The AI skills boom is real. The market for those skills is not

Every week, another accelerator announces a new cohort. Another NGO publishes a report on the digital skills gap. Another multinational hosts a bootcamp in Lagos or Abuja, training three hundred young Nigerians to “leverage AI for economic transformation.” The photographs are always the same: bright faces, laptops open, a facilitator gesturing at a slide deck bearing the logo of a Silicon Valley sponsor. Everyone claps. Everyone posts. Everyone moves on.

And yet, if you walk into the average Nigerian SME, and ask the owner whether they have integrated any form of artificial intelligence into their operations, the honest answer, nine times out of ten, is no. Not because they lack trained staff. Not because nobody told them AI exists. But because, for most Nigerian businesses operating in the real economy, artificial intelligence is not solving a problem they are currently willing to pay to fix.
This is the crisis nobody wants to name. Nigeria has a demand problem, not merely a supply problem.

We have confused activity for progress. The supply side of the AI economy, the training programmes, the certifications, the hackathons, has expanded dramatically in recent years. Nigeria now produces thousands of young people annually who can prompt a language model, build a basic machine learning pipeline, or describe the use cases of computer vision with reasonable fluency. This is not nothing. But skills without a functioning market to absorb them are credentials, not capital. They look impressive on a CV and do very little for GDP.

The uncomfortable truth is that most Nigerian businesses do not yet have the foundational infrastructure that makes AI adoption worthwhile. AI tools do not operate in a vacuum. They require reliable electricity to power the devices that run them. They require broadband internet that does not cut out midway through a query. They require the digitisation of business processes, invoices, inventory, customer records, that can serve as the raw material for machine learning. They require the organisational maturity to implement change and the financial stability to absorb the cost of experimentation. Strip away these preconditions, and even the most AI-fluent workforce in the world becomes largely decorative.

We are, in other words, training people to drive cars on roads that have not yet been built.

The demand problem goes deeper than infrastructure. It is also cultural and structural. Nigerian businesses, particularly the SMEs that form the backbone of the non-oil economy, operate in conditions of radical uncertainty. Interest rates are punishing. The naira remains volatile. Regulatory environments shift without notice. When survival itself demands all available cognitive and financial bandwidth, the adoption of an unfamiliar technology whose returns are diffuse and long-term is a luxury that most owners rationally decline. This is not ignorance. It is sound prioritisation under constraint.

The AI evangelists will counter that this is precisely why skills training matters, that awareness precedes adoption, that education creates the appetite for change. There is partial truth in this. But awareness programmes that are not paired with concrete demand stimulation, access to affordable tools, and business models that demonstrate near-term, tangible returns are at best premature and at worst a form of development theatre: visible, photogenic, and structurally ineffective.

What would genuine demand creation look like? It would begin with the public sector, which remains the largest single actor in the Nigerian economy and whose procurement decisions shape the behaviour of entire industries. If federal and state ministries were required to use AI-assisted tools for procurement, budgeting, or service delivery, and if local firms were given a preferential pathway to supply those tools, demand would be created at scale, and the private sector would follow. This is not a novel idea. It is how industrial policy has worked in every country that has successfully built a technology sector, from South Korea to Estonia.

It would also require the financial system to lower the cost of experimentation. Venture debt and innovation grants targeted specifically at businesses piloting AI solutions, not just at startups building AI products, but at traditional businesses adopting them, would shift the risk calculus meaningfully. At present, the cost of failure for a mid-sized manufacturer trialling an AI-powered inventory system is borne entirely by the business. Socialise a portion of that risk, and adoption accelerates.

And it would require the technology community itself to develop products that meet Nigerian businesses where they are, rather than where Silicon Valley imagines them to be. A language model that functions reliably on a 3G connection, that handles Yoruba and Igbo without condescension, that integrates with the informal accounting practices of a market trader: this is harder to build than a generic chatbot, and far more valuable to the Nigerian economy. The gap between what is being built for Nigeria and what Nigeria actually needs remains wide, and it will not close simply by training more users.

None of this is to suggest that skills development is unimportant. A technically literate population is a genuine asset, and the investments being made in that direction are not wasted. But skills without demand are seeds without soil. The question Nigeria’s business community, its government, and its development partners must ask, urgently, honestly, and without the comfort of a press release, is not how do we train more people to use AI, but how do we build an economy in which there is actual reason to use it.

Until that question moves to the centre of the conversation, we will continue to produce beautiful cohort photographs and stubbornly flat adoption curves. The skills gap is real. The demand gap is larger. And we are spending most of our energy on the wrong one.