94 Economists Agree AI Spreads. 32 Say Africa Won't See It.

A Financial Times poll surveyed 94 economists across Asia, the US and Europe. Fifty-four percent expected AI to deliver significant gains in many or nearly all G20 countries over the next decade.

Ask the same panel about Africa, and optimism curdles. Nearly a third warned that AI adoption may concentrate wealth among shareholders in large economies while disrupting routine cognitive jobs in vulnerable African countries. The verdict: this is not a diffusion story. It is a sorting mechanism, and the people doing the sorting are describing it out loud.

The Electricity Excuse

The bullish case leans on history. Gains can come through adoption, AI-enabled imports, knowledge spillovers and cheaper goods and services, economists said. You do not need to build frontier AI, just buy it cheap once someone else perfects it.

That logic holds for South Korea or Poland, economies with fiber, grid capacity and capital markets already built. It breaks on contact with the actual electricity analogy: electricity's diffusion took over 40 years to reach a fraction of the population that mobile phones reached within 15. The constraints were ports, wires, regulators, money. AI runs into the same wall, just with GPUs instead of copper.

Africa has 1.3 billion people and more internet users than North America's 328 million. It has barely as many data centers as Switzerland. You cannot adopt your way around a continent that must route its own data through someone else's servers to compute anything at scale.

Who Holds The Pen

A G20 spread is twenty economies, not 195 countries. It is concentration with better PR. Research on Africa's own AI productivity math backs the panel's pessimism: early winners narrow to a handful of economies that already have digital groundwork.

No economist in the FT panel is quoted calling for a fix. A survey of people whose job is forecasting consequences produced a consequence and no plan. That is what it looks like when people with leverage have already priced in the outcome.

If your portfolio, your product roadmap or your policy brief treats "AI spreads globally" as settled, the FT's own economists just told you which third of the planet that sentence excludes.

What to watch: whether the African Union or IMF push concrete AI-infrastructure financing at the next multilateral forum; which G20 nations attach actual capital, not just language, to African AI readiness; and whether China or India, both with deep African economic footprints, move first to fill the gap the US and Europe are leaving open.