How o9 and NVIDIA are Driving Faster Supply Chain Planning

o9 is collaborating with NVIDIA to ensure faster solve times for its supply chain operations.

In today's complex world, solutions need to be fast and efficient at simplifying workflows, with agility when operating amid volatility.

By embedding optimisation and demand planning into operations, businesses can see greater time and cost efficiency, as well as better resilience when faced with ongoing disruption.

Technical collaboration

Around the world, o9 is using its Digital Brain platform to drive enterprise planning. To transform this workflow even more, it is collaborating with NVIDIA by applying NVIDIA's GPU-accelerated optimisation solver, NVIDIA cuOpt, on its platform.

When NVIDIA cuOpt is integrated into the o9 platform, as well as being run on the NVIDIA Blackwell computing infrastructure, the companies have found more than 10x faster solve times for large-scale supply chain challenges.

To test the efficiency of this collaboration, the trial was conducted on a supply chain planning dataset, made up of approximately 30 million variables, 15.7 million constraints and 94 million nonzeros.

This reflects the complexity and the scale of large-enterprise inventory optimisation and production planning workloads. 

By trialling it on this dataset, o9 and NVIDIA were able to test the legitimacy of their collaboration, finding out whether it really did transform efficiency, and to what scale.

"Solve time directly limits how many scenarios a planning team can evaluate in a given planning window, especially for large-scale LP problems with tens of millions of variables and constraints," Ashwin Rao, Chief Technology Officer at o9, explains to Supply Chain Digital.

"But speed alone isn't the full story — a fast system can still recommend a replenishment quantity that violates a contractual minimum order, or mistake correlation for causation, if it lacks structured reasoning about how decisions connect across the business.

"That's why o9 pairs GPU-accelerated solve speed with a neurosymbolic architecture, where the "middle 80%" traces plans, forecasts and inventory positions to identify the actual cause behind a number, not just a fast answer. The result is optimisation that's both faster and more trustworthy for decisions with real financial consequences."

Increasing efficiency

The trial demonstrated an optimisation solve in 57.4 seconds, meeting optimal solution status. A CPU-based LP solver required 661.7 seconds to solve the same problem – this reflects a 10x reduction in solve time. By reducing an almost-11-minute compute time to just under one minute shows the levels of efficiency that this partnership can introduce. 

Despite this immense reduction of time-taken, the objective value remained within 0.008% of the CPU solver's optimal solution, demonstrating no sacrifice of solution quality. When a solution can operate more efficiently, with the same level of high quality, it drives cost savings and greater worker productivity. 

By minimising time-taken for each solution, there is much less idle-time, therefore less time for customers to wait. Workers become much more efficient and productive. Clients across a range of industries are deploying these solutions in order to solve problems and grow to scale, without the need to invest in other platforms or tools. 

"The world of optimisation is evolving rapidly, and our Digital Brain platform for planning and decision-making is built to evolve with it," says Chakri Gottemukkala, Co-Founder, CEO and Chairman at o9.

"Demonstrating a 10x-plus speedup on a real, 30-million variable supply chain LP—with best-in-class solution quality—is a meaningful proof point for what GPU-accelerated planning can deliver. Our architecture is uniquely positioned to allow o9 enterprise clients to immediately benefit from our collaboration with NVIDIA."

Growing resilience

Optimisation in planning helps transform large enterprises, solving a range of issues that prevent them from meeting demand. The increase in planning agility also helps organisations avoid risk – teams can enact more scenario planning, respond to disruptions faster and be more resilient. 

"Faster solves let planning teams iterate multiple times in a day instead of relying on overnight batch runs, meaning they can respond to disruptions as they emerge rather than waiting for the next scheduled cycle," adds Ashwin.

"While no system can reliably predict every black swan event like a port strike or geopolitical shock, what matters is how quickly the organisation can understand what changed and decide what to do next. In practice, this pairs neural AI interpreting incoming signals with a symbolic layer connecting them to the inventory, capacity and financial outcomes already modeled in the business.

"Speed, in other words, is what makes real-time disruption response possible, but it's the underlying reasoning architecture that makes the response trustworthy enough to act on."

The collaboration between o9 and NVIDIA is working to transform efficiency throughout organisations, making operations more time-efficient and cost-efficient. Through the demand planning capabilities that are being expanded on, organisations are also more resilient amid ongoing geopolitical turbulence.