Every year at Cisco Live U.S., I spend time with the people who keep the world running. The IT teams behind hospitals, power grids, banks, airlines, and governments. The operators who get paged at 2 a.m. The security teams racing against adversaries who never take a day off.

This year there’s more urgency than ever. The environment we’re operating in has changed faster in the last 18 months than in the previous decade. In that context, agentic AI has massive implications for every enterprise. When it comes right down to it, every agentic action is a routing challenge, a trust decision, and a telemetry event.

That’s what we’re focused on at Cisco.

We are entering a networking supercycle

We’re transitioning from the era of intelligent chatbots to one of agentic AI, where agents act autonomously on our behalf. This shift isn’t an incremental upgrade — it requires us to refactor our infrastructure entirely.

Most people are still modeling AI like software. It is not. A SaaS app waits for humans to click. Agents do not. They reason, act, and loop back…continuously, at machine speed, without pause. They generate sustained infrastructure demand and traffic volumes at unprecedented levels.

As a result, agentic AI isn’t just compute-bound or memory-bound. It’s network-bound in a way that nothing before it has been. Our latest research on AI traffic makes this concrete. Without agentic AI, we expected enterprise WAN traffic to grow roughly 2.5x over the next decade. With agentic AI, that number jumps to approximately 9x. And I think even that may prove conservative.

A single agentic task generates 450% more traffic than a human doing equivalent work. Roughly 70% of that is inference. And nearly 10% of AI flows now carry more data upstream than downstream — versus 0.5% for typical web traffic — because context continuously moves back into models.

That difference is creating a networking supercycle. And because the network is so central to all the AI infrastructure the world is building right now, we need a totally new way to manage our networks and the systems they connect.

Introducing Cisco Cloud Control

Today we are introducing Cisco Cloud Control — the first unified platform to bring every IT management experience across our portfolio into one operational environment. Compute, data center networking, workplace networking, security, observability, collaboration, third-party integrations: one login, one shared view. The design principle is simplicity without sacrificing sophistication.

Importantly, all the rich telemetry from across your operation (networking, observability, security, collaboration, and more) comes together in Cloud Control so that humans and agents are working from the same information across the full estate, without losing context as they move between tasks.

Cloud Control is more than a product. It represents what Cisco has become. Over the last two years, we rebuilt this company. We moved from a collection of individual products, each with its own interface and data model, to a platform where every product makes the others more powerful.

Like any true platform, Cloud Control works with your existing stack doesn’t ask you to rip out what you’ve already built. And it’s also designed for you to leverage an open ecosystem through what we call Cloud Control Studio, a customization engine, with two environments inside it.

There’s Agent Builder, which lets you build agents tailored to your own policies and workflows, connecting them to third-party tools including AWS, Google, Microsoft, and ServiceNow through native connectors or the open Model Context Protocol (MCP). And there’s App Builder, which lets you build and publish apps from natural-language prompts with built-in agentic coding assistants, starting with OpenAI Codex. Everything built in Studio can then be published to the Cloud Control Marketplace.

There is so much more to learn about Cloud Control. Dive in here and check out how we’re bringing new AgenticOps capabilities to some of the most important and common workplace networking workflows here.

Cloud Control is available now in the U.S. with global availability planned for later this year.

Mythos has entered the chat — and nothing will ever be the same

The only change that’s potentially more radical than the one we’re seeing in infrastructure is how AI is revolutionizing cybersecurity. When AI can find in hours what traditional scanning missed for decades, the old patch-and-pray model is finished. We need a whole new operating model for securing our infrastructure.

Cisco joined Project Glasswing from Anthropic and OpenAI’s Daybreak as a founding partner of both because we believe the only credible response is to be proactive: stress-testing our own products before adversaries do. We are using the most capable AI vulnerability discovery tools in the world to find and fix our own issues first. And we’re not keeping that advantage to ourselves. Through the recently open-sourced Foundry Security Spec, every defender can apply that same rigor to their own AI-driven security evaluations.

At Cisco Live, we shared how we’re working to create tools and defenses that match the new pace of attack, starting with Live Protect. It acts as a digital immune system for Cisco products, using compensating controls to shield their vulnerabilities from exploitation at runtime. No patch required, no reboot, no maintenance window.

Now available in N9000 series switches, Live Protect is expanding to more products in the Cisco Portfolio in the coming months, starting with campus and branch smart switches, and followed by secure routers later in the year.

Ultimately, Cisco Cloud Control is where this new mode of defense plays out in real time. It’s your command center for a post-Mythos world. One place to see your full vulnerability posture, manage your security policy, and deploy protections like Live Protect the moment they’re needed.

From Live Protect and Hybrid Mesh Firewall, to how you can protect and secure the AI agents that are already operating inside your organization today, check out all our security news here.

What a platform advantage actually means

What makes Cisco’s approach different isn’t any single product. It’s the architecture underneath all of them. We build our own silicon, design the systems it powers, write the software that runs on them, and provide the security and observability that watches over it all.

Every product we build is designed to make the others more powerful. The more of Cisco you run, the more the value of the platform compounds. You don’t have to run all of it.  Your environment is a mix of vendors and always will be. But our job is to make it work better, not to replace it.

That’s what we mean when we say platform advantage. And this week, you are seeing just how powerful it can be.