Cisco Systems and AMD have expanded their corporate technology partnership by introducing a joint hardware and security software package designed to protect, deploy, and manage distributed artificial intelligence resources at enterprise scale. The collaboration pairs AMD’s high-performance Ryzen AI Halo architecture with a suite of Cisco networking, observability, and governance tools, targeting the operational challenges of running AI workloads locally on employee devices.
The new framework was detailed during AMD’s Advancing AI event. Jeetu Patel, Cisco’s president and chief product officer, spoke at the gathering about how artificial intelligence inference is becoming widely distributed across enterprise networks, requiring a cohesive software and hardware stack developed jointly by the two companies.
The integrated architecture builds on AMD’s Ryzen AI Halo hardware platform, which is designed to support local AI inference on AI personal computers utilizing a central processing unit, a graphics processing unit, and an XDNA neural processing unit. According to AMD statements, resilient local AI platforms must maintain functionality even when network connectivity is limited, models undergo changes, or enterprise workloads shift.
“AMD provides the deskside/local AI platform. At the foundation is AMD Ryzen AI Halo hardware, an isolated agent sandbox and the services needed for local-first inferencing, including model routing and token limits via AMD’s Semantic Router and local inference on Lemonade,” wrote Yash Sheth, Cisco’s senior director, engineering and research, in a corporate blog post outlining the deployment package.
Enterprise Security and Observability for Deskside AI
To secure local processing nodes, Cisco integrates several proprietary management and defense layers directly onto the hardware. The security harness includes Splunk Agent Observability and Splunk Infrastructure Monitoring, giving enterprise IT departments full-stack visibility to track agent behavior, tokenomics, and compute operations.
Additionally, Cisco incorporates its AI Defense tool for model and agent security, alongside DefenseClaw for on-device security policy enforcement. These guardrails operate directly within the agent harness on the local machine. Cisco Cloud Control rounds out the package by providing a single pane of glass for unified network policy and administrative control.
“To make deskside and local AI computing work at enterprise scale, every AI node must be treated as a secure, managed node in the enterprise network,” Sheth wrote, emphasizing that token efficiency and data sovereignty are driving organizations to deploy ambient AI agents directly alongside workers.
Industry Shifts Toward Edge and Inference Workloads
During the Advancing AI keynote presentation, AMD leadership outlined broader macroeconomic and technological trends shaping the semiconductor and enterprise software sectors. Industry adoption of artificial intelligence continues to accelerate across all verticals, with agentic AI driving a massive surge in compute demand.
Workloads are increasingly shifting away from model training toward inference, which is projected to account for 60 percent of global AI compute capacity by 2026. At the same time, computing infrastructure is expanding beyond centralized cloud data centers, positioning edge devices and personal computers as critical nodes for real-time distributed intelligence.
Jack Huynh, senior vice president and general manager, computing and graphics group at AMD, noted that pairing Ryzen AI Halo systems with Cisco infrastructure allows enterprise customers to deploy scalable AI systems that remain secure, performant, and observable.
“Running more AI locally can help improve responsiveness, keep sensitive data closer to users, and reduce dependence on cloud-only approaches, but enterprises also need a way to monitor and manage these systems at scale,” AMD stated regarding the collaborative initiative.
Market projections shared during the corporate event indicate that the global AI accelerator market could reach $1.4 trillion by 2030, nearly tripling prior forecasts. While graphics processors are expected to dominate that expansion, CPUs are finding new growth vectors driven by the operational demands of agentic AI. Concurrently, the server CPU market is forecasted to grow by more than 50 percent to reach $200 billion by 2030, supported by the growing need for robust enterprise infrastructure.
Organizations evaluating enterprise AI deployments can monitor upcoming updates through official corporate announcements from Cisco and AMD. Readers are encouraged to share their thoughts or join the conversation in the comments below.
Keep reading