In the era of agentic AI, the network is becoming the critical foundation for trust, security and the seamless operating models through which people and AI can build the future.
However, as organisations move quickly to gain an advantage from AI, a significant gap has emerged between their ambitions and the capabilities of their legacy infrastructure.
These are among the central findings of Cisco’s new study, No time to wait: The accelerating impact of AI on campus and branch networks. Conducted by Cisco and Foundry Research, the report is based on a survey of more than 3 400 senior IT and networking decision-makers across 15 countries.
The findings highlight the urgency of addressing the growing infrastructure challenges associated with AI adoption.
According to the research, 73% of organisations are already experiencing, or expect to experience, campus and branch network capacity constraints within the next two years. Respondents reported that their network traffic increased by 34% over the past year. They expect it to grow by a further 96% over the next 12 months and by 209% over the next three years.
The rising demand will be driven by agentic AI, generative AI and physical AI, including robotics and AI-driven Internet of Things technologies. The adoption of these technologies shows no sign of slowing. Trust will be essential for organisations seeking to succeed with AI. This includes trust in their infrastructure, security and AI governance, all of which ultimately depend on the network.
Cisco president and chief product officer Jeetu Patel has described this period as a “networking supercycle”, reflecting the network’s central role in the AI infrastructure currently being developed. The Cisco research reinforces the scale of the challenge. More than 90% of respondents fear that their organisations could face financial and competitive risks if campus and branch networks are not adapted to meet AI-driven demand.
Despite these concerns, only 30% of even the most aggressive AI adopters say they are fully prepared for the projected growth of AI across their networks.
Securing trust in AI
Security is fundamental to establishing trust in AI, and the network provides a natural point from which organisations can anchor that security.
Cisco’s study found that 80% of respondents expect security risks to increase further as AI adoption expands beyond generative AI applications. Another 80% said AI had already expanded their organisation’s attack surface during the previous 12 months. As a result, 61% are holding back from scaling their AI initiatives until they have greater confidence in their security.
The leading challenges include the additional complexity involved in securing AI models and the increasing sophistication of AI-driven cyberthreats. A retail executive quoted in Cisco’s study described the difficulty organisations face in establishing appropriate security guardrails for every AI tool they may need to use.
The network can help organisations address several of these challenges.
Cisco’s platform strategy is designed to reduce complexity by integrating networking, security, visibility, governance and data intelligence. This integration enables IT and security teams to monitor and protect their organisations from the data centre through to the network edge.
In an interview with VentureBeat, Michael Dickman, Cisco Senior Vice President and General Manager for Campus Networking, explained how Cisco’s platform strategy can create new opportunities to strengthen security, use data and support additional applications. The strategy includes advanced solutions such as Cisco Cloud Control.
Dickman explained that a platform approach enables organisations to share data in ways that were not previously possible. Once that capability is available, he said, “anything you want to do is now built on a solid foundation.”
According to Dickman, that network foundation enables the different building blocks required for trust to be integrated, simplified and streamlined. He also stressed the importance of ensuring that an organisation’s first agentic AI use cases succeed from a trust perspective. This requires organisations to incorporate role-based access control, privileged access management and micro-segmentation into their approach.
Making AI visible and simple
Autonomous systems and highly distributed technology environments are becoming increasingly complex. Against this background, 71% of respondents in Cisco’s study reported growing blind spots in their monitoring and visibility.
Technology leaders increasingly see the network as a platform through which they can connect fragmented environments, extend visibility and maintain policies at scale. Dickman explained that when information is brought together through a shared data fabric, it can support a broad range of applications. These include security-related use cases, the optimisation of business processes and technology spending, and the centralisation of visibility.
Removing fragmentation through a common data fabric can therefore provide organisations with significant new capabilities. This is why Cisco is placing increased emphasis on its One Cisco platform. The approach brings together the breadth of Cisco’s portfolio within a unified platform.
As the demands created by AI continue to increase, the network platform is becoming increasingly essential. Building trust in AI while reducing complexity and fragmentation is no longer optional. Respondents in Cisco’s study identified business and operational risks their organisations could face if they fail to meet AI’s infrastructure demands.
Patel summarised the stakes by warning that companies will increasingly be divided between “those that are AI companies, and those that are irrelevant.”