Africa Deep Tech Foundation Launches Laptop LLM Challenge
The Africa Deep Tech Foundation has opened applications for the Africa Deep Tech Challenge 2026, calling on engineers, researchers and startups to build AI language models that run fully offline on standard laptops. The competition offers over $20,000 in prizes and support to help winning teams grow
The Africa Deep Tech Foundation has opened applications for the Africa Deep Tech Challenge (ADTC) 2026, introducing this year's theme, "The Laptop LLM Challenge." The initiative is calling on engineers, researchers, students, startups and AI builders across Africa to develop powerful artificial intelligence that runs entirely offline on the laptops many people already own.
Unlike many popular AI tools that depend on cloud servers and constant internet access, every submission must work completely offline on a standard laptop with 8GB RAM. The goal is to make AI more accessible, affordable and practical for communities where internet access, cloud computing costs and high-performance hardware remain major barriers.
The challenge comes as interest in generative AI continues to grow across Africa. While developers are building new AI applications in healthcare, finance, education and agriculture, many of these solutions still depend on expensive cloud infrastructure. Through ADTC 2026, the foundation hopes to encourage the development of AI models that people can use anywhere, even without an internet connection.
Applications are open to individual developers and teams of up to three members, including engineers, AI researchers, university students, startups and innovators building solutions for African challenges.
Participants are expected to develop a Large Language Model (LLM) application that runs fully offline while solving real-world problems. Projects can focus on sectors such as healthcare, agriculture, education, coding assistants, enterprise productivity, scientific research and creative writing. The organisers are also encouraging developers to build AI systems that support African languages, with additional recognition for meaningful localisation.
Beyond producing accurate responses, the competition places strong emphasis on efficiency. Every entry will be tested on a standard laptop equipped with an Intel Core i5 processor, integrated graphics and 8GB RAM running Ubuntu Linux. During evaluation, models must remain within a 7GB memory limit, operate without internet access, respond efficiently and avoid excessive system resource usage. Entries that exceed the memory limit or fail to meet the offline requirement will be disqualified.
Submissions will be judged on several factors, including offline performance, memory efficiency, speed, technical innovation, usability and real-world impact. Judges will also consider how well projects address African challenges and support local languages where applicable.
The competition offers a prize package worth more than $20,000. The overall winner will receive $8,000 in cash alongside a six-month scale-up programme designed to help transform promising ideas into viable products. The programme includes mentorship, pilot opportunities, legal support and exposure to investors. Second place will receive $4,000, while the Best Integration Award carries a $3,000 prize and the Best Localisation Award offers $1,500. Finalists and semifinalists will also receive GPU credits to support further AI development.
Applications are now open through the Africa Deep Tech Foundation's official challenge portal. Applicants will be required to submit information about themselves or their team, describe their proposed solution, explain the technical approach and outline the expected impact of their project.
The first application stage, known as Gate 1, closes on August 25, 2026. Shortlisted teams will proceed to technical audits before presenting their projects at a live final in October, where winners will be selected.
The challenge arrives at a time when advances in model compression, quantisation and efficient inference are making it possible to run capable AI models on ordinary computers. For Africa, where connectivity and computing resources remain uneven, offline AI has the potential to expand access to education, business tools, healthcare support and digital services.
Rather than encouraging developers to build another cloud-based chatbot, the Africa Deep Tech Challenge is pushing innovators to rethink how AI should be built for African realities.
The solutions produced through this competition could help shape a future in which advanced AI runs on everyday laptops, bringing powerful technology closer to the people who need it most.