New research by Cisco indicates that the rapid expansion of artificial intelligence (AI) is stressing enterprise campus and branch networks, necessitating increased capacity, security, and modernization efforts. A survey conducted by Cisco and Foundry, which involved 3,472 IT and networking leaders across 15 countries, revealed that AI is already altering traffic patterns within these environments.
Jeetu Patel, Cisco’s president and chief product officer, emphasized the importance of networking in AI infrastructure: “We have entered a networking supercycle, because the network is so central to all the AI infrastructure the world is building now.” The study highlights the need for businesses to prepare their networks as they move from generative AI projects to implementing AI agents that require constant communication with other systems.
Key findings from the survey include:
- A 34% increase in AI-related traffic over the past year in campus and branch networks.
- An expected 209% rise in traffic over the next three years.
- 73% of organizations anticipate or are currently facing capacity constraints in their network systems.
- 67% reported that AI workloads are increasing internal traffic demands.
- 80% indicated that AI usage has broadened their potential attack surface.
- 61% are postponing further AI deployment until their security measures are improved.
- 85% foresee moderate to significant increases in AI agent deployment within two years.
The changing traffic landscape presents considerable challenges for network teams. Traditionally, networks are designed for consistent traffic types, yet the unpredictable nature of AI communications complicates this arrangement. Cisco identified "aggressive AI adopters" as those with extensive generative AI use; however, only 30% of these organizations feel fully prepared for anticipated growth in network demands. Consequently, 93% of IT decision-makers are prioritizing network modernization initiatives.
The study also points out a gap in observability, with many IT departments unable to track what AI services are operating on their networks due to the increasing pace of AI experimentation within organizations.
Security continues to be a critical barrier. As companies juggle numerous AI tools, establishing proper governance around them becomes challenging, leading to potential vulnerabilities.
As enterprises develop their AI strategies, it is clear that plans must encompass not just data centers, but also campus and branch networks, which play a crucial role in AI applications. Cisco has urged companies to adjust their infrastructure planning to include these areas to successfully scale AI initiatives in the coming years.
For further insights from Cisco, check out their research on AI’s impact on campus and branch networks and learn about how to support AI workloads while managing security risks.
