Cassava Artificial Intelligence deploys NVIDIA Nemotron to accelerate the adoption of agent-based AI in its market.
Cassava is deploying NVIDIA Nemotron models via its locally hosted, managed inference service on the Cassava AI Multi-Model Exchange Platform (CAIMEx).
Berlin, October 21, 2026 /PRNewswire/ — Global technology company Cassava Technologies today announced it has integrated NVIDIA’s Nemotron open models into its AI Factory capabilities to help businesses and service providers achieve growth and innovation through AI more efficiently and cost-effectively.
Cassava is deploying NVIDIA Nemotron models through its locally hosted, managed inference service on the Cassava AI Multi-Model Exchange Platform (CAIMEx). By offering open models via API on Cassava’s AI Factory, CAIMEx enables enterprises to flexibly select optimal model combinations for their applications without having to build and operate their own model infrastructure.
Ahmed El Beheiry, Chief Technology and AI Officer at Cassava Technologies, said: “Expanding our AI Factory across our markets hinges on delivering the capabilities and cost-efficiency enterprises need to build and deploy AI solutions at scale. By introducing NVIDIA Nemotron into CAIMEx, enhancing the efficiency of every GPU cycle, and launching flexible token options, we are giving enterprises and service providers more choices in how they access these capabilities and deliver them to their customers.”
NVIDIA Nemotron brings efficient inference and tool-calling capabilities to CAIMEx, enabling customers to build AI agents capable of handling complex tasks and interacting with enterprise tools. Its open model weights, training data, and training protocols provide developers with a customizable foundation upon which they can fine-tune and create specialized agents tailored to specific use cases.
As enterprises mature and AI agents begin to handle longer-running tasks and access enterprise systems, security and protection mechanisms will become critical requirements. These capabilities must be built into the infrastructure. Cassava says the company is using NVIDIA OpenShell to integrate model compute, governed agent execution, and isolated security execution into its AI Factory to support customers’ agent-based AI workloads.
The CAIMEx Token Optimizer enhances the economics of generative AI and agent-based AI by identifying previously processed information and reusing it when appropriate, rather than reprocessing the same content. This reduces computational demands and helps customers use their AI tokens more efficiently. The approach works with both Cassava-hosted models and third-party AI models, enabling businesses to reduce unnecessary processing as their AI usage scales.
Cassava stated that this expansion is its latest step toward building an integrated AI ecosystem that provides businesses with on-premises infrastructure, models, and services, enhancing trust and confidence in data security and control while supporting business growth.
This press release is issued by APO Group on behalf of Cassava Technologies.
About Cassava Technologies:
Cassava Technologies is a technology company that provides vertically integrated digital services and infrastructure ecosystems to drive digital transformation. Through business units such as Liquid Intelligent Technologies, Liquid C2, Africa Data Centres, Distributed Power Africa, and Sasai Fintech, Cassava operates in key growth markets including Africa, the Middle East, Latin America, and the United States. Cassava delivers relevant services to customers in 94 countries, helping them grow, transform, and expand their businesses.
Media contact: [email protected]