Cisco Expands Secure AI Factory with NVIDIA for Rack-Scale Era

Cisco expands its Secure AI Factory with NVIDIA, adding Supermicro rack-scale compute to deliver validated, full-stack AI infrastructure for enterprises, neoclouds and sovereign clouds.

AI models keep evolving, placing new performance demands on networks, power and cooling. Enterprises need infrastructure that not only delivers raw compute but also protects data, controls costs and accelerates time to value. The next era of AI infrastructure is defined by validated, full-stack architectures that customers can deploy quickly, run efficiently and secure with confidence. Cisco is addressing this shift by expanding its Secure AI Factory with NVIDIA to the rack-scale era.

What was announced

On August 25, 2026, Cisco announced an expansion of the Secure AI Factory with NVIDIA through a new partnership with Supermicro. The collaboration brings high-density, liquid- and air-cooled compute systems into Cisco's full-stack AI infrastructure portfolio for enterprises, neoclouds and sovereign clouds.

The expanded architecture supports use cases ranging from trillion-parameter model training to edge inferencing. It includes NVIDIA Cloud Partner (NCP) compliant solutions, Cisco AI networking systems, and support for platforms such as NVIDIA Vera Rubin NVL72 and NVIDIA HGX Rubin NVL8. Cisco will begin offering Supermicro compute solutions as part of the Secure AI Factory with NVIDIA in October 2026.

We are at the beginning of one of the largest datacenter buildouts in history. Every organization is racing to scale AI, but speed only counts if it comes with control of data, managed token costs and real ROI

Why this matters

For CIOs, CISOs, IT directors and procurement leaders, AI infrastructure decisions now carry strategic weight. GPU and memory availability, power and cooling constraints, and data sovereignty requirements all shape deployment plans. A validated, secure, full-stack approach reduces integration risk, shortens time to production and aligns AI investments with measurable business outcomes.

By combining Cisco networking, security and observability with Supermicro compute and NVIDIA accelerators, organizations gain a single architecture that scales from core data centers to edge, with NCP-compliant designs suited to neocloud and sovereign cloud operators.

Technical details

• Rack-scale compute: Supermicro liquid- and air-cooled systems validated within the Cisco AI infrastructure portfolio.
• NVIDIA platforms: Support for NVIDIA Vera Rubin NVL72 and NVIDIA HGX Rubin NVL8 for trillion-parameter training and high-throughput inference.
• Unified networking: Cisco Silicon One switches for the front-end and NVIDIA Spectrum-X based switches for the back-end, unified by Cisco Nexus One.
• NCP compliance: NVIDIA Cloud Partner compliant solutions for neocloud and sovereign cloud customers.
• Liquid cooling: Rack-to-fabric liquid cooling combining Cisco liquid-cooled networking and Supermicro liquid-cooled servers.
• Services: New Cisco Validated Infrastructure Services (CVIS) aligned with NVIDIA Infrastructure Services (NVIS).
• Operations: NVIDIA AI Enterprise software plus AgenticOps through Cisco Cloud Control for end-to-end observability.

Softprom and Cisco

Softprom is the official distributor of Cisco. Our team helps enterprises, service providers and public sector organizations design, deploy and support Secure AI Factory architectures aligned with real business goals.

This content was prepared as part of the Softprom DistriFlow project — an automated system for monitoring and adapting vendor news. Original source: original article.