Monitor, an annual report, notes that over the most recent decade for which it has data, “fruits and nuts” became the continent’s largest agricultural export, overtaking cocoa. The shares of traditional cash crops like cocoa, cotton, tobacco and sugar all fell. Africa still lags behind Latin America, which exports four times as much in agricultural goods (an average of $341bn per year in 2019-23 compared with Africa’s $81bn). Peru, which barely exported blueberries in 2010, was the world’s largest exporter a decade later. But if Africa can learn from such success stories, there will be fruitful returns. “Blueberries are a great example of agriculture as high-tech industry,” argues Daniel Hulls, the boss of AgDevCo, an investment fund. Driscoll’s, an American agribusiness giant, supplies seeds to farmers in Zimbabwe that are the result of years of cross-pollination. Polyethylene covers protect bushes. Storage rooms have optical sorting machines. To maximise profits, firms time production for the summer, after American supply dwindles but before Peru’s picks up. Blueberries are not the only example of Africa’s “industrialisation of fresh”. Today Kenya and Ethiopia account for 6% and 2% of global cut-flower exports. Their farms feature machines that spray fine mist, purify water by reverse osmosis and point radium bulbs at petals to prettify them. Africa’s exports of tropical fruits, such as avocados, mangoes and pineapples, have more than tripled over the past decade, a faster growth rate than in any other region, according to the Food and Agriculture Organisation, a UN agency. Last year Europe consumed more than 1m tonnes (roughly the weight of ten aircraft carriers) of avocados for the first time. Much of its demand was met by Morocco and Kenya.
In May, for the first time, South Africa replaced Spain as the world’s largest exporter of citrus fruit. Since overtaking wine in 2010, citrus has been the country’s most valuable agricultural export, and recently has earned it more than it gains through diamond exports (see chart). The success can be traced back to 1997 and the abolition of state-run marketing boards. Before then the government divided export revenues among farmers; thereafter each farmer kept their own profits and had more incentive to invest in their fruit. Industry experts also stress the role of the Citrus Growers Association (CGA), which takes levies from farmers to carry out research, lobby for better policies and disseminate know-how. The special benefits some economists ascribe to manufacturing seem to apply to tech-savvy farms. Their exports boost trade balances. They employ hundreds of people. They help create other products and firms along supply chains. “These are integrated manufacturing and logistics businesses,” says Wandile Sihlobo, a South African agricultural economist.
Jane Maina, the boss of Vert, a tropical-fruit firm in Kenya, says she has moved into exporting dried snacks. The new range means fewer mangoes and pineapples sourced from smallholders are wasted, raising farmers’ productivity. The cold-storage facilities built at east African airports for the flower business are now used by exporters of herbs, beans and vegetables. The blueberry rush in Zimbabwe has led to new “agritech” apps monitoring weather conditions and yields. In South Africa the citrus boom has produced several spin-off research firms. Sadly, bad roads and clogged ports still make it hard to deliver perfect fruit. African farmers, like many elsewhere, complain about inadequate political support. Some have a point. When Ethiopia’s flower business started, the government at the time provided armed guards for airport-bound lorries. Now exporters complain of being stopped by armed bandits. The good news is that potential markets are opening. In May China stopped applying tariffs to most African imports. In July the first punnets of blueberries from Zimbabwe were sent there. Yet non-tariff barriers, such as food-safety and plant-health regulations, often prove trickiest. Managing them needs concerted diplomacy and dedicated industry bodies, such as South Africa’s CGA.
