A platform designed for Mauritian schools could soon be assessed in South Africa. For Mauritius Telecom, the stakes go beyond selling an educational tool: the aim is to circulate AI services developed on the continent, while drawing on computing infrastructure that remains highly concentrated.
A partnership between Mauritius and South Africa
Mauritius occupies a distinctive place in regional exchanges. The island is a member of the Southern African Development Community (SADC), which brings together countries in southern Africa and the Indian Ocean. South Africa, the first destination being considered for the platform, is also a member. This regional framework facilitates exchanges, but does not guarantee that the tools will be adapted or adopted in schools.
On 18 November 2026, Mauritius Telecom and Cassava Technologies signed a memorandum of understanding on the sidelines of the Africa Tech Festival. The agreement provides for an assessment of two avenues: possible access for the Mauritian operator to Cassava’s computing capacity and AI models, and potential opportunities for myt Learn in South Africa’s public and private education sectors. Expansion into other African markets is also being considered.
This is an exploratory agreement, not an announcement that a commercial rollout has already been decided. The two companies still need to assess requirements, the terms of access to infrastructure and the platform’s suitability for the targeted school systems. No launch timetable or investment amount has been announced in the information available.
This caution matters. It distinguishes an intention to partner from a service actually deployed in classrooms. Details reported by TelecomTV are corroborated by Mauritian outlet Défi Media: both partners describe the text as a memorandum of understanding and say they will assess opportunities in South Africa.
An educational platform to adapt
Developed by Mauritius Telecom, myt Learn is used in some year groups at schools in Mauritius. It offers lessons aligned with school curricula, a multilingual AI tutor and personalised revision exercises. Tools for teachers can also help them prepare lessons, create assessments and mark work.
These functions are not enough to ensure a direct transfer to South Africa. The country has several languages, varied school contexts and its own curricula. The platform would therefore need to be adapted to local educational content, teachers’ needs and the practical conditions for accessing devices and connectivity. Information published by the partners indicates that this adaptation is among the possibilities to be assessed, not that it has already been carried out.
For pupils, the potential benefit lies in supplementary support and personalised exercises. For teachers, the challenge is to have useful tools without transferring pedagogical responsibility to an automated system. Content quality, protection of children’s data and the ability to verify generated answers will be crucial before any rollout.
Computing power: the material prerequisite for AI
AI services require significant computing capacity. Mauritius Telecom already operates myt AI Cloud in Mauritius. The memorandum opens the way to a study of access to Cassava’s AI Factory in Cape Town, equipped with NVIDIA technologies, as well as to models available through CAIMEx, the company’s AI platform. Subject to arrangements yet to be defined, these resources could supplement Mauritian capacity for certain uses.
The continental context helps explain this approach. A 2026 report by the United Nations Economic Commission for Africa estimates that the continent had about 249 data centres, including 61 in South Africa, 25 in Nigeria and 19 in Kenya. These three countries together accounted for just over 42% of the facilities surveyed. The report also highlights South Africa’s importance in colocation capacity. These figures describe geographical concentration; they do not mean that every centre is equipped for AI-specific workloads.
An International Monetary Fund research paper on AI in sub-Saharan Africa reports, citing earlier research, that nearly half of the region’s data centres are in South Africa, Nigeria and Kenya. It also notes that access to electricity and network reliability remain obstacles for businesses. These details provide context for the agreement, without confirming the separate estimate that only 5% of African data centres are suitable for AI. That figure, cited in the initial material, could not be corroborated with an identifiable primary source.
An opportunity, but also choices to watch
For Mauritius Telecom, the agreement could open up new opportunities for products developed in Mauritius. For Cassava, it represents a chance to connect its regional infrastructure with services already designed by an African operator. But access to computing power guarantees neither viable demand nor lasting adoption. Costs for schools, connectivity and teacher training will matter as much as server capacity.
The location of infrastructure also raises questions about control. The partners say that the use of Cassava’s capacity and models will depend on use cases and the authorisation of the customers concerned. It remains to be specified where data will be processed, which rules will apply and how school information will be protected. For Mauritius and South Africa alike, these arrangements will shape trust in the platform.
At SADC and continental level, the agreement illustrates a path for cooperation between African companies: combining infrastructure concentrated in a major regional market with products developed elsewhere on the continent. The next concrete indicator will be the publication of the results of the assessment of myt Learn in South Africa. Until then, the memorandum marks a commercial and technological avenue, not yet a presence for the platform in South African schools.



