How CircuitNotion Is Building African AI That Outperforms Western Giants

CircuitNotion, founded in 2024 by Ntirushwa Jean Marc, claims a 35 percent market share among AI companies operating in Rwanda, making it one among the top country's premier artificial intelligence infrastructure provider. Operating from Kigali, the company has built Incuti AI, described as one of the best performing Kinyarwanda chatbots.

Incuti AI: African Language Models

Incuti AI is CircuitNotion's flagship African Large Language Model designed to understand African languages, culture, and context. According to founder Ntirushwa, the platform's models are trained specifically on African context and African languages, addressing a significant gap in global AI development.

The platform powers intelligent business assistants, AI customer support automation, voice AI for call centers, government and enterprise knowledge systems, and education tutoring systems. Incuti AI is accessible through API calls and a chat interface.

The Niche Strategy: Smaller, Specialized Models

When asked about competing with giants like OpenAI and Google, Ntirushwa explains that CircuitNotion has adopted a specialized approach rather than building general-purpose models. "Recent research showed that small AI models that are specialized in one particular niche outperform those giant models which are trained on huge data because they lack clarity," he states.

The company's model focuses specifically on Kinyarwanda and other African languages, with fewer parameters than general models. This specialization enables CircuitNotion to price its services at 50 percent of GPT model costs, as fewer parameters require less GPU power to generate tokens. Despite having fewer datasets in African languages, Ntirushwa notes these specialized models outperform general models at specific African tasks.

Facing Infrastructure and Cultural Challenges

Ntirushwa identifies five major challenges facing African AI developers: GPU server access, expensive internet infrastructure, high electrical power costs, dataset scarcity for African languages, and what he calls a "continental confidence gap."

What Should Change

To address this confidence gap, Ntirushwa proposes several solutions. He believes governments should play a key role in adopting and advocating for domestic products, citing the Chinese system as an example. He emphasizes teaching local AI systems and products in schools, where mindsets can be more easily changed, and genuine investment in research beyond political rhetoric.

Why Build in Rwanda

I was curious about why he chose Rwanda specifically; he explains the strategic advantage of developing AI models in Rwanda: "We are solving real African problems directly from the ground, using local context and data that global companies don't have access to. Being based here means we can build AI that understands African languages like Kinyarwanda and Swahili, supports African accents, and fits our business environments, culture, and infrastructure."

Business Model and Investment

Most startups burn cash with and build up a bigger user base, hence spending years in development, but no profits; and because of this, I asked whether CircuitNotion is profitable.

To respond, he said that CircuitNotion is profitable, he also stated, "If I'm not profiting I would have already closed. Of course profit is there." The company raised Series A funding in 2025, and is actively raising more money to expand operations to East Africa. CircuitNotion also operates an IoT monitoring platform for smart devices and connected infrastructure across Rwanda, diversifying beyond pure AI applications.

Aligning with Rwanda's AI Vision

Rwanda is investing heavily into becoming Africa's AI hub; presenting massive opportunity to companies like CircuitNotion. In 2023, Rwanda also approved The Rwanda's National Artificial Intelligence Policy focused on AI literacy, infrastructure, adoption and more

CircuitNotion's work aligns with the policy. The policy emphasizes long-term investment in human capital beginning at primary education, with school curricula adapting for the age of AI and digital technologies.

Part of Broader African AI Movement

CircuitNotion is not alone in developing African language models. Lelapa AI from South Africa launched InkubaLM, supporting Swahili, Yoruba, IsiXhosa, Hausa, and isiZulu. Additionally, French telecom Orange announced a 2025 project where OpenAI and Meta will collaborate to fine-tune Whisper and Llama models to translate regional African languages.

Without going too far, it is worth noting that, in Rwanda, there are other companies like Digital Umuganda who have models in use. We can give an example of Mbaza Chatbot that was used during COVID-19 period and is still used today.

Building Africa's AI Future

Ntirushwa expresses optimism about the future: "I'm super happy that Africans will not be the user in future but also the builder and researchers who are not left behind." By building African-centric AI infrastructure from Rwanda, CircuitNotion addresses both technical challenges and questions about technological sovereignty and cultural representation in artificial intelligence.

Next in Series: Stay tuned for our next article exploring other Rwandan companies contributing to the global AI race. Share your suggestions at our contact page.