Cisco Expands Secure AI Factory With NVIDIA For Rack-Scale AI

Cisco is adding Supermicro rack-scale systems to its Secure AI Factory with NVIDIA, targeting high-density AI workloads across enterprise and cloud deployments.
Cisco is expanding its Secure AI Factory with NVIDIA through a partnership with Supermicro, adding high-density rack-scale computing systems to its AI infrastructure portfolio.
The company announced the expansion on August 25, saying the new architecture is designed to support AI deployments across enterprises, neoclouds and sovereign clouds. The offering combines Supermicro liquid- and air-cooled server systems with Cisco's networking, security and infrastructure services.
The move extends Cisco's existing Secure AI Factory with NVIDIA architecture, which was introduced in 2025 to combine AI compute, networking and security into a validated infrastructure stack. Cisco has also expanded its AI infrastructure business as demand for networking and computing systems for AI workloads has increased.
Cisco Adds Rack-Scale Compute To Its AI Infrastructure
Under the expanded offering, customers will have access to Supermicro systems designed for high-density GPU deployments. Cisco said the systems will support both liquid- and air-cooled configurations and can be deployed alongside Cisco AI networking systems.
The architecture will support NVIDIA platforms including the NVIDIA Vera Rubin NVL72 and NVIDIA HGX Rubin NVL8. The NVL72 is a rack-scale system built around 72 NVIDIA Rubin GPUs, while the HGX Rubin NVL8 uses eight Rubin GPUs.
Cisco said the addition is intended to cover workloads ranging from trillion-parameter model training to high-throughput inference. The company is also positioning the infrastructure for customers that need to operate AI systems alongside conventional workloads.
The expanded architecture includes Cisco Silicon One-based switches for front-end networking and Cisco switches based on NVIDIA Spectrum-X for the back end. Cisco said these components will be unified through Cisco Nexus One, creating a common networking architecture across the stack.
Networking is becoming a larger consideration as AI systems increase in scale and power requirements. Cisco has previously highlighted the relationship between AI infrastructure, networking capacity and power constraints in its work around Silicon One and AI data-center infrastructure.
Focus On Cloud Providers And Deployment Speed
Cisco said the expanded Secure AI Factory will offer solutions compliant with NVIDIA Cloud Partner (NCP) requirements. The company is targeting neocloud and sovereign cloud operators that are building infrastructure for large-scale AI workloads.
The architecture will also include Cisco Validated Infrastructure Services (CVIS), which Cisco said will certify that deployed infrastructure matches its intended design and relevant reference architectures. Cisco is investing in a dedicated AI lab to develop tools and test software for the service.
For operations, Cisco said customers will be able to use NVIDIA AI Enterprise software alongside AgenticOps capabilities through Cisco Cloud Control. The company said these tools will provide visibility across compute, network interface cards, optics and network performance.
The expansion follows Cisco's broader effort to package AI infrastructure into validated architectures rather than individual components. Its work with NetApp on validated infrastructure for enterprise AI is another example of that approach.
Cisco said it will begin offering Supermicro compute solutions as part of Secure AI Factory with NVIDIA in 10/2026.
Key Takeaways
- Cisco expands its Secure AI Factory by integrating Supermicro rack-scale systems for high-density AI.
- The partnership with Supermicro enhances Cisco's NVIDIA AI infrastructure for enterprise and cloud deployments.
- New architecture combines Supermicro servers with Cisco's networking, security, and infrastructure services.
- This move extends Cisco's existing Secure AI Factory framework, introduced in 2025.