AI-Driven Networks Need Capacity, Complexity, and Security

Cisco's chief architect of provider mobility, Bob Everson, testified before the US Senate subcommittee on telecommunications and media about the impact of AI on networks. He emphasized that AI is profoundly changing enterprise and service provider networks, requiring more capacity and complex designs to run efficiently.

The hearing focused on how widespread AI use has forced networks to evolve. According to Everson, Cisco measured a fourfold increase in AI inference traffic over eight months, with networks optimized for content flowing downstream but now facing two-way and uplink-intensive AI workloads.

Everson highlighted the challenges of managing complex networks, citing a 34% increase in traffic tied to AI workloads over the last 12 months and an expected 96% increase this coming year. He also emphasized the importance of data sovereignty and security, with enterprises and governments concerned about moving sensitive information across public internet.

Cisco's AI-native tools enable networks to reroute traffic, adjust capacity, or reconfigure nodes when performance degradation or hardware failure is detected. Everson suggested that automating repetitive tasks can close the workforce talent gap in managing software-defined networks.