The Agentic Supercycle: Inside Cisco’s Blueprint for the Next Era of Enterprise Infrastructure

Lessons from Cisco Live 2026: Where Silicon, Security, and Autonomy Converge

The Cisco Live event was not only about key updates and product releases, but also about how Cisco has been positioning itself as an AI infrastructure provider. Throughout the event, Cisco advanced a consistent message: AI is no longer just a feature within infrastructure but a tool that runs the overall infrastructure autonomously. AI has moved from experimentation into mainstream enterprise deployment. However, it is only as powerful as the network it connects with. As Chuck Robbins, CEO and chairman of Cisco, said, every model, every inference, every agent ultimately runs on infrastructure. And that infrastructure is now being asked to do far more than it was ever designed to.

Robbins further mentioned that AI-associated network traffic is expected to triple over the next three years, driven by robotics, manufacturing automation, physical AI, and autonomous digital workflows. According to Cisco, more than 90% of enterprises acknowledge they must modernize their infrastructure to prepare for this transition.

Now the key question for every CIO, CTO, and CISO is: How should enterprises redesign their networks, security postures, and operating models to thrive in the agentic AI era where infrastructure demand is surging, cyber threats are striking, and trust must extend beyond humans to autonomous digital agents?

Scaling Networking Infrastructure for the Agentic Era

Cisco has claimed silicon as one of its most visible bets in 2026. It launched two AI networking chips, the G300 and P200. These are designed to power and scale AI data centers for high-demand training, inference, and agentic workloads.

• • G300 is designed to support ultra-high-performance AI clusters, providing the bandwidth and throughput needed as agentic AI applications drive exponential increases in east-west traffic within data This helps enterprises build scalable AI infrastructure that supports data-intensive workloads without creating network bottlenecks.
  • P200 enables geographically distributed AI It enables multiple data centers to operate as a single logical environment over long distances, helping enterprises overcome locality, power, and capacity constraints while supporting resilient, distributed AI deployments.

Avasant views this development as key for enterprises as it helps to modernize network infrastructure to efficiently support large-scale, distributed AI workloads and avoid network bottlenecks as AI adoption accelerates.

Reimagining Security for a Post-Mythos, Agentic World

To start with, we need to understand that Mythos is a generative AI model from Anthropic, built for complex, multistep reasoning tasks that run autonomously over extended periods. It can locate bugs, analyze network systems, and generate exploits faster than human researchers. However, after its launch, a data leak, caused by company staffers inadvertently exposing material about the LLM, has raised concerns about the model’s potential implications for cybersecurity.

According to data cited in the Avasant Worldwide IT Spending and Staffing Outlook for 2024, security and privacy are top-of-mind issues for most enterprises. Cybersecurity threats have been ranked as the greatest concern, with 102 of nearly 200 responses, followed by privacy, which comes in second with 87 of the top three votes.

At Cisco Live 2026, the company also unveiled certain strategic portfolio enhancements that define its post-Mythos incident security and infrastructure vision:

1. 1. Vulnerability management: The company launched Live Protect, a digital immune system for Cisco products. It uses runtime security controls to mitigate vulnerabilities without requiring immediate patching, reboots, or service It enables teams to reduce risk while maintaining Cisco infrastructure operations and completing permanent remediation through normal patch and change-control processes. It is available in Cisco Nexus One orchestrated through Nexus Dashboard for Cisco’s data center and cloud networks, giving security and infrastructure teams centralized control over their network devices and vulnerability statuses.
   2. Firewall: Since 2025, Cisco has been offering its Hybrid Mesh Firewall that unifies distributed enforcement points, including firewalls, smart switches, agents, AI guardrails, and third-party firewalls across customers’ data centers, branches, clouds, campuses, and IoT environments. It enables customers to protect the development and deployment of AI models, shield vulnerabilities, and stop unauthorized lateral movement.

Further to this they have also made some new enhancements to its firewall offerings, that were announced during their Live event, which are as follows:

• • Launched Cisco Firewall 10.1, an advanced threat protection for encrypted traffic and expanded support for virtual firewall instances across hybrid environments.
  • Providing direct integration between Cisco firewalls and Splunk through built-in Splunk integration wizard across its firewall management center 10.0 version or later to stream logs securely via syslog. This enables seamless export of advanced, Zeek-like logs to improve network visibility and accelerate threat prevention through pre-built detections.
  • Expanded its Cisco Secure Firewall 6100 Series by introducing the 6120 and 6130 models to support scalable, AI-ready data center and enterprise security.

Reinventing Operations Through Cisco Cloud Control

During its Cisco Live 2026 event, the company also launched Cisco Cloud Control, a unified AI-native operations platform designed for humans and AI agents to manage networks, security, and data centers under a single command center. This includes Cisco Security Cloud Control, which covers entire IT infrastructure.

While Cisco Cloud Control serves as the AI-native operational brain for the enterprise, Cisco is bringing the same agentic operating model to the workplace, enabling human and AI agents to collaboratively diagnose, remediate, and optimize employee experiences through AI Canvas and autonomous workflows.

