Extreme Networks unveiled updates to its AI-powered Platform ONE, signaling a shift from AI-managed networks to networks that actively run AI workloads using idle hardware resources. The approach promises enhanced developer workflows, reduced cloud dependency, and improved operational reliability.
- On-prem AI zones utilize spare network device compute, reducing cloud inference reliance
- Agent ONE Coworker saved customers 1,000 troubleshooting hours in the first month
- Autonomous Operator mode ensures safe, auditable AI-driven network operations
Infrastructure signal
Extreme Networks introduces Exchange zones enabling AI workloads to run directly on network hardware such as access points and switches, supplementing on-prem servers. This architecture taps into unused compute resources at the edge, reducing latency and lowering cloud inference usage. By processing AI locally, organizations can better control data flow and reduce cloud compute expenses tied to AI operations.
These zones rely on open-weight AI models by default, shifting the AI workload closer to data sources. The combination of Extreme Universal Compute Platform hardware and network device compute creates a hybrid AI infrastructure deployed by mid-2027. This approach highlights a new balance between cloud and edge computing, encouraging infrastructure teams to rethink site-level capacity planning and hardware utilization.
Developer impact
Agent ONE Coworker has demonstrated strong engagement and operational impact, with 41% of Platform ONE customers using it daily and 98% responsiveness to critical automated alerts. It reverses traditional AI assistant roles by proactively notifying admins rather than waiting for queries, enhancing developer and admin workflows with actionable insights.
What teams should watch
IT and network operations teams must monitor the rollout of Exchange zones and Operator mode for opportunities to optimize on-prem compute use and reduce reliance on cloud AI inference. Teams should evaluate how AI-driven proactive troubleshooting affects incident response and network reliability metrics, as early results show meaningful time savings and issue resolution speedups.
Security, compliance, and platform governance teams should focus on maintaining strict policy controls embedded in Extreme Exchange rather than core Platform ONE to ensure safe AI tool operation and credential protection. Observability improvements from AI-assisted monitoring also require adaptation in incident workflows to integrate AI-generated nudges and alerts effectively.