Hardware
Cisco Live: The router that
runs the factory
Cisco’s new ruggedised, AI-powered routers bring machine learning to the factory floor, turning IoT into industrial intelligence, writes ARTHUR GOLDSTUCK.
When Cisco launched its latest industrial networking innovations at Cisco Live 2025 in San Diego yesterday (11 June), the real stars emerged far from sleek devices or flashy demos. They stood in tough metal casings, built to endure extremes of heat, cold, dust and vibration.
These are Cisco’s new AI-powered routers and ruggedised switches, designed to embed intelligence directly into the harshest industrial environments.
The new lineup features 19 additions to the networking giant’s Industrial Ethernet family, from IP30 to IP67-rated devices, built for conditions found in rail networks, underground mines, and remote substations.
Cisco calls it a generational leap, and it shows how AI now resides at the edge, where the action unfolds.
The IE3500 Rugged Series is a compact, robust switch built to power industrial networks with high data demands. The IE3500H Heavy Duty Series takes that resilience further with reinforced design for extreme conditions. Both offer high Power over Ethernet (PoE) density, enabling them to support smart cameras, Lidar units, and Wi-Fi 7 access points.
Combined with Cisco Silicon and advanced networking protocols like VXLAN MP-BGP EVPN, these switches create consistent, secure fabrics across wildly diverse use cases. Cisco has engineered these switches to thrive in their environments, acting as decision-making hubs in ecosystems previously reliant on centralised data centres.

Photo courtesy Cisco.
In France, energy provider Enedis is using the IE9300 Rugged Series – a high-performance switch built for mission-critical power applications – to upgrade 2,300 substations. With support for IEC 61850-3 compliance and remote zero-touch deployment via Cisco Catalyst Center (a cloud-based network management tool), Enedis gains smarter substations that are easier to scale. Nutrien, a global fertiliser giant, cites the same gear for keeping operations running underground, praising the integration of control and monitoring into their infrastructure.
The offering includes more than hardware. Cisco is blending Wi-Fi 7 and its Ultra-Reliable Wireless Backhaul (URWB) into a single dual-channel access point, designed for resilient, dynamic connectivity in high-interference environments. This unit automatically shifts data loads, reroutes signals during interruptions, and supports real-time machine-to-machine collaboration.
Cisco Cyber Vision is the security intelligence layer of this ecosystem. Now powered by AI clustering, it identifies and groups devices into security zones, applies segmentation policies, and detects threats like irregular access or identity misuse. Paired with Cisco Secure Equipment Access (SEA), users can ask an AI assistant for threat explanations or network behaviour. This replaces static dashboards with interactive, intelligent insights.
At Cisco Live, a demo pulled in curious engineers and CIOs alike: a guitar refurbishment factory where robots scanned for flaws, applied GenAI-driven repairs, and tuned strings based on pitch detection algorithms. The entire production loop ran on an AI-powered network backbone – real-time robotics, vision systems, predictive diagnostics – coordinated by industrial-grade routers.
Samuel Pasquier, VP of product management for Cisco Industrial IoT Networking, spelled out the process: “First, precision-robots scan the guitar’s body for any imperfectons like scratches or dents (AI machine vision). Generatve AI then steps in, suggesting creative fxes for these flaws, which are encoded into RFID tags on each pallet. At the next station, the magic of music is found. As the guitar pauses before pickup, a robot plucks a string, and Gen AI listens to the tone, adjusting the tuning pegs to perfection.”
This marks a new phase for Industrial IoT: moving from passive data collection to active decision-making. Devices are no longer endpoints, but participants. Routers and switches, embedded with intelligence, now orchestrate operations, enhance security, and manage performance in real time.
In mining, that means spotting equipment fatigue before a breakdown. In transport, it enables dynamic routing based on weather, load conditions or AI-generated risk maps. The router becomes the bridge between operational technology (OT) and IT analytics, with no middleware required.
Meanwhile, a protocol called VXLAN MP-BGP EVPN allows for scalable and secure network segmentation across locations. It ensures a mine shaft’s data stays isolated from a refinery’s or a rail depot’s, while giving enterprise teams centralised visibility.
Engineers on the ground benefit from secure remote access, mobile dashboards, and collaboration tools overlaying digital models on physical systems. They troubleshoot in real time, connect with global experts, and receive AI-generated recommendations tailored to their site.
The net result: AI becomes infrastructure. It lives inside routers, automates security policies, visualises workflows, and predicts failures. It adapts constantly, and all from within rugged casings that look more like industrial batteries than networking gear.
AI Strategy in the past year or so has often been about sprinkling AI on top of existing products. At Cisco, as the San Diego event shows, it is about baking it into the very fabric of industrial networking.
* Arthur Goldstuck is CEO of World Wide Worx, editor-in-chief of Gadget.co.za, and author of The Hitchhiker’s Guide to AI.
