AMD has selected Cisco to supply observability and governance capabilities for its new Ryzen AI Halo developer personal computers. The collaboration was announced at the Advancing AI event in San Francisco, with Cisco’s Splunk monitoring tools and Secure Network infrastructure integrated into the $4,000 compact devices.
The partnership targets organizations deploying these high-performance machines for intensive artificial intelligence and agentic workloads. Cisco will offer fleet-wide monitoring that tracks agent behavior, compute utilization, and token usage across the hardware. Additionally, the company’s open-source DefenseClaw security tool will enforce policy guardrails directly on the devices.
Cisco Chief Product Officer Jeetu Patel stated that the rise of distributed inference has made security a primary concern for customers. He explained that the integration allows enterprises to maintain enforcement guardrails during runtime if an agent begins to deviate from its assigned tasks or misbehave.
Patel also highlighted the financial implications of managing AI agents. He noted that enterprises need to monitor token consumption to prevent excessive costs. The unified management plane, powered by Cisco’s Cloud Control platform, provides visibility into the entire IT estate, including data centers and cloud environments, alongside the Halo fleet.
The announcement underscores the growing need for tools that can manage the complexity and cost of AI agents running on desktop hardware. Details regarding the specific timeline for widespread availability of these integrated features were not provided in the source material.