The AI Hub for Sustainable Development, working with partners across Africa, Italy and the G7, has launched the Compute Accelerator Programme to support the next generation of AI innovators solving development challenges. The goal is simple but important. Many African teams have already built strong AI solutions for health, education, agriculture, language access and industrial systems. What they lack is consistent access to compute. By giving these ventures the compute power, mentorship and partnerships they need, the programme is expected to unlock growth across sectors and show what becomes possible when access is no longer the barrier.
This six month programme running from October 2025 to March 2026 brings together two tracks. The first is the Compute Ready Track which supports twenty AI ventures that already have validated solutions and immediate compute needs. The second track focuses on infrastructure builders. Together they show the impact that compute access can have when linked to real development challenges.
The first group of twenty ventures reflects the strength and range of Africa’s AI ecosystem. They work in health, agritech, education, Earth observation, multilingual systems, enterprise intelligence, finance and sovereign AI. Their work shows how compute can unlock scale by improving model training, deployment, performance and reliability. They also provide insights into the kind of infrastructure the continent needs including supported languages, real workloads and energy efficient systems. Six of these ventures are already building infrastructure themselves creating a full cycle between innovators and the future of African compute.
Strengthening language access, public services and enterprise systems
WideBot AI (Founded by Mohamed Nabil & Mohamed Mostafa) is building an Arabic-first platform with its own 7B model AQL. The team already serves governments and large enterprises across Saudi Arabia and the wider region with voice and language technologies that support banking, healthcare and telecom services. With compute support from the programme WideBot will optimize its models and scale across Saudi Arabia, the UAE and Egypt by March 2026.
Vambo AI (Founded by Chido Dzinotyiwei and Isheanesu Misi) focuses on African language inclusion. With its Jua-Tanga model family trained with the support of more than eighty linguists, the platform provides translation, transcription, semantic search and voice-driven interfaces across forty four African languages. Through the programme Vambo AI will improve model performance, fine-tuning, inference speed and large-scale deployment with support from Cassava Technologies’ NVIDIA GPU clusters and global cloud partners.
Transat Management Company (TMC) (Founded by Mini Kouame) brings AI to SME finance by matching verified African businesses with global investors through NLP, predictive analytics and a human-in-the-loop engine. The company is seeking partnerships with DFIs and institutional investors across the UNDP network as it strengthens multilingual modelling and scaling across Africa. Compute support will improve risk scoring, deal engine performance and data processing for diverse markets.
SCADAians is transforming Egypt’s energy sector. The company builds AI-powered analytics and monitoring tools for the country’s fuel supply network using locally hosted GPUs. As part of the programme it will expand into a national-scale AI platform in line with Egypt’s Vision 2030 improving logistics, fraud detection and demand forecasting.
PAWA AI (Founded by Winnie Mang’eni, Innocent Charles and Dr. Michael Mollel) focuses on small language models that run on edge devices, private clouds and sovereign cloud environments. The no-code platform helps governments and schools build localized models for learning, agriculture and public services. Through the programme PAWA AI will train multilingual models with GPU resources while advancing ethical and energy-efficient AI systems for low-resource environments.
PCS AGRI (Founded by Tahar Hamdani) brings AI to seed quality and agricultural productivity through computer vision, deep learning and IoT. The Track Seeds platform automates germination counting, seedling classification and traceability. The company also supports tomato yield analytics and pest detection. Compute access will support large scale deployment and improved real-time decision support for farmers across Morocco, Spain, Turkey and the Netherlands.
NextAV (Founded by Hichem Mokni) is building AI super-resolution for public satellite imagery improving Sentinel and Landsat data by up to sixty times. This work opens new use cases in energy, climate and infrastructure monitoring and depends on high compute to scale across new regions and models.
Muna Kalati (formerly Muna Kalati) is advancing Muna TV’s AI recommendation engine for African children’s content now in live deployment applying AI to transform learning and culture. Founded by Christian Elongue and a Ghana based team the company addresses a critical gap: the cultural disconnect in children’s media. By optimizing its NLP models, subtitling and dubbing tools through the accelerator the company plans to expand access to local language content across the continent while improving real-time inference for low-bandwidth conditions.
