How will Cisco N9100 Address AI Networking Bottlenecks?

Data centre operators deploying AI workloads at scale face networking bottlenecks. The infrastructure connecting thousands of GPUs must handle bandwidth demands whilst managing congestion and maintaining low latency. At GTC in Washington DC, Cisco introduced the N9100 series switch, becoming the first partner to deliver a Nvidia Cloud Partner-compliant reference architecture.

The N9100 is the first Nvidia partner-developed data centre switch based on Nvidia Spectrum-X Ethernet switch silicon. Customers can order the switch before the end of the year, with a choice between Cisco NX-OS or SONiC operating systems.

Cisco’s Nexus data centre switching solutions deliver a unified operating model through Cisco Nexus Dashboard across Silicon One, Cloud-scale ASICs and switches built on Spectrum-X Ethernet switch silicon. This attempts to maintain operational consistency as organisations scale beyond hundreds of GPUs.

“We’re at the beginning of the largest data centre build-out in history,” says Jeetu Patel, President and Chief Product Officer at Cisco. “The infrastructure that will power the agentic AI applications and innovation of the future requires new architectures designed to overcome today's constraints in power, computing and network performance.”

Cisco Cloud Reference Architecture targets neocloud infrastructure demands

Cisco has developed the Cisco Cloud Reference Architecture for neocloud and sovereign cloud customers, based on Nvidia's Cloud Partner reference architecture. The architecture utilises Cisco’s Silicon One and Cloud-scale ASIC offerings, including the Cisco 8223 based on Silicon One P200, Nvidia BlueField-4 DPUs and Nvidia ConnectX-9 SuperNICs.

“Nvidia Spectrum-X Ethernet delivers the performance of accelerated networking for Ethernet,” says Gilad Shainer, SVP of Networking at Nvidia. “Working with Cisco’s Cloud Reference Architectures and Nvidia Cloud Partner design principles, customers can choose to deploy Spectrum-X Ethernet using the newest Cisco N9100 series or Cisco Silicon One based switches to build open, high-performance AI networks.”

Cisco Secure AI Factory with Nvidia expands security capabilities

Cisco has expanded the Cisco Secure AI Factory with Nvidia, which launched at GTC in March 2025. Cisco AI Defense now works with Nvidia NeMo Guardrails for AI application cybersecurity. Organisations can deploy Cisco AI Defense on-premises.

Splunk Observability Cloud monitors the AI application stack, tracking performance, quality, security and cost. Splunk Enterprise Security extends this to protect AI workloads.

Cisco Isovalent has been validated for inference workloads on AI PODs. Cisco Nexus Hyperfabric AI with a Cisco G200 Silicon One switch delivers 800G Ethernet. Cisco UCS 880A M8 rack servers with Nvidia HGX B300 and Cisco UCS X-Series modular servers with Nvidia RTX PRO 6000 Blackwell Server Edition GPUs are orderable as part of AI PODs.

Nvidia Run:ai software is available through Cisco, enabling AI workload and GPU orchestration. The ecosystem includes Nutanix Kubernetes Platform and Nutanix Unified Storage. Nutanix Enterprise AI solution simplifies building containerised inference services.

Cisco is collaborating with Nvidia on Nvidia AI Factory for Government, a reference design for AI workloads in regulated environments.

Cisco and Nvidia develop AI-native wireless stack for 6G networks

Cisco, Nvidia and telecom partners have developed the first American AI-RAN stack for mobile networks that integrates sensing and communication. The stack allows telecom providers to integrate AI into mobile networks, starting with 5G services. The technology combines Cisco’s user plane function and 5G core software with Nvidia AI Aerial platform.

“Cisco’s N9100 series powered by Nvidia Spectrum-X Ethernet switch silicon, provides a solution for high-performance, open infrastructure to meet our AI cloud demands,” says Yih Leong Sun, Head of Infra at GMI Cloud. “The capability to run NX-OS or SONiC under a unified operating model on Nexus Dashboard delivers more flexibility to our customers with operational simplicity.”

“As the demand for computing power continues to grow, our GPU clusters are expanding rapidly in scale,” says Junfeng Cheng, Head of Networking Infrastructure at Xiaohongshu. “With the Cisco Nexus platform, we can fully leverage the AI networking capabilities of Nvidia Spectrum-X Ethernet switch silicon without incurring additional operational or development costs and seamlessly integrate with our existing systems.”