Enterprise AI is moving rapidly from piloting to production. As organizations deploy larger models and agentic AI systems, they quickly realize that AI is only as fast—and as secure—as the network it runs on. AI workloads demand massive bandwidth, low latency, and robust security that traditional networks struggle to provide.

To bridge this gap, infrastructure teams need solutions that deliver unprecedented scale, operational simplicity, and pervasive security. We are excited to showcase exactly how we are making this a reality at the ONUG AI Networking Summit Dallas, taking place May 13–14, 2026, in Frisco, Texas.

Here is a look at what you can expect from Cisco at the event and how we can help you build your AI-native enterprise.

Catch Cisco’s thought leaders on the main stage and beyond

The shift toward AI-ready infrastructure requires new frameworks and operational models. Our experts will be sharing actionable insights across several key sessions:

When Agents Run the Network: Security, Access, and the AI Data Center Ahead
Main Stage Keynote | May 13
The rise of AI agents changes how networks are built and secured. Tom Gillis, Cisco SVP & GM of the Infrastructure and Security Group, will deliver a forward-looking keynote on the AI-native data center. He will explore how networks must adapt for machine-to-machine intelligence and what security models are required when autonomous agents drive operations.

From AI Ambition to AI Infrastructure: Your Blueprint to Success for Enterprise Networks
Cisco Triple-T Session | May 13
Before you can run AI efficiently, you need a smart, secure network built to carry it. This session by Sai Natarajan, Sr. Director, Cisco Data Center product management, cuts through the complexity, offering a practical blueprint for deploying AI infrastructure without disrupting existing workloads. Learn how to simplify, future-proof, and supercharge your data center for the Agentic Era.

Hybrid AI Infrastructure and AgenticOps Security in Practice
Panel Session | May 14
Tom Gillis joins a panel of industry leaders to discuss building hybrid AI infrastructures that balance public AI services with private model deployments. Discover how aligned architectures and operational models create secure, adaptable environments for changing AI consumption patterns.

Defending Against Adversarial AI Agents – From Digital Darwinism to Guardrails
Panel Session | May 13
As AI agents become more autonomous, they introduce new attack vectors. Craig Connors, VP and CTO of the Cisco Security Business Group, will join this panel to discuss how adversarial agents change threat modeling and what defensive patterns, guardrails, and data controls are required to mitigate these emerging risks.

Innovations powering the AI enterprise

At ONUG, we will highlight the latest Cisco innovations designed to scale out and scale across for the agentic era, ensuring your transition to production AI is seamless and secure.

Unprecedented scale and simplicity

AI workloads generate traffic intensity that pushes conventional networks to the breaking point. We are answering this challenge with Cisco Nexus One, powered by the Silicon, Systems, Optics, Software, Operational Models, Security and Observability.

Furthermore, we are leveraging AgenticOps via AI Canvas, bringing AI-driven troubleshooting and optimization to your enterprise infrastructure using the Cisco Deep Network Model.

To ensure every AI job runs at peak efficiency, Cisco’s architecture brings together advanced AI job monitoring, intelligent mixed mode load balancing, and built-in GPU optimization across the fabric. Real-time monitoring provides visibility into job status and resource allocation, so infrastructure teams can identify and resolve congestion or bottlenecks before they impact performance. Mixed mode load balancing dynamically shifts workloads across network paths and clusters for optimal flow, minimizing latency and maximizing throughput—even during periods of unpredictable demand. Combined with deep integration into GPU orchestration, this enables organizations to boost compute utilization, speed up AI model convergence, and ultimately drive greater business value from every GPU investment.

The result: your AI workloads are executed more efficiently, reliably, and with agility, empowering you to move faster from proof-of-concept to impactful production deployments.

Deep, pervasive security for AI workloads

You cannot scale AI without securing it. Traditional security models often create bottlenecks or tax the very CPUs needed for AI processing. We are changing the game with a unified security approach embedded directly into the network.

Our Cisco N9300 Series Smart Switches fuse networking and security, supporting L4 segmentation to ensure policies follow the workload. Combined with Cisco Hypershield, organizations gain air-gapped, distributed segmentation across the fabric. To protect against vulnerabilities in real time, Cisco Live Protect instantly deploys eBPF-based shields without requiring reboots or downtime.

For the ultimate in server-level security, we are showcasing our integration of Cisco Hybrid Mesh Firewall with NVIDIA BlueField DPUs. By running stateful segmentation and firewall policies directly on the DPU, we deliver line-rate protection without compromising GPU or CPU performance. This ensures your AI front-end fabrics remain secure, isolated, and highly efficient.

Let’s build the future of AI together

The AI revolution is here, and the network is its foundation. Whether you need to scale massive GPU clusters, secure multi-tenant AI environments, or simplify operations with agentic automation, Cisco has the blueprint to get you there.

We look forward to being a part of the ONUG community and seeing you in Dallas. Be sure to attend our sessions and connect with our experts to discuss your AI networking and security strategies.

Explore how Cisco can help you build secure, scalable AI infrastructure today