Cisco has expanded its Secure AI Factory with NVIDIA to support the rack-scale era of artificial intelligence infrastructure. The new architecture, developed with Supermicro, adds high-density liquid- and air-cooled compute systems designed for demanding AI workloads, including large-scale training, inference and agentic AI.
The expanded platform brings together compute, networking, security, storage, software and observability as an integrated architecture. It supports NVIDIA Vera Rubin NVL72 and HGX Rubin NVL8 configurations and uses Cisco AI networking technologies, while Cisco says the infrastructure is designed to help enterprises, neoclouds and sovereign-cloud operators deploy AI more efficiently and securely.
Cisco is expected to begin offering the Supermicro compute solutions as part of Secure AI Factory with NVIDIA from October 2026.
Key Takeaway
Cisco is moving AI infrastructure toward integrated rack-scale systems designed to combine high-performance computing, networking, security and operational control.