Kinaxis Advances Large-Scale Supply Chain Optimization with NVIDIA AI

Executive Summary

• Kinaxis announced that GPU‑accelerated optimization using NVIDIA cuOpt™ delivers up to 12× faster end‑to‑end planning performance on a large‑scale semiconductor model.
• The improvement shrank total calculation time from >3 hours to ~17 minutes, with core solve time improving 23×, enabling interactive scenario iteration.
• Kinaxis will showcase the advancement at NVIDIA GTC 2026, highlighting its impact on the Maestro™ platform and the company’s agent‑driven orchestration strategy.

Key Details

• Tested on a semiconductor planning model containing ≈50 million decision variables (40,000 SKUs, six‑quarter horizon).
• End‑to‑end calculation time reduced by up to 12×; core optimization solve time cut by 23×, >95% reduction in compute time.
• Planning cycles now run in roughly 17 minutes versus previously >3 hours.
• GPU acceleration performed on NVIDIA AI infrastructure via NVIDIA cuOpt™.
• Faster solves enable planners to evaluate many more alternatives within operational decision windows, supporting Kinaxis’ concurrent supply‑chain orchestration and agent‑driven workflow.
• Kinaxis serves over 400 global enterprises, managing >$200 billion in inventory and generating >250,000 scenarios per month.
• The company will co‑present the results at NVIDIA GTC 2026.

Notable Quotes

• “This milestone demonstrates how accelerated computing can change the way large‑scale planning problems are solved,” – Gelu Ticala, Chief Technology Officer, Kinaxis.
• “The increasing complexity of global supply chains demands a fundamental shift to accelerated decision‑making,” – Alex Fender, Director of Decision Intelligence, NVIDIA.