A new IMF commentary says artificial intelligence could add about half a percentage point to global growth each year, but warns the gains may bypass low-income countries unless they fix power, skills and financing gaps.
Money is pouring into artificial intelligence (AI) at a scale that may eventually dwarf what the world spent building railways, electricity grids and the internet. But a commentary published by the International Monetary Fund (IMF), the global lender that advises countries on economic policy, cautions that the benefits are not being shared evenly — and that the growth boost in rich countries could be more than double what low-income countries see.
The piece, carried by Norvan Reports, estimates that AI could lift global growth by roughly half a percentage point a year over the next decade. That gain, however, depends on firms genuinely redesigning their products, processes and workflows rather than simply buying the technology. The IMF draws a comparison with the internet: the real returns came from new business models, not from putting up a webpage.
Where the risk sits
Much of today's spending is going into AI infrastructure, especially data centres built by so-called hyperscalers — the very large cloud and technology companies. The IMF notes these firms are increasingly borrowing to fund the build-out. Their share of investment-grade bond issuance has jumped from under 1 percent to about 10 percent in just two years.
The Fund also flags that deal-making in the sector is often circular, with hyperscalers investing in AI model developers who then buy computing capacity back from them. If earnings disappoint, or a serious cyber incident shakes confidence, a market correction could spread through that leveraged chain — and, because AI assets are held across borders, the losses would not stay in one country.
Jobs and the skills question
IMF research cited in the piece puts global job exposure to AI at about 40 percent, rising to roughly 60 percent in advanced economies. Exposure does not automatically mean job losses; some roles will be strengthened by the technology while others are replaced. The Fund says there is no clear impact on employment so far, but the nature of work is already shifting.
In Asia-Pacific, employer demand in highly exposed occupations has fallen by about 10 percent, even as demand grows for workers who can use AI tools. New roles such as prompt engineering have emerged within a few years.
What it means for Africa
The commentary is direct about Sub-Saharan Africa: attracting private money for AI infrastructure will require a better business climate alongside targeted government subsidies and guarantees. Advanced economies are three times better equipped in digital infrastructure than low-income countries, based on the IMF's AI Preparedness Index.
Power is the other constraint. The International Energy Agency expects data centres to consume about 3 percent of global electricity by 2030 — more than Japan uses today. The IMF urges countries to build medium-term plans to raise reliable power supply, invest in skills, and modernise tax and insolvency rules so firms can adapt.