Jacaranda Health uses AI to support pregnant and postpartum mothers through PROMPTS and its AI assistant UlizaMama. The system handles more than 12000 daily questions across Swahili, English and Sheng. With support from the programme Jacaranda will optimize Twi voice models for expansion in Ghana reduce training costs and strengthen its multilingual AI pipeline for maternal health.
Hexastack is developing an open-source framework for multilingual conversational AI. With a visual flow builder and multi-channel integration it helps enterprises deploy privacy-first agents across WhatsApp, Facebook and web platforms. Compute support will improve performance in Arabic, French and African languages.
Hasab AI (Founded by Kidus Yared) is building Africa’s audio intelligence stack. The platform converts conversations in Amharic, Afaan Oromoo and Tigrigna into structured insights for enterprises. With GPU support the team will optimize ASR and predictive models and expand infrastructure to reach more sectors across Ethiopia.
Crane AI (Founded by Ugandan engineers Kato Steven Mubiru and Bakunga Bronson) Labs is advancing sovereign, offline-first AI for governments and NGOs working in health, education and agriculture. The company builds on-device inference tools that do not rely on cloud connectivity and integrates a governance benchmark recognized by the UK Government. Through the programme Crane AI Labs will work with governments and development banks to design national-scale digital public infrastructure supported by Africa-based compute.
DeepLeaf (Founded by El Mahdi Aboulmanadel) brings AI to crop diagnostics with 96% accuracy and the ability to process ten thousand images per hour even offline. With GPU support DeepLeaf plans to strengthen model training cycles and expand partnerships in national agriculture programmes across Africa and the Middle East.
Artificial Intelligence Centre of Excellence Africa (Founded by John Kamara) focuses on sovereign enterprise intelligence through its DataInViz platform. The system connects fragmented data sources and produces real-time insights for institutions in health, agriculture and finance. Compute support will expand model performance and improve the processing engine Moesha for large-scale government deployments.
Armstrong Education (Founded by Ali Wahba and Ashraf Darwish) is advancing AI-powered STEAM and robotics education across the MENA region. Its platform personalizes learning for children aged four to eighteen. The programme will help Armstrong improve its adaptive learning engine and expand partnerships with schools and development agencies.
ADEX Technology (Founded by Djaoued Allal) is scaling its sovereign AI automation platform ADEXGENIE.ai. With models built for Arabic, French and African dialects the company focuses on workflow automation for regulated industries. Compute support will enhance performance inside its sovereign cloud infrastructure ADEXCLOUD.dz.
Udu Technologies (Founded by Alexander Tsado) provides affordable GPU access through its GPUaaS platform and Udu-Rayda marketplace. Operating across Nigeria, Ghana, Kenya and South Africa the company supports researchers, startups and governments. As part of both tracks UduTech will expand infrastructure while showcasing how local compute unlocks innovation.
Horus Labs (Founded Raymond U. Ononiwu alongside co-founders Ludovic Bernad and Dennis) builds modular, renewable-powered edge data centers for AI workloads. With partnerships with Schneider Electric and others the company delivers Tier III ready data centers tailored to African climates and needs. The accelerator will support Horus Labs as it expands its infrastructure into more countries.
A turning point for Africa’s AI ecosystem
The Compute Accelerator Programme shows that Africa already has the talent, the ideas and the systems to build world-class AI solutions. What has been missing is access to compute at the right cost and scale. This programme directly addresses that gap. It strengthens innovators, supports infrastructure builders and creates new pathways for partnerships between governments, universities, donors and private sector leaders.
The ventures in the Compute Ready Track represent a new chapter for African AI. They are building multilingual language systems, AI-powered health tools, sovereign cloud infrastructure, Earth observation models, agritech solutions and next-generation education platforms. With the right compute and the right network their impact will stretch across the continent.
The programme is more than an investment in technology. It is an investment in inclusive development driven by African innovators who understand local needs and are ready to build for the future.
