In the agentic AI era, the network is the critical foundation for trust, security, and the seamless operating models from which humans and AI can build the future.

 

But as organizations scramble to gain an AI advantage, there's a wide gap between their ambitions and what legacy infrastructure can support.

These are critical takeaways from a new Cisco study, “No time to wait: The accelerating impact of AI on campus and branch networks.” Based on a survey of more than 3,400 senior IT and networking decision-makers across 15 countries, the report — conducted by Cisco and Foundry Research — represents a stark call to action amid the rising challenges of AI adoption.

In the survey, 73% of organizations are already facing or anticipate campus and branch network capacity constraints within two years. That’s not surprising, since they reported that their network traffic had risen 34% in the past year. Over the next 12 months, traffic is expected to increase 96%, and over the next three years 209%.

These demands will be driven by agentic AI, as well as generative AI, and physical AI (robotics and AI-driven Internet of Things), all of which show no sign of slowing.

For organizations to succeed with AI, trust is essential. That is, trust in infrastructure, trust in security, and trust in AI governance — all of which lead back to the network.

As Cisco president and chief product officer Jeetu Patel stressed, we have entered a “networking supercycle, because the network is so central to all the AI infrastructure the world is building now.”

This was underscored by Cisco’s survey, in which more than 90% feared financial and competitive risk if campus and branch networking is not adapted for AI-driven demand. Yet only 30% of even the most aggressive adopters say they are fully prepared for projected AI growth across the network.

Securing trust in AI

Of course, security is a foundational element to trust. And the network is its natural anchor point.

In Cisco’s survey, 80% expect security risks to further increase as AI adoption expands beyond generative use cases, and another 80% said the technology has already expanded their attack surface in the last 12 months. As a result, 61% are holding back from scaling AI initiatives until they have greater confidence in their security.

Key challenges include added complexity in securing AI models and the increased sophistication of AI-driven cyber threats.

A retail executive quoted in Cisco’s study outlined the problem. “The issue from a security standpoint,” he said, “is that it’s hard to create the guardrails for every possible AI tool that your organization must use.”

Again, the network can solve a multitude of problems.

Cisco’s platform strategy, for example, cuts through that complexity. By integrating networking, security, visibility, governance, and data intelligence, IT and security teams can monitor and protect, from the data center and beyond, out to the edge.

In a VentureBeat interview, Michael Dickman, Cisco’s SVP and general manager for campus networking, spoke of how Cisco’s platform strategy — now including advanced solutions like Cisco Cloud Control — unlocks all-new possibilities for ensuring security, leveraging data, and more.

“When you have that platform approach,” he said, “all of a sudden you have data sharing in a way that wasn’t possible, and that's the unlock — anything you want to do is now built on a solid foundation.”

Dickman stressed that that solid network foundation enables the building blocks for trust to be integrated, simplified, and streamlined.

“Make sure those first examples of agentic AI use cases are successful from a trust standpoint,” he said, “and that means building in role-based access control, privileged access management, micro-segmentation.”

Making it all visible — and simple

Given the complexity of autonomous systems and highly distributed environments, it’s not surprising that 71% of respondents in Cisco’s study described growing blind spots in monitoring and visibility.

But increasingly, leaders see the network as the platform from which fragmentation can be connected, visibility extended, and policy maintained at scale.

“When all that information comes into that one data fabric,” Dickman said, “it can then be used to inform all kinds of use cases. I'm excited about a lot of security-oriented use cases, but there's so much more. There's optimization of business process, there's optimization of technology spend, there's centralization of visibility.”

“When fragmentation goes away through that common data fabric,” he continued, “it's very powerful. And that's why at Cisco you'll see so much focus on this One Cisco platform, which is explicitly bringing together the breadth of what Cisco does in the portfolio into a true platform.”

That network platform is increasingly essential.

Because building trust in AI, while lessening complexity and fragmentation, is no longer optional. The high stakes were underscored throughout Cisco’s survey, with respondents citing business and operational risks if they failed to meet the infrastructure demands of AI.

Or, as Patel put it: “Eventually there will be only two kinds of companies: those that are AI companies, and those that are irrelevant.”