Industrial automation is a notoriously capital-intensive sector where mistakes in physical planning can cost millions of dollars and months of lost production. At the NVIDIA GTC 2026 conference, Rockwell Automation presented a software-driven solution to that physical problem. The company showcased how physics-based digital twins are moving out of siloed engineering departments and into accessible, collaborative environments.
Rockwell Automation is not a software startup testing new waters. With roots tracing back to 1903 as the Compression Rheostat Company, the Milwaukee-based giant now employs roughly 27,000 people and serves customers in more than 100 countries. However, its recent presentations highlight a deliberate pivot toward simulation and cloud infrastructure. John Pritchard, who leads the digital twin software business at Rockwell, explained in an interview at the event that these virtual models are designed to help manufacturers plan new facilities or upgrade existing infrastructure before any steel is cut.
According to Pritchard, the business case for this level of simulation is straightforward and measurable. He noted that utilizing physics-based digital twins has a massive impact on the timeline of heavy industrial builds. “It will typically give you 50 percent reduction in the time and the cost of delivering these projects,” Pritchard said.
From Silos to Web Browsers
The core of Rockwell’s presentation at GTC centered on how these digital twins are delivered to clients. Historically, reviewing high-fidelity industrial models required specialized workstations and heavy, proprietary software. By integrating technologies from Microsoft Azure and NVIDIA, Rockwell has fundamentally changed the accessibility of its simulations.
The company is leveraging cloud compute and application programming interfaces to stream complex environments directly to standard devices. Pritchard emphasized that this shift eliminates the traditional bottlenecks of client reviews. “When we have teams of engineers collaborating on projects, we can now deliver a client review experience to a web browser,” Pritchard said. He added that a client can quickly understand a project’s status simply by clicking a URL, allowing distributed teams to evaluate milestones without installing dedicated software.
Immersive Design Reviews
Beyond browser-based collaboration, Rockwell is integrating extended reality to give manufacturers a spatial understanding of their future facilities. While most engineering work remains on standard monitors, Pritchard noted that the ability to load a model into a virtual reality headset transforms the review process. Customers can walk onto their actual, empty factory floor and place a full-scale digital model of the machinery into the physical space.
This immersive approach bridges the gap between technical blueprints and practical reality. It allows executives and facility managers to visualize the exact footprint and ergonomics of a production line. “As soon as you can see it, people understand it,” Pritchard said. “As soon as they can understand it, they can ask meaningful questions.”
A Jumpstart on Workforce Training
The utility of the digital twin extends past the design and construction phases. One of the most practical applications Rockwell highlighted is workforce readiness. A persistent challenge in manufacturing is the downtime required to train operators on newly installed equipment.
Rockwell is using its virtual models to solve this operational bottleneck. Because the physics-based digital twins accurately reflect the behavior of the real machinery, factory workers can use them as training simulators. Pritchard noted that manufacturers can take the operators who will eventually run the equipment and introduce them to the model early in the process. “They can start to actually get a jump on the training,” Pritchard said, adding that workers can learn how to interact with the machines before the hardware even arrives on site.
The Hardware Reality
Despite the seamless browser integration and virtual reality capabilities, powering these industrial simulations requires immense computational resources. Rockwell has consistently expanded its relationship with NVIDIA since early 2024 to support the integration of artificial intelligence and advanced rendering in its software stack. However, the localized hardware demands remain significant.
Pritchard was direct about the infrastructure required to run high-fidelity simulations locally, acknowledging that the software “inconveniences quite a few electrons.” While cloud solutions are becoming more viable, many manufacturers still prefer localized data control. “Customers either have to invest in on-prem compute and actually also typically a pretty powerful NVIDIA GPU,” Pritchard said. He stressed that adopting this technology requires both capital investment and dedicated time for human workers to master the tools, but argued that the long-term efficiency gains always justify the initial friction.
#rockwellautomation #robotics #NVIDIAGTC
