Komprise Secures Patent to Boost GPU Utilization for AI-Driven Unstructured Data Workloads

New updates have been reported about Komprise.

Komprise has secured a new U.S. patent (#US12,566,637) for its Elastic Shares technology, positioning the company to address a core cost issue in AI and unstructured data processing: underused GPUs, memory, and network capacity. The patented approach dynamically reallocates compute resources across large unstructured datasets, aiming for near-linear scale-out and materially higher utilization than traditional, static load balancing, which typically balances work only once at job start.

This innovation directly targets the budget pressure CIOs face as unstructured data grows more than 20% annually and AI infrastructure spending rises, by reducing waste in expensive GPU clusters that research shows can sit idle 14–76% of the time even during active jobs. For Komprise customers, Elastic Shares is designed to lower total cost of ownership for AI data pipelines, migrations, tiering, and sensitive data management while accelerating data mobilization into AI workflows. The new patent expands Komprise’s IP portfolio in analytics-driven unstructured data management, reinforcing its strategy to connect data lifecycle management with AI through its Intelligent Data Management platform, Smart Data Workflows, and Global Metadatabase, which together provide governed, context-rich access to petabyte-scale file and object data across hybrid environments.