African AI Startups Worth Watching Right Now

There is a version of Africa’s AI story that gets told mostly in superlatives in a continent leapfrogging, a generation rising, a market awakening. The actual picture is more specific and more instructive.

A TechCabal Insights analysis tracking 207 AI startups across 17 African countries found that Nigeria, South Africa, and Kenya account for 63% of all tracked activity, which is where the talent pools, investor networks, and technical infrastructure tend to cluster. But the more telling shift is geographic. Egypt grew from three tracked AI startups in 2022 to eleven by 2025. Ghana, Tunisia, Rwanda, and Senegal are consolidating their own clusters. The ecosystem is broadening.

What separates the startups generating serious attention right now from earlier waves of African tech is the nature of the problems they are solving. These are not generic software platforms adapted from templates built elsewhere. The companies are drawing investor interest. Increasingly, commercial traction is working on problems that are large in scale, thin on existing solutions, and deeply specific to African contexts: infant mortality screening, underrepresented language processing, agricultural post-harvest loss, and environmental data scarcity. The AI is the mechanism. The problem comes first.

The Startups to Watch

Ubenwa is one of the more striking examples of that approach. Founded by Charles Onu, who first encountered birth asphyxia while volunteering with Enactus in Nigeria before pursuing a computer science PhD at McGill University, the startup builds AI models that analyse an infant’s cry to detect birth asphyxia — a leading cause of neonatal death. The diagnosis takes roughly ten seconds and requires nothing beyond a smartphone audio recording. Birth asphyxia causes close to one million neonatal deaths annually, according to WHO data, and Ubenwa’s clinical validation studies are running in collaboration with neonatologists in Nigeria, Canada, and Brazil. The startup is backed by partners including the WHO Africa office, MIT, and the Québec AI Institute. In a health system where specialist neonatal care is sparse and equipment costs are prohibitive, a tool that runs on existing consumer hardware addresses the constraint directly.

Lelapa AI is tackling a different gap. Co-founded by Jade Abbott and Pelonomi Moiloa and based in Johannesburg, the company builds language models for African languages that global AI platforms have largely overlooked. In 2023, Lelapa raised $2.5 million in seed funding from Mozilla Ventures, Atlantica Ventures, and angel investors, including Jeff Dean and Karim Beguir of InstaDeep. Its flagship product, InkubaLM, is Africa’s first multilingual small language model, supporting Swahili, Yoruba, isiXhosa, Hausa, and isiZulu. By mid-2025, a global compression challenge hosted by Lelapa and Zindi had reduced InkubaLM’s size by 75% without loss of performance, a meaningful milestone for deployment in low-connectivity environments. The business model is language-model-as-a-service, which positions Lelapa to sell into any enterprise that needs to serve African consumers in languages they actually speak.

Releaf is a Lagos-based agritech company that combines proprietary hardware with AI to address post-harvest losses in Nigeria’s oil palm sector, a $3 billion market dominated by an estimated four million smallholder farmers. Founded by Ikenna Nzewi and Uzoma Ayogu, the startup raised $4.2 million in seed funding in 2021 from Samurai Incubate Africa, Future Africa, and Consonance Investment Managers, and later closed a $3.3 million pre-Series A backed by several of the same investors. Its Kraken de-shelling technology and SITE geospatial application help factories identify quality raw material and optimise processing locations. Since launching, Releaf has processed more than 10 million kilograms of palm nuts and secured over $100 million in supply contracts from manufacturers including PZ Cussons and Presco. Its valuation has tripled since the seed round.

Amini, founded in 2022 by former NVIDIA executive Kate Kallot and based in Nairobi, is building infrastructure for Africa’s environmental data gap. The continent remains the world’s most data-scarce region for environmental intelligence, soil health, drought patterns, flood risk, and crop yield data are either locked in government paper files or unavailable at the resolution African farms require. Amini aggregates data from satellites, ground sensors, and weather stations, then uses AI to produce hyper-local environmental insights for agricultural insurers, FMCG companies, and governments. It raised a $4 million seed round led by Salesforce Ventures and the Female Founders Fund. Among its customers is Aon, the global insurance broker, whose collaboration with Amini and the African Development Bank has reportedly cut crop insurance management costs by up to 30%. Amini also became the first African company accepted into the Seraphim Space Accelerator.

Indicina, based in Lagos, operates in the credit infrastructure layer of Nigeria’s fintech ecosystem. The startup uses AI to generate risk assessments that help lenders evaluate borrowers who lack formal credit histories, a structural problem in a market where the majority of the population has no credit file. It has raised $3.2 million, with investors including LoftyInc Capital, Silverbacks Holdings, and Unicorn Growth Capital. The value proposition is expanding credit access while managing default risk, a harder problem than it sounds in an environment where traditional underwriting models fail by design.

What the Funding Picture Actually Shows

The capital environment for African AI remains substantially constrained. African AI startups raised a cumulative $803 million in external funding between 2019 and mid-2025, a figure smaller than many single funding rounds completed by US AI companies during the same period. Global private AI investment ran between $100 billion and $130 billion in 2024 alone. Africa’s share was a fraction of a percent.

That gap has direct operational consequences: it limits compute access, constrains hiring, and makes sustained R&D difficult. The InstaDeep acquisition by BioNTech in 2023 for approximately $682 million remains the continent’s landmark AI exit and the benchmark that others are measured against, and still an exception rather than an established pattern.

Concrete steps are being taken at the infrastructure level. At the Unstoppable Africa 2025 conference in New York, Strive Masiyiwa announced plans for Africa’s first network of AI factories powered by NVIDIA GPUs, targeting completion by end of 2026. Compute access has been one of the binding constraints for serious AI development on the continent; addressing it as an infrastructure problem rather than a startup-by-startup problem is the right framing.

Where Things Actually Stand

The ecosystem is producing technically credible companies. The founders building in health AI, language technology, and agritech are not simply applying tools built elsewhere to African problems. They are generating original datasets, designing for infrastructure constraints, and building models that would not exist without deliberate effort to serve African contexts. That is a different kind of work, and it is producing different kinds of defensibility.

What is still unresolved is whether the capital environment can support these companies to scale at the pace their market opportunities warrant. The founding momentum is real. The follow-on funding environment, at the growth stage, remains the persistent bottleneck that the next few years will test.