The Global AI Spending Boom and What It Means for Nigeria

Analysts now put projected global spending on data centres at somewhere around $3 trillion through 2028. That is not a technology budget. It is larger than the annual economic output of most countries on earth, committed to buildings full of chips.

Most Nigerians will read that and conclude, sensibly, that it has nothing to do with them. I want to explain why it does, and why the connection is not the one people expect.

The loop

Start with what makes this boom unusual. In a normal investment cycle, a company raises money from investors, buys equipment from a supplier, sells a product to customers, and pays the investors back from the proceeds. Four different parties.

In this cycle, those roles have collapsed into each other.

A chipmaker invests billions in an AI company. The AI company uses that money to buy chips from the chipmaker. The chipmaker books the sale as revenue. That revenue lifts the chipmaker's share price. The higher share price makes it easier to fund the next investment in the next AI company, which will buy more chips.

This is not hypothetical. Nvidia invested roughly $30 billion in OpenAI earlier this year and around $10 billion in Anthropic. Reporting this month indicated Nvidia was in talks over a backstop of up to $250 billion for an OpenAI data-centre campus in Ohio, and Nvidia's own shares fell more than 4% on the news. Analysts have tried to size the web of mutual commitments across the sector and arrived at figures around $745 billion.

Meanwhile, OpenAI is reported to be on track to lose something like $14 billion this year.

I am not claiming fraud, and I do not think anyone is lying. Every one of these deals may be defensible individually. But when a large share of an industry's demand is funded by its own suppliers, the revenue figures stop being an independent signal of how much customers want the product. They become a measure of how much the industry is willing to lend itself.

The awkward evidence

Here is the finding that ought to be central to this conversation and mostly is not.

A National Bureau of Economic Research working paper published in February 2026 found that around 90% of firms surveyed reported no measurable impact of AI on workplace productivity, while executives at those same firms projected productivity gains of roughly 1.4%.

Read that twice. The companies buying the technology largely cannot detect it in their own output numbers, and the people running them expect small gains rather than transformative ones.

That gap between the spending and the measured return is the entire risk. It may close — many genuinely transformative technologies took a decade to show up in productivity statistics, electricity among them. Or it may not close fast enough to justify $3 trillion of buildings, in which case a lot of capital has been committed to capacity nobody needs at prices nobody can recover.

Why this reaches Lagos

Three channels, in order of how quickly they bite.

The cost of money. This is the immediate one. The AI build-out is financed with borrowed money, and borrowing got more expensive this month: the US Federal Reserve raised rates on 16 September for the first time since 2023, with more signalled. When global capital gets nervous about a large leveraged bet, it does not reduce its exposure to that bet first. It reduces its exposure to everything it considers risky — and emerging market assets sit near the top of that list. A wobble in a data-centre trade in Texas shows up as a wider spread on a Nigerian eurobond and a harder fundraising conversation for a Lagos founder. African startups raised about $1.44 billion in the first half of 2026, an increasingly large share of it as debt. Debt is priced off global rates. The link is direct.

Outsourced work. Nigeria has built a real business in remote services — customer support, content moderation, data annotation, back-office processing, junior software work — sold to foreign clients. These are exactly the tasks that current AI systems handle most competently. If the productivity gains eventually materialise, they will materialise first in precisely the category of work Nigeria has been selling.

Infrastructure. Data centres need reliable power, cheap land and fibre. Nigeria has one of those three. So while capital floods into digital infrastructure globally, very little of it is coming here, and the constraint is not talent or ambition. It is 4,200 megawatts of national generation.

What Nigeria should actually take from this

Three things, and none of them is "build a Nigerian OpenAI."

Stop competing on tasks and start competing on judgement. The work most exposed is the work that follows a defined procedure with a defined output. The work least exposed is work requiring context, accountability and local knowledge — a person who understands why a Nigerian customer's complaint is really about a failed transfer at month-end, or why a lending model trained abroad misreads informal income. That is a different skill from transcription, and it is the one worth building.

Treat the compute gap as an infrastructure problem, not a technology problem. Nigeria will not close the gap by writing an AI policy document. It closes it by generating and distributing more electricity. Everything else is downstream.

Be honest with young people about where the ladder is breaking. The traditional path into professional work — the junior role where you do the repetitive tasks while you learn — is the rung most at risk. Anyone advising a 22-year-old to seek an entry-level role doing routine processing for a foreign client should be prepared to explain what happens to that role in three years.

And one thing in Nigeria's favour

There is an unexpected advantage in not being inside the loop.

If the AI trade turns out to be substantially overbuilt, the countries that will hurt most are the ones whose pension funds, stock indices and tax bases are heavily exposed to a handful of very large technology companies. Nigeria's exposure to that concentration is close to zero. The NGX is banks, cement, consumer goods and, shortly, a refinery. None of those is priced on a bet about artificial general intelligence.

The indirect exposure through global risk appetite is real and it is the one to watch. But the direct hit that would land on a retirement account in California is not going to land on one in Abuja.

That is not a strategy. It is not something to be proud of. But it is a real cushion, and it is worth understanding correctly rather than either panicking about a boom Nigeria is not in, or pretending the boom cannot reach it.

The honest summary: the most consequential financial story in the world right now is a circle of very large companies buying from each other on borrowed money, watched by customers who cannot yet measure what they are getting. Nigeria is outside that circle. Whether the circle holds still decides the price of Nigeria's money next year — and that, rather than any chatbot, is how this reaches your job.