Cisco Says Networking Now Critical as AI Scales Up in Organisations

As AI adoption scales up in organisations, a critical challenge has emerged: building infrastructure that can reliably support AI workloads in production. According to Arun Shetty, CTO & Senior Director of Solutions Engineering at Cisco India & South Asia, this shift is forcing technology leaders to rethink long-standing assumptions about networking, security, observability, and data architecture.

Cisco's role is evolving beyond just connecting AI infrastructure, as the company becomes an integral part of the infrastructure that enables AI to operate at scale. Shetty explains that the transition from pilots to production changes everything, with organisations no longer looking for isolated AI deployments but rather an integrated infrastructure stack that combines compute, networking, security, storage, and observability in a unified way.

Networking is becoming a strategic differentiator as AI workloads require predictable performance, efficient GPU utilisation, lower power consumption, and consistent throughput. Shetty notes that many AI initiatives fail to scale because of data centre limitations rather than model performance, highlighting the importance of addressing infrastructure constraints, trust, and data fragmentation.

Cisco is integrating Splunk's observability and security capabilities into its AI strategy to provide unified, real-time visibility across AI applications, infrastructure, and networks. This enables organisations to distinguish between performance issues and security risks by correlating telemetry data and applying AI to it.