The partnership integrates Kasm’s containerized workspace platform with Intel’s OpenVINO toolkit and Advanced Matrix Extensions (AMX). This architecture allows organizations to execute inference tasks—such as code assistance, document retrieval, and autonomous agents—directly on local hardware. By utilizing CPU-based acceleration, the system eliminates the reliance on GPUs for standard workloads, allowing companies to keep sensitive data within their own security perimeter.

Jaymes Davis, Chief Technical Evangelist at Kasm, noted that this approach provides a viable path for regulated industries like finance, healthcare, and defense to scale AI access across their entire workforce. Beyond CPU operations, the platform acts as a unified control plane capable of orchestrating workloads across NPUs and discrete GPUs. For high-demand tasks, Kasm 1.19 supports SR-IOV bifurcation of Intel Arc Pro cards, which divides a single physical GPU into multiple isolated virtual functions.

Economic feasibility remains a core pillar of the rollout. According to the company, the architecture reaches cost parity with standard per-seat AI subscriptions once a deployment hits approximately 40 users per node. As organizations move beyond this threshold, the operational costs become increasingly favorable compared to external SaaS alternatives, providing a predictable financial model for large-scale internal AI deployment.