from someone else: somebody else’s cloud, fibre, data centre and somebody else’s model. As usage grows, the cloud bill increases. If the company also depends on an external model provider, another portion of the value created by every new customer can leave the business almost as quickly as it enters.
This is where Nigeria’s recently announced National Digital Cloud Policy becomes important. The policy is not simply about housing Nigerian data on computers located inside Nigeria. It is also intended to stimulate investment in data centres, connectivity, cloud infrastructure and AI compute. The government says it wants to mobilise $250 million in private investment within 12 months and $750 million within 24 months.
Nigeria aims to combine demand from ministries, departments, and agencies rather than having each public body purchase its own cloud capacity. In essence, the idea is to use government purchasing power as “anchor demand”. The policy also wants to position Nigeria as a cloud and digital services hub for West Africa. If data centres built in Lagos can reliably serve customers in Accra, Dakar and elsewhere in the region, compute begins to move from being only an expense for Nigerian companies to becoming a service Nigeria can sell. A data centre, however, has to be paid for long before the first customer begins using it. There’s a large upfront investment in servers, cooling systems, security, fibre-optic connections and backup power. Investors are more likely to build when there’s a big customer guaranteed to purchase capacity, hence government demand could make new infrastructure easier to finance.
Now an anchor can, in fact, anchor the right thing or the wrong thing. What happens if the market grows, but the economic weight shifts even further towards a handful of operators already dominant in many countries? Even though servers may sit in Nigeria, the more profitable parts of the chain like the specialised chips, software, technical know-how and pricing power, may still rest outside the country. Hosting infrastructure in Nigeria is useful, but that doesn’t mean owning more of the economics surrounding it.
Nigeria is not operating from zero. AWS already operates a Local Zone in Lagos, while Equinix has continued expanding its Nigerian data-centre presence after entering through its acquisition of MainOne. These investments help customers seeking lower latency and resilient connections, but the presence of a data centre does not automatically make it an AI factory. AI workloads can require specialised graphics-processing hardware, while startups may focus on price, banks on security, government on sensitive data, and smaller businesses on whether the final bill remains affordable in naira.
And then there’s the cost of running these machines. Data centres rely on uninterrupted electricity, cooling and backups. Much of the technology required comes from abroad and is priced internationally, which affects infrastructure already expensive to set up when the exchange rate is unfavourable. This is compounded by an already lopsided world in data-centre capacity, as the World Bank estimates that high-income countries held 77% of global colocation capacity as of June 2025, while lower-middle-income countries held only 5%. Nigeria is unlikely to bridge this gap by producing every chip or competing in every part of the global AI stack, and frankly, it doesn’t do much good to pretend otherwise.
A more relevant question is this: where can Nigeria establish its position in that chain? Will the country engineer and maintain complex infrastructure or only remain a consumer? Can Nigerian companies expand beyond shared hosting into higher-value cloud and AI services? Can infrastructure built in Nigeria serve customers across the wider African market? Can public funding fuel competition, rather than quietly making government a guaranteed customer to established providers without knowing where the larger share of value eventually goes?
Nigeria does not necessarily need to be present in the entire AI economy to profit from the boom, but taking advantage of the areas within its reach could go a long way. The likely winners of Nigeria’s AI boom might be quite unlike the companies getting the most media attention. Some profitable businesses may never design a sleek consumer app or develop a famous new algorithm. They may simply come to own more of the infrastructure we all continue to pay to use.
