Agentic AI Is Driving the Next Expansion in Data Center CPUs

Dimitrios (Dimi) Ziakas, PhD, CTO & VP, Architecture

For the past several years, AI infrastructure discussions have focused almost entirely on GPUs. Accelerators remain essential for training and inference, but the rapid emergence of agentic AI workflows is changing the compute balance inside modern data centers.

Agentic systems do far more than execute a single model call. They plan, orchestrate tools, retrieve data from multiple sources, maintain memory across interactions, and coordinate multiple services across a distributed infrastructure. As a result, a growing portion of the AI pipeline now sits across both the GPU and the CPU. These workloads rely heavily on CPUs for orchestration, data preparation, memory management, storage interaction, networking services and overall system coordination.

In practice, this means the next phase of AI infrastructure requires balanced systems. GPUs remain the engines for model execution, with CPUs increasingly powering the surrounding control plane and data plane that allow agentic workflows to operate efficiently at scale.

At Hyve Solutions, we see this shift clearly in how customers are designing next-generation AI infrastructure. Organizations are building environments where accelerated compute and general-purpose compute work together. GPU platforms drive training and large-scale inference and CPU infrastructure handles orchestration layers, retrieval pipelines, storage services and other data-intensive tasks that enable production AI systems.

Hyve is well positioned for these systems requirements. Our experience delivering hyperscale server platforms and rack-level infrastructure allows us to support both sides of the AI compute equation. As a System Partner in the Nvidia Partner Network (NPN), we work closely with the Nvidia to deliver high-performance accelerated platforms while also providing robust CPU-based systems optimized for cloud-scale and data-intensive workloads.

As AI moves toward persistent, tool-using, multi-step systems, infrastructure requirements will continue to evolve. The future of AI data centers will not be defined by GPUs alone. It will be defined by well-architected systems that combine GPU acceleration with powerful, scalable CPU infrastructure.

That is the direction Hyve is helping customers build today.

‍