The Year Africa Set the Terms

For a generation, the terms that governed African economies were written elsewhere and read here. In 2025 that arrangement cracked, in both directions at once.

Start with the letter that arrived in Maseru. In April 2025, from a podium in Washington, the American president declared a national emergency over the trade deficit and imposed what he called reciprocal tariffs on the world. Lesotho, a mountain kingdom of roughly two million people, was assigned a rate of 50 percent, the steepest levied on any nation on earth. A country most of the world could not place on a map had been handed the highest tariff in it. The terms of access to the largest consumer market on the planet had been rewritten overnight, for an economy that did nothing to invite it.

Now set beside it a second scene from later the same year. In November, heads of state landed in Johannesburg for the first G20 summit ever held on African soil, under a South African presidency that put debt, inequality and a fairer cost of capital at the centre of the agenda. The two most powerful leaders in the room stayed away. The meeting produced a leaders declaration regardless. For once the continent was not lobbying at the edge of the table. It was setting it.

Between those two scenes runs the argument of this issue. The old model had a single author. Trade preferences, aid budgets, the price of money and the rules of the digital economy were set in foreign capitals and delivered to African governments as weather, something to be endured or exploited but never negotiated. We have argued before that markets are made by whoever sets their rules, not by whoever trades in them, and 2025 was the year that question stopped being abstract. Our December brief on the global election supercycle called this the exposure that matters most, the settings reset abroad in the same window that decide more for the continent than any single vote cast on it. 2025 tested that thesis and mostly confirmed it. But it also showed the first serious counter current in years.

The counter current is visible in the numbers the African Development Bank presented in Abidjan on 30 March 2026. The continent grew 4.2 percent in 2025, up from 3.5 percent in 2024, and outpaced the 3.1 percent world average. Twelve of the twenty fastest growing economies on earth were African. Growth topped 5 percent in twenty two of them and 7 percent in six. Behind the average sits a spread wide enough to make the single continental figure close to meaningless: Ethiopia at 9.8 percent, Rwanda at 7.5 percent, Uganda at 6.4 percent, against oil economies barely moving. There is no African economy. There are more than fifty, running on different clocks.

The counter current is visible in money as well. After a two year freeze, African sovereigns walked back into international markets. Cote d'Ivoire opened in January 2025, Benin and Kenya followed, and by year end eight countries had raised 15.7 billion dollars across fourteen eurobond issues, up 15 percent on 2024, most of it oversubscribed. The S&P Africa Sovereign Bond Index returned more than 20 percent in the year to October and spreads narrowed to their tightest since 2019. The price of money is still set abroad. But in 2025, for the first time in a while, the market came back to Africa on improving terms, not punitive ones.

And it is visible in silicon. In the same twelve months, the physical layer of artificial intelligence began landing on the continent, no longer only rented from abroad. Google opened a cloud region in Johannesburg, Microsoft and G42 committed a billion dollars to Kenya, and Cassava Technologies signed a deal with Nvidia to build Africa's first AI factory, a 700 million dollar rollout beginning with 3,000 GPUs in South Africa and reaching 12,000 across five countries. Whether this becomes ownership or merely a better class of tenancy is the open question we return to later. But the machines are now here, with African addresses.

Hold the two scenes together and the shape of the year appears. Africa was on the receiving end of the harshest terms anyone imposed in 2025, and it authored more of its own terms than in any year this century. Both are true, in the same twelve months, and the tension between them is not a contradiction to be resolved. It is the condition to be managed.

Which brings us to the question this issue is organised around, and the one we would ask of any institution reading it at the start of a decisive year. Not whether the terms are fair, because they are not and were never going to be. The question is narrower and more useful: on the decisions that will actually shape the next ten years, whose terms are you accepting by default, and which ones could you be writing yourself?