Twenty-six projects have been selected under LINGUA Africa, covering more than 50 African languages, dialects and sign languages spoken by over 500 million people across 47 countries — with each awardee eligible for up to $250,000 in cash and $400,000 in compute credits.
The initiative brings together the Microsoft AI for Good Lab, the Gates Foundation, the Masakhane African Languages Hub and Google.org, supporting African researchers, startups, universities, nonprofits and community organisations building language resources and applications. It follows an open call that closed on 15 June.
The selected languages span widely spoken ones including Kiswahili, Hausa, Amharic, Wolof and Luganda, alongside far less digitally represented languages such as Awngi, Tjwao, Doma, Medumba and Ghomala’.
Built around use cases, not datasets alone
What separates LINGUA Africa from most African language-AI funding is that projects are tied to specific sectors rather than general corpus-building.
Healthcare is the largest focus. The World Health Organization’s Kenya-based AFYA-LINGUA project will build a community-grounded African-language layer for health covering Kiswahili, Hausa, Yorùbá, Lingala, Amharic, isiZulu, Chichewa, Wolof, Kinyarwanda and Crioulo. Makerere University in Uganda will build a maternal and child health speech dataset covering Luganda, Acholi, Dholuo, Dinka and Juba Arabic. Crails will develop an open Luganda medical speech corpus and foundational models for clinical AI, while projects in Sierra Leone and Rwanda address AI-powered telemedicine and community health worker tools.
In agriculture, the Mohamed bin Zayed University of Artificial Intelligence will build a multimodal vision-language advisory system across 10 African languages including Hausa, Yoruba, Igbo, Twi, Wolof, Kiswahili, Amharic, Afaan Oromo, Luganda and Kinyarwanda. Kenya’s AntuGrow will develop SemaShambani, a voice-first agricultural advisory and credit-literacy platform in Kiswahili, Gĩkũyũ, Kikamba, Dholuo, Luhya and Kalenjin. The Alliance for a Green Revolution in Africa will build MlimiVoice Malawi for smallholder farmers in Chichewa and Tumbuka.
“AI only delivers value when people can actually use it, and language is the bridge,” said Inbal Becker-Reshef, managing director of the AI for Good Lab.
Kenya features heavily
Beyond the WHO and AntuGrow projects, Algedi Intelligence Labs was selected for RADA, covering Kiswahili, Gĩkũyũ, Kamba, Dholuo, Maa, Kalenjin and Somali. Ushahidi will develop open language resources for health and education in Tonga, Tjwao and Doma.
Dialects, sign language and the long tail
Several projects target gaps that even African language-AI work has tended to skip.
HausaNLP will build DialectBridge, focusing on Hausa dialects spoken in Kano, Sokoto, Zaria and Katsina — dialect-level variation being a layer below where most language modelling operates. The Bantu Language Initiative will develop open voice infrastructure for Lingala, Kituba and Kikongo in the Congo Basin. Arusha Technical College in Tanzania will build TSL-Open, a community-led resource for Tanzanian Sign Language.
Sign language is among the most neglected categories in African AI. The iHUB and Mastercard Foundation EdTech cohort selected this month includes Signvrse working in the same space in Kenya, but funded resources remain scarce relative to need.
Other projects cover Awngi in Ethiopia and Ghomala’, Medumba and Fe’efe’e in Cameroon.
Where it sits in a crowded field
African language AI has become one of the continent’s most heavily funded research areas, and LINGUA Africa now joins several parallel efforts.
The Gates Foundation-backed African Next Voices dataset recorded 9,000 hours of speech across 18 languages on a $2.2 million grant. Google’s WAXAL covers 21 sub-Saharan languages. The GSMA and Pleias released CommonLingua in April, an open-source model identifying 61 African languages — addressing the foundational problem that African-language text is routinely mislabelled as English or French. Lelapa AI’s InkubaLM serves five languages on modest hardware. Two weeks ago, the GSMA launched ATLAS Umoja AI with five governments, built on Nigeria’s open-source N-ATLAS model.
Each attacks a different layer. ATLAS Umoja adds a governmental tier; African Next Voices and WAXAL build raw speech and text resources; CommonLingua handles identification. LINGUA Africa’s contribution is the applied layer — funding the specific tools that turn those resources into a health worker’s transcription system or a farmer’s voice advisory.
Whether that produces deployed services rather than well-documented pilots is the test. The Masakhane Hub’s stated goal is to reach one billion Africans with locally relevant AI tools by 2029.