AI Canvas controlled availability was also announced at Cisco Live 2026 as a part of Cisco Cloud Control. AI Canvas is a multiplayer agentic workspace where IT teams and AI agents investigate and resolve issues across every domain. It merges real-time telemetry and cross-team collaboration into a single, interactive workspace to accelerate incident investigation and response. AI Canvas utilizes Cisco’s specialized Deep Network Model, an LLM designed specifically for network operations management, to understand complex IT infrastructure. Anurag Dhingra, senior vice president and general manager of the Enterprise Connectivity and Collaboration group, demonstrated how, through Cisco Cloud Control, an operator will be able to work with autonomous agents that follow a structured path from signal to action: spotting trouble, identifying causes, carrying out fixes, testing changes before deployment, and confirming the user experience has recovered.

According to Avasant, the platform’s greatest value lies not in unified management but in its ability to automate cross-domain operations, accelerate troubleshooting, enable AI-driven remediation, and establish the operational foundation needed to support agentic AI at scale.

Elevating Digital Resilience with Cisco IQ and Splunk

During Cisco Live 2026, Liz Centoni, executive vice president and chief customer experience officer, showcased Cisco IQ, an AI-powered digital experience platform that unifies Cisco’s support and professional services into a single interface. It provides IT teams with real-time insights, predictive asset tracking, and proactive troubleshooting to prevent outages and optimize network infrastructure.

To exemplify this, GEODIS, a transport and logistics firm, showcased how it leverages Cisco IQ to shift from reactive network management to proactive, AI-driven infrastructure visibility, which is critical for maintaining uptime in logistics operations. Cisco IQ provided real-time, end-to-end visibility across GEODIS’ global network estate, enabling precise tracking of hardware health, life cycle status (especially end-of-support risks), and vulnerabilities exposed by emerging AI-driven threats such as Mythos. This visibility is complemented by AI-assisted insights and planning capabilities, allowing GEODIS to identify at-risk devices, prioritize refresh cycles, and create actionable, site-level remediation road maps without disrupting operations.

In addition to this, Cisco leverages Splunk capabilities, and since its acquisition in March 2024, Cisco has continued investing in augmenting its security capabilities, focusing on delivering advanced security and AI-powered data analytics. Splunk’s Senior Vice President and General Manager Kamal Hathi highlighted how Splunk is playing a key role across Cisco Cloud Control and Cisco Data Fabric.

• • Data unification: Launched in September 2025, Cisco Data Fabric, powered by the Splunk platform, is an architecture designed to unify, contextualize, and analyze volumes of enterprise machine Its purpose is to correlate telemetry to drive action, providing a single unified data layer required for cross-domain observability, agentic operations, and machine-speed response.
  • Cross-domain observability: Splunk also enables AI-driven root cause analysis across infrastructure and applications delivered through AI Canvas.
  • Agentic security: Splunk supplies the unified data spine, agentic analytics, and AI observability that convert Cisco’s infrastructure into an autonomous, self-defending, self-governing security operating system.

Avasant views this as a key development by Cisco, with Cisco IQ and Splunk collectively signaling a structural reset in enterprise IT operations. Together, they enable enterprises to continuously monitor their systems, resolve issues with precision, automate incident response, and measure digital trust.

AI Observability and Tokenomics

Jeetu Patel, president and chief product officer at Cisco, has described observability as a core pillar of the company’s platform stack and mentioned that traditional infrastructure monitoring is no longer sufficient in the agentic era. To further exemplify this, Patel also cited Cisco’s acquisition of Galileo (April 2026), an AI observability and evaluation company for multi-agent systems, working as a core engine for assessing agent behavior within Cisco Cloud Control. Galileo has also strengthened Cisco’s Splunk Observability portfolio and its AI agent monitoring capabilities across Splunk Observability Cloud, providing customers with real-time visibility into the full agent development life cycle.

Additionally, for tokenomics, Cisco has built an agent observability and tokenomics app (jointly powered by Galileo and Splunk), delivered natively inside Cloud Control. This app provides enterprises with visibility into every agent running across the environment and the tokens each agent consumes.

In April 2026, Cisco also introduced Luna, a family of small language models, delivering the same evaluation accuracy as large LLMs at 95% lower cost and millisecond latency. For enterprises running millions of agent transactions daily, this makes continuous agent evaluation economically viable at scale.

Avasant views tokenomics observability as the next strategic inflection for enterprises. As agents scale, cost, trust, and control converge at the agent level, making per-agent visibility and 100% evaluation coverage non-negotiable pillars of any AI operating model.

Conclusion: Architecting the Infrastructure for the Agentic Era

Cisco Live 2026 did not merely announce new product innovations but more around how enterprise AI readiness will increasingly depend on the strength of the underlying infrastructure layer. Cisco has marked a broader shift in how enterprises must think about AI infrastructure. As AI agents become active participants across enterprise operations, infrastructure must evolve from a collection of connected systems into a unified platform that can securely operate, govern, and optimize AI-driven workflows. Cisco’s strategy spanning AI networking, runtime security, AI-native operations, data unification, observability, and tokenomics reflects this transition toward autonomous infrastructure.

By Gaurav Dewan, Research Director and Siddharth Mehta, Research Leader