Almost every organisation on the continent runs on infrastructure it does not own — foreign clouds, foreign platforms, foreign models, and cables and compute controlled elsewhere. Sovereignty is usually reduced to a question of data residency: keep the servers onshore and the problem is solved. It isn't. If the analytics layer, the models, the distribution and the talent all sit outside the region, ownership of the storage is a technicality. The harder questions are the ones companies rarely put on a slide — what actually breaks if a vendor changes its terms, whether "compliant" means anything three suppliers deep, whether an enterprise is anywhere near data-ready for the AI strategy its board has already announced, and what any of this is worth if it cannot be secured.
This panel takes those questions to four people who deal with them in practice rather than in theory. Anousha Mahadea (CEO, Currimjee Informatics) brings the boardroom view of data, governance and where the budget actually goes. Dylan Harbour (Director of Technology, Ringier South Africa) speaks to platform dependency and what sovereignty costs at the architecture level. Dante Sassenberg (Head of Pentesting, Integrity360) spends his working life finding out what holds under test and what doesn't. Naveesh Doolhur (Head of AI & Data Analytics, Mauritius Telecom) sits where national infrastructure meets applied AI. Across five rounds — dependency, risk and regulation, AI strategy, and execution — the session moves from diagnosis to what an organisation can actually do about it on Monday morning. Audience Q&A follows.
