Versa Networks and Intel teamed to apply AI-powered security, networking, and analytics directly at the enterprise edge.

The collaboration combines Intel’s Xeon 6 processors – the company’s data center-focused central processing units (CPUs) — with Versa’s secure access service edge (SASE) networking security platform to allow organizations to run AI inference workloads closer to users, devices, and applications.

The joint architecture pairs the Xeon 6 processors, equipped with Advanced Matrix Extensions, with the VersaONE Universal SASE Platform and Versa Operating System. These technologies provide the compute and software foundation to support AI-driven workloads at the edge.

The focus is on enabling inference for a range of AI models – including traditional machine learning, deep neural networks, and small language models (SLMs) – directly within branch, campus, and other distributed environments.

Running these workloads locally can improve performance for applications like advanced threat protection (ATP), data loss prevention (DLP), and network analytics, where sending data back to a centralized cloud can introduce delays and increase risk exposure.

Intel said that AMX acceleration can improve throughput for these types of edge AI workloads compared to previous generations, although the companies did not disclose specific benchmarks.

The partnership is an extension of Versa’s “intelligent edge” strategy – an architecture that integrates networking, security, and AI into a single operational layer. The idea is to treat the network itself as a control plane for AI-driven operations, rather than a separate transport layer, as enterprises are increasingly looking to operationalize AI.

The collaboration comes at a time when companies are seeking to meet a growing convergence between networking and compute. As AI workloads become more distributed, vendors are increasingly integrating processing capabilities directly into networking platforms, blurring the lines between infrastructure layers.