How Kinaxis Brings GPU Power to Supply Chain Planning

Kinaxis has announced a key performance milestone for its Maestro platform as it pushes further into large-scale supply chain optimisation using GPU-accelerated computing.

The update centres on integrating NVIDIA’s cuOpt solver and AI infrastructure, enabling faster processing of increasingly complex planning models. As supply chains grow in scale and complexity, planning models increasingly need to reconcile tens of millions of variables across extended time horizons and multiple planning levels.

Companies are managing hundreds of thousands of SKUs across multiple regions while facing disruptions from trade policy, climate events, and demand volatility. Organisations are constrained not by visibility, but by how quickly they can run and iterate scenarios to respond to real-world changes.

While many organisations have adopted AI-driven visibility tools, the next frontier is computational speed. Traditionally, enterprise planning relied on batch-based optimisation that could take hours or even days.

Kinaxis’ integration of NVIDIA cuOpt signals a shift towards interactive, agent-driven planning that rivals advanced in-memory and cloud-native optimisation platforms.

Gelu Ticala, Chief Technology Officer at Kinaxis, says: “This milestone demonstrates how accelerated computing can change the way large-scale planning problems are solved.

“When optimisation shifts from hours to minutes, organisations gain the ability to iterate more frequently and evaluate more alternatives.

“That iterative speed is essential to enabling concurrent supply chain orchestration and advancing our agent-driven strategy.”

Kinaxis applied the enhanced platform to a large-scale semiconductor planning model comprising nearly 50 million decision variables.

The model spanned more than 40,000 SKUs across a six-quarter daily planning horizon, one of the most demanding use cases in industrial supply chain planning.

Results showed up to a 12-fold reduction in total end-to-end calculation time, with planning cycles dropping from more than three hours to around 17 minutes.

Core optimisation solve time improved by 23 times, cutting compute requirements by over 95% while maintaining comparable solution quality.

This marks a shift from traditional batch-based planning processes to more interactive, iterative workflows, allowing planners to evaluate multiple scenarios within operational decision windows.

The development aligns with Kinaxis’ broader strategy around concurrent supply chain orchestration, where decisions across planning levels are continuously updated as new data becomes available.

Faster optimisation enables more frequent recalculation, improving responsiveness to supplier delays, fluctuating demand, and production bottlenecks, directly reducing stockouts, excess inventory, and missed revenue opportunities.

Kinaxis supports more than 400 global enterprises, orchestrating decisions across over US$200bn in inventory and generating upwards of 250,000 scenarios each month.

GPU acceleration also underpins the company’s agent-driven workflows, which rely on running multiple optimisation scenarios simultaneously to evaluate alternatives.

By reducing solve times, more scenarios can be explored within practical decision windows, strengthening decision intelligence and adaptive planning.

Alex Fender, Director of Decision Intelligence at NVIDIA, says: “The increasing complexity of global supply chains demands a fundamental shift to accelerated decision-making.

“By integrating NVIDIA cuOpt into its Maestro platform, Kinaxis is empowering customers to achieve planning agility and scenario iteration to help navigate rapid change.”

Kinaxis and NVIDIA will co-present these advancements during NVIDIA GTC 2026, highlighting how GPU-accelerated optimisation is reshaping supply chain planning at enterprise scale.

As GPU acceleration becomes standard in supply chain planning, companies can run larger models, integrate real-time data, and explore predictive simulations, potentially redefining the speed, scale, and sophistication of enterprise decision-making.