Physical Address
4 Elgon Terrace, Kololo, Kampala, Uganda
At the Africa Soft Power Summit 2025 in Nairobi, tech giants Google and Meta called for greater AI sovereignty on the continent, urging an approach where AI systems are built with African datasets, grounded in local languages, and aligned with regional priorities.
What’s Driving the Sovereignty Move
One of the driving factors is that many global AI models have been trained primarily on English or widely spoken languages, leaving regional and indigenous tongues under-represented. Google emphasized its investment in AI development to change this — for example, its support to the Masakhane Research Foundation, which works on natural language tools for over 40 African languages. Orange, in collaboration with Meta and OpenAI, has also initiated projects to fine-tune large language models (LLMs) like LLaMA and Whisper to better understand regional languages, starting with Wolof and Pulaar in West Africa.
Infrastructure & Local Data Systems
Sovereignty is not just about languages. It’s also about where data is stored, how it’s processed, and who controls it. At the summit, Google’s representatives spoke about growing local capacity — research hubs, community centers, scholarships, and academic partnerships — to ensure AI infrastructure isn’t just imported but developed locally.Meta has similarly expressed interest in ensuring that AI solutions are culturally relevant, with local input in design, data handling, and use-cases.
Challenges & Commitments
There remain challenges: building high-quality datasets, securing enough computing power and storage locally, ensuring privacy and responsible data use, and avoiding biases in models that don’t understand African context.
To address these, Google is committing significant funding (over US$37 million) toward language inclusion, AI education, startup support, and infrastructure like its new AI Community Center in Accra. Orange’s language model work is being released as open source for non-commercial use, which helps academic, public, and nonprofit sectors.
Bottom Line
The call from Google and Meta for AI sovereignty in Africa is more than rhetoric. Already, we’re seeing tangible steps — local datasets, language models, funding, infrastructure, and policy dialogues. If these efforts scale effectively, AI in Africa can become more inclusive, more relevant, and more resilient, driven by Africans, for Africans.