The success of Zimbabwean blueberries, Kenyan flowers and South African oranges suggests there is huge potential for African agriculture to play the sort of role long envisioned for manufacturing. But to maximise the fruits of their labours African policymakers will need to help ensure that their firms are globally competitive. Time for fresh thinking. ■ Sign up to the Analysing Africa, a weekly newsletter that keeps you in the loop about the world’s youngest—and least understood—continent. This article was downloaded by zlibrary from https://www.economist.com/middle-east-and-africa/2026/08/02/forget-gold-and-copper- africas-latest-boom-is-in-fruit
Middle East & Africa | Lean and local An African vision of artificial intelligence Tech pioneers favour models suited to specific tasks that use little compute Aug 6th 2026 What would African artificial intelligence (AI) look like? Olubayo Adekanmbi, a Nigerian entrepreneur, reels off ideas: as fluent in African languages as in English; as comfortable with speech as with text; and lean, so it can flourish where chips are scarce and internet patchy. By contrast, frontier labs in Silicon Valley train gargantuan models on English-heavy datasets. “The current gameplay for AI globally is completely un-African,” he says. Which is why two years ago, after a career in telecoms, Mr Adekanmbi co- founded EqualyzAI. The startup has designed tools to advise mothers on health, help traders with accounts and handle customer calls for banks, in
English and four Nigerian languages. Similar outfits are springing up all over the continent. They might seem redundant, given the spread of frontier models. Microsoft estimates that among people of working age 8% of Ethiopians, 9% of Congolese and 23% of South Africans have used generative AI. OpenAI says the number of ChatGPT users in Africa is growing faster than in any other region, albeit from a low base. Chinese models are also popular. But even the best large language models (LLMs) falter in African languages, typically scoring 10-30 percentage points lower on benchmarking tests than they do in English. They struggle with dialects and “code-switching”, such as when speakers skip between English and a mother tongue. For smaller languages, much of their training is on Bible verses. Asmelash Teka Hadgu of Lesan AI, which works on machine translation and speech recognition for languages from the Horn of Africa, argues that existing models are “broken” for everyday speech. That is especially true for languages like Amharic and Tigrinya, which use non-Latin scripts; in other major tongues gains have come faster. The root problem is lack of data. Common Crawl, a database used to train AI models, scraped 872m English web pages in July, but just 80,000 in Amharic. Across African languages, only Swahili was better represented than Welsh. High-quality audio for speech recognition is also scarce. African researchers are therefore recording how people talk, with backing from the likes of Google and the Gates Foundation. Audace Niyonkuru of Digital Umuganda, in Rwanda, says his firm strives to get a balance of ages, genders and dialects. What then? African researchers use new datasets to fine-tune open-weight models (whose training parameters are public), which uses less compute than building from scratch. Jacaranda Health, a Kenyan NGO, trained Meta’s Llama2 in Swahili, so it could advise on maternal health. But even after fine-tuning, such models are often worse at general tasks than a proprietary model.
Another focus for African innovation is so-called “small language models” for specific tasks, which have at most a few billion parameters (frontier LLMs have trillions). Lelapa AI, a South African lab, named its 400m- parameter model InkubaLM, after the isiZulu word for dung beetle, which can move many times its own weight. Smaller models can be run locally, or even offline. That keeps sensitive data on the continent. EqualyzAI has built a local model for the government of Kano state, in northern Nigeria, which allows officials to query documents and recordings of meetings. “If you want sovereignty, keep it small,” advises Mr Adekanmbi. All this amounts to a somewhat counter-cultural vision of AI, prizing modesty over scale and context over omniscience. It may not survive the all- conquering ambitions of Silicon Valley. But for now African researchers have found a niche. ■ Sign up to the Analysing Africa, a weekly newsletter that keeps you in the loop about the world’s youngest—and least understood—continent. This article was downloaded by zlibrary from https://www.economist.com/middle-east-and-africa/2026/08/06/an-african-vision-of- artificial-intelligence
Middle East & Africa | Of sea walls and motorways Nigeria’s builder-in-chief A friend of the president is racking up infrastructure contracts Aug 6th 2026 During flash floods in Lagos last month the water crept up to car doors and poured into homes, disrupting life in Nigeria’s commercial capital for days. Climate change is the main culprit for worsening flooding. But in poorer parts of Lagos some residents wonder if other factors could be exacerbating the damage. They worry that the construction of a coastal motorway and a sea wall, built to protect a new luxury waterfront district, may be eroding the seashore, worsening the impact of floods along the coast. Yet residents may find it hard to get information about the impact of either project. Both are part of a $20bn infrastructure portfolio overseen by firms controlled by Gilbert Chagoury (pictured), a Lebanese-Nigerian billionaire