South Africa may lose top AI talent to global markets due to weak local demand for homegrown technology, warns Dr. Vukosi Marivate, professor of computer science at the University of Pretoria and co-founder of Lelapa AI.

Despite the country’s ranking among the top 13 globally in AI compute capacity, Marivate says South Africa lacks the deep-tech ecosystem needed to retain AI researchers and developers.

“The main issue is the lack of local demand for deep-tech solutions built from the ground up,” he told MyBroadband. “Many organisations opt to import technology, pushing our best talent abroad.”

Lelapa AI, based in Johannesburg, developed InkubaLM-0.4B, a small language model trained on African languages including isiZulu, isiXhosa, Swahili, Hausa, and Yoruba. But Marivate said the lack of local investment has slowed progress.

He emphasized the importance of local teams creating AI tools suited to African languages and contexts, which are often overlooked in global models from companies like OpenAI and Google.

Other startups tackling similar gaps include Enlabler, a youth-focused data labelling firm supporting AI training, and Community Wolf, a platform that uses a multilingual chatbot to report suspicious activity to authorities and private security.

While local innovation shows promise, experts say South Africa must build real market demand and invest in homegrown solutions to stop the AI brain drain and foster a self-sustaining tech ecosystem.