Cisco Warns Agentic AI Will Put New Pressure on Enterprise Networks

agentic AI – At Cisco Live in Las Vegas, Cisco leaders warned that enterprise AI is shifting from chatbots to agentic systems, bringing far higher and steadier network traffic. CEO Chuck Robbins said networking demand could at least triple in three years, while executives

When Cisco Chair and CEO Chuck Robbins stepped on stage at Cisco Live in Las Vegas, he didn’t talk about which chips companies should buy next. He talked about what happens when they try to run their AI ambitions over networks that weren’t designed for machine-speed decision making.

Robbins argued that the next major constraint for enterprises may be the network itself. The logic was simple—and uncomfortable for IT teams who’ve spent years betting their AI plans on model access and compute upgrades. He said AI traffic could at least triple over the next three years as enterprises move from chatbots to agentic systems that can act. communicate. and make decisions at machine speed. “Every technical innovation over the past 40 years and all the current buzz about GPUs. inferencing. agentic AI. and the latest models — it is all dependent on the network. ” Robbins said. “We expect at least a tripling of networking traffic over the next three years due to the evolving capabilities and consumption demands of AI.”.

Starbucks’ network, rebuilt for prediction
Robbins ended his keynote by introducing Brian Niccol, chairman and CEO of Starbucks Coffee Company, and the conversation quickly turned from abstract infrastructure to day-to-day operations.

Niccol described how the “original function” of the network in Starbucks was to provide customers with a reliable Wi‑Fi experience in every store to attract them. Now, he said, the network has to do more work: help forecasting, scheduling, and supply chain management.

He offered a specific example of what that looks like in practice. Instead of sending full cases of coffee. syrup. and other supplies to stores. Starbucks is transitioning to sending smaller amounts based on real-time and near-real-time knowledge. That includes in-store inventory, traffic volumes, supply chain availability, and predicted consumption.

More information, Niccol said, makes predictions more accurate—down to when and what people will order. He put a human-sized example behind the math: “If it’s Tuesday on the East Coast and 80°F, we have much better predictability on how many bottles of water a specific store will need,” he said.

It’s a familiar promise in retail, but Niccol’s point landed because it depends on connectivity that can deliver up-to-date signals fast enough to matter.

Agents don’t browse—they swarm
Jeetu Patel. Cisco president and chief product officer. carried that theme into what he framed as a market pivot: from AI chatbots to agentic AI “coworkers” that can operate autonomously. For Patel, the shift isn’t only about software behavior. It changes the network’s rhythm.

“Chatbots have a very spiky network demand pattern, whereas agentic AI and physical AI have a high, sustained demand signal,” he said. “With agents operating at machine speed, the volume of traffic will be meaningfully higher than before. Humans click, but agents swarm.”

Patel also laid out a longer horizon, saying he foresees a near future where trillions of agents work together to assist humans in decision-making and task execution. He described those agents as especially consumptive, adding that they generate “450% more traffic than a human,” according to Patel.

That magnitude, Patel argued, should be viewed as a “networking supercycle” requiring massive upgrades—not just in data centers, but also in workplace and service provider infrastructure—to prepare for the surge in agent-driven traffic.

A cybersecurity supercycle moves at agent speed
Cisco’s messaging didn’t stop with performance. It folded directly into security, with Robbins warning that the same AI-driven acceleration is changing the pace of defense—and the pace of attack.

“AI is changing the speed of cyber defense, empowering our adversaries at a pace we have never seen,” Robbins said.

He gave a concrete example of how quickly attackers can act when they have AI-style speed. Cisco, he said, scanned 1.8 billion lines of code across 25 programming models in a matter of days. Robbins said that capability would have taken eight years in the past. and he warned that bad actors can access similar speed. In response, he said many Cisco customers are now “clamoring for an air-gapped, AI-ready, secure, and resilient infrastructure.”.

Liz Centoni, Cisco executive VP of Customer Experience, reinforced the urgency during the second day’s keynote. “40% of exploits directly impact end-of-life devices,” she said. Then she pointed to a newer problem: AI can map a network in minutes to find vulnerabilities “no one has ever spotted before.”.

Her description of risk was paired with a product pitch. Cisco’s show highlighted Cisco IQ, built into Cisco Cloud Control. The company said it lets IT teams and CISOs see their entire landscape with clarity. inventory every asset. understand current security status. and take automated remedial actions.

One argument stitched through all of it
Robbins and Patel described the move from compute-first upgrades to something broader—an infrastructure phase focused on connection. monitoring. security. and automation. Their core claim was that AI readiness is no longer limited to a data center or cloud checklist.

For IT teams. they framed the shift as a network architecture question: enterprises must be able to connect. observe. secure. and automate the systems that compute makes possible. Robbins said the first wave was about compute; the next one. in Cisco’s view. may be about whether enterprises can keep up when agentic AI raises both traffic levels and operational stakes.

Cisco Cisco Live agentic AI enterprise networks AI traffic cybersecurity air-gapped infrastructure Cisco IQ Cisco Cloud Control Chuck Robbins Jeetu Patel Brian Niccol Starbucks Liz Centoni