Cassava Technologies has announced the expansion of its AI Factory capabilities by deploying NVIDIA Nemotron open models as a locally hosted, managed inference service. The models are being made available to enterprise customers and service providers through the Cassava AI Multi-Model Exchange.
Managed inference expansion
Cassava Technologies announced it is expanding its AI Factory capabilities by incorporating NVIDIA Nemotron open models into its infrastructure. The models are hosted locally and delivered as a managed inference service via the Cassava AI Multi-Model Exchange (CAIMEx).
By providing open models as APIs on the AI Factory platform, Cassava said CAIMEx allows enterprises to select and combine models for their applications without needing to construct and operate dedicated model-serving infrastructure.
Capabilities and agent security
NVIDIA Nemotron provides reasoning and tool-use capabilities designed to help customers construct AI agents that complete complex tasks and interact with enterprise software tools. Developers can fine-tune the open model weights, training data, and recipes to create specialized agents.
To manage security and policy enforcement as AI agents interact with business systems, Cassava is incorporating NVIDIA OpenShell into its AI Factory. The integration combines model compute, governed agent execution, and security enforcement for agentic workloads.
Efficiency and token reuse
The expansion also includes the CAIMEx Token Optimiser, which is designed to reduce compute requirements for generative and agentic AI. The feature identifies data that has already been processed and reuses it where applicable rather than reprocessing the same content.
Cassava stated that the optimization process applies to both Cassava-hosted and third-party AI models. Group Chief Technology & AI Officer Ahmed El Beheiry said the deployment and token options aim to provide businesses with flexible choices for scaling AI solutions.
Key facts and where they come from
- Cassava Technologies is expanding its AI Factory capabilities using NVIDIA Nemotron open models.
Cassava Technologies (www.CassavaTechnologies.com), a global technology leader, has announced the expansion of its AI Factory capabilities with NVIDIA Nemotron open models
- The models are delivered as a locally hosted managed inference service on CAIMEx.
Cassava is deploying NVIDIA Nemotron (https://apo-opa.co/4dRrqnx) models as a locally hosted, managed inference service through the Cassava AI Multi-Model Exchange (CAIMEx)
- NVIDIA OpenShell is used to provide security and governance for agentic workloads.
Cassava is using NVIDIA OpenShell to bring model compute, governed agent execution and independent security enforcement together for customers’ agentic workloads to its AI Factory.
- The CAIMEx Token Optimiser reuses previously processed content to cut processing needs.
The CAIMEx Token Optimiser is designed to improve the economics of generative and agentic AI by identifying information that has already been processed and, where appropriate, reusing it
