Nvidia, Cisco look to deepen AI innovation across 6G, telecoms (www.computerweekly.com)

Cisco and Nvidia announced a broad partnership to accelerate secure, scalable AI infrastructure across datacentres, enterprises and telecoms. Central to the announcement is the Cisco N9100 — billed as the first Nvidia partner-developed datacentre switch using Nvidia Spectrum‑X Ethernet silicon — available to order before year‑end and offered with either Cisco NX‑OS or Sonic. Cisco is also publishing an Nvidia Cloud Partner‑compliant reference architecture aimed at neocloud and sovereign cloud deployments that combines Cisco’s Silicon One and cloud‑scale ASICs. Complementing the hardware, Cisco’s Secure AI Factory with Nvidia (unveiled March 2025) layers security and observability across AI stacks, built on Cisco AI PODs and Silicon One‑powered Nexus switching, while new integrations target core AI infrastructure, ecosystem expansion and government readiness. The move is significant because it ties high‑performance Ethernet switching, cloud reference designs and end‑to‑end security directly to Nvidia’s AI ecosystem, lowering friction for organizations building large‑scale, regulated or sovereign AI clouds. Cisco and partners also unveiled what they call the industry’s first AI‑native wireless stack for 6G (an American AI‑RAN) that integrates sensing and communications for pre‑6G applications — a signal that telcos will be targeted for real‑time, high‑density AI workloads at the edge. For the AI/ML community this means more standardized, carrier‑grade building blocks for scaling model training and inference, tighter security/observability controls, and an explicit push to modernize networks to meet the power, latency and connectivity demands of agentic and distributed AI.