Dell’Oro Group Finds Agentic AI and Inference Drive Front-End Switch Growth
Dell’Oro Group reports that agentic AI and inference workloads are increasing demand for front-end data center switching, with more than half of near-term front-end switch growth tied to AI-related opportunities. The shift matters to network planners because inference and orchestration workloads change how front-end capacity is sized and deployed.
Market Overview
The report links the demand increase to a transition from large-scale training to inference and agentic workflows, which it says create network requirements different from those tied to training workloads. It projects that front-end data center switch sales growth over the next few years will include net-new expansion for both established vendors and new market entrants.
Key Findings
Dell’Oro Group estimates that more than half of growth in Front-end Data Center Switch sales over the next few years will come from AI-related opportunities. It also forecasts adoption of 800 Gbps and 1.6 Tbps in front-end networks driven by AI-related build-outs.
Technology and Workload Implications
Sameh Boujelbene said, “As AI infrastructure shifts from large-scale training to inference and agentic workflows, there is an increasing demand for general-purpose infrastructure, and expanded front-end network requirements.” He added that the “traditional assumption of a 10-to-1 ratio of XPU to CPU no longer applies across all deployments, with some environments moving closer to a 1-to-1 ratio.”
On compute and data movement, Boujelbene said, “CPUs are becoming increasingly important for workloads orchestration and data movement.” He also noted that “networking for KV caching storage rack is also needed for inferencing applications.”
Supplier and Deployment Beneficiaries
The report names Accton, Arista Networks, Celestica, Cisco, HPE/Juniper, H3C, Huawei, NVIDIA and others as primary beneficiaries of the forecast growth. It frames these vendors as capturing demand from the AI-related front-end network build-outs.
This Analyst Signals brief reflects a neutral, fact-based summary of the original research note.