Cohere announced NVIDIA ecosystem-native models and an optimized version of its agentic North platform for NVIDIA DGX Spark. The partnership targets regulated industries that need high-performance AI running securely on-premise within national borders.
Cohere, an enterprise AI platform, is building custom models optimized for NVIDIA's latest architecture and software ecosystem. These models will ship alongside an optimized version of North, Cohere's agentic platform, on NVIDIA DGX Spark — a high-performance, locally run enterprise workstation. The collaboration targets specialized enterprise workloads where frontier models still face limitations across specific hardware footprints.
This partnership addresses a growing gap in enterprise AI: regulated industries like finance, healthcare, and public sector need AI systems that operate securely within national borders. Running North locally on DGX Spark gives organizations a full stack — agents, models, and search — that handles sensitive data at low latency without leaving their infrastructure.
Evaluate DGX Spark as an on-premise AI deployment option for workloads where data residency and security requirements rule out cloud-hosted models.
