Cassava Technologies, the pan-African infrastructure group, will offer Nvidia's Nemotron open models as a managed service on its AI Factory, Techweez reported on 21 October. Businesses and service providers can call the models as APIs through the Cassava AI Multi-Model Exchange (CAIMEx), without building and running their own model infrastructure.
Cassava describes Nemotron as built for reasoning and for using software tools, the abilities AI agents need to work through multi-step tasks inside company systems. Nvidia has released the models' weights, training data and methods, so developers can adapt them.
Two more pieces
- The CAIMEx Token Optimiser, which Cassava says spots content a model has already processed and reuses it. It works with Cassava-hosted and third-party models
- Nvidia OpenShell, which Cassava is using to combine computing, controlled agent activity and independent security enforcement as agents gain access to business systems
"Expanding our AI Factory across our markets is about creating the capacity and economics needed for businesses to build and deploy AI solutions at scale," said Ahmed El Beheiry, Cassava's group chief technology and AI officer.
Why where the model runs matters
Cassava announced its AI Factory in March 2025, with data centres in South Africa, Egypt, Kenya, Morocco and Nigeria lined up for Nvidia hardware and a stated aim of keeping data inside Africa. It later added GPU rental, and in November 2025 agreed to provide computing capacity to Rockefeller Foundation grantees in eight countries.
For many companies the question is compliance as much as cost. Techweez notes that Kenyan firms using foreign cloud AI send customer data abroad, which can complicate compliance with Kenya's Data Protection Act, 2019. A model hosted locally and reached as an API is one way around that.
