Körber Supply Chain, NVIDIA deal advances digital twin capabilities

Körber Supply Chain’s newly announced collaboration with NVIDIA is another sign that the push to bring digital twins to the fore in supply chain has moved beyond the theoretical stage for most businesses. The concept of building a digital twin for your supply chain is no longer something that only the largest organizations can accomplish, and that means the technology is now a true decision-making tool.  

The partnership centers on combining Körber’s logistics data and operational expertise with NVIDIA’s Omniverse platform to create accurate digital twins mirroring real-world warehouse and logistics operations with increasing precision. The goal is not just better visualization, but the ability to simulate, test, and optimize systems before they are ever deployed.

That ambition aligns closely with what industry leaders have been working toward for years. As Helena Garriga, president of Körber’s supply chain business, explains, the key enabler is not the concept itself, but the computing power now available to support it.

“The amount of data that you need to analyze and read per second requires a GPU power that is not the norm that we have in a computer,” she said in an interview with Supply Chain Management Review. The partnership with NVIDIA opens new doors to access the power necessary, she added.

From simulation to physical AI

What differentiates this next phase of digital twins is the emergence of what Körber and NVIDIA describe as “physical AI.” These are systems that don’t just analyze data, but interact with realistic, physics-based simulations of the physical world.

Historically, that level of capability has been out of reach, the companies noted in a release.

“The way the simulation was portrayed in the real world was not accurate at all for us to really use it at the customer base,” Garriga said.  

That gap is now beginning to close. Advances in GPU computing and simulation platforms are enabling organizations to incorporate environmental variables such as movement, positioning, system interactions, and physical constraints into their models.

“The amount of data you need to analyze to really have a simulated world that is the same as what we have here is huge,” Garriga said.  “The system will be able to simulate [the accuracy and speed of a supply chain] almost at the same time as it happens.”

The result is a digital environment that can increasingly mirror operations in near real time.

Expanding beyond design into execution

As these capabilities improve, digital twins are no longer confined to network design or long-term planning. They are starting to influence multiple stages of the supply chain lifecycle. One of the earliest and most practical applications is in the sales and solution design process. Instead of relying on static proposals, organizations can now simulate how a system will perform before it is built, Garriga said.

“What we are able to do now is simulate pretty accurately what the customer wants,” she said, noting that it allows the customer to how the solutions will perform before installing.

“It allows them to really understand what they are asking and bridge the gap between what they’re asking and what they really need when they see it running,” Garriga said.

The same approach is accelerating research and development. Instead of relying on physical testing cycles, engineers can iterate in virtual environments. The engineers can test online and see which changes were the right ones. Previously, somebody had to go into a lab and make those tweaks physically.

“That speeds up the R&D process by months,” Garriga said.

Reducing risk in live operations

Where digital twins may ultimately deliver the most value is in live operations. The ability to test changes, whether layout adjustments, throughput increases, or automation strategies, before deploying them reduces both risk and disruption.

This is particularly relevant in high-volume environments where even small disruptions can have outsized impacts. It also ties directly to cost reduction, especially in maintenance and downtime, Garriga said.

Adoption still depends on maturity

Despite the growing capabilities, adoption is far from uniform. Some organizations are still exploring digital twins conceptually, often driven by broader interest in AI. Others are embedding them into core workflows with clearly defined use cases. Garriga points to a common disconnect between interest and execution—the desire to incorporate AI and digital twins and what technology is actually needed to solve the problem.

More mature organizations are taking a different approach, starting with the problem and working backward to the technology, she said.

Technology is not the differentiator

The Körber–NVIDIA collaboration, though, is another step in the evolution of technology, moving simulation technology into the mainstream. Platforms are improving and computing costs are declining, enabling scalable capabilities.

“Thresholds are going down as we speak,” Garriga said. “It will be easier for simple setups to get there, where before that was not economically viable.”

But that doesn’t guarantee value, she said. As with any technology investment, Garriga advised clearly defining the problem. Technology is advancing quickly, but installing the wrong solution won’t be beneficial if it doesn’t solve the underlying problems.

As with any AI-driven initiative, success ultimately depends on clarity of purpose.

SC
MR