Skip to content
‹ Back to IndustriesIndustrial IoT

Rockwell's Emulate3D adds NVIDIA Omniverse to test whole factories before startup

Rockwell Automation has built NVIDIA Omniverse, RTX rendering and OpenUSD into Emulate3D, stretching its virtual commissioning software from single machines to whole factories. Bigger models raise the stakes on scope. Plant engineering teams now have to decide how much of a factory is worth modeling, and who will build and maintain that model.

This story was produced through MarketScale. See how Industrial IoT teams put it to work with AI Visibility (GEO).

By MarketScale Newsroom · Rockwell AutomationEmulate3dNvidia OmniverseOpenusd
Share
Rockwell's Emulate3D adds NVIDIA Omniverse to test whole factories before startup

Key takeaways

01

Emulate3D's core job hasn't changed: it hands a real controller a convincing virtual plant so the logic gets debugged before startup. What changed is how much plant it can hand over.

02

Software is the small part of a digital twin’s cost; the bigger cost is the engineers who build and run it. With factory-scale models, staffing and scope decisions need to be budgeted alongside the license.

03

ABB has published performance claims for its Omniverse-based RobotStudio HyperReality. Treat those claims as a reminder to ask any simulation vendor for measurable, application-specific evidence—without assuming ABB’s results translate to factory-scale Emulate3D projects.

Free workspace

Turn your Industrial IoT expertise into content.

Record interviews, organize footage, and write with AI on a free trial of the MarketScale platform for qualifying companies. No demo required, no credit card.

Try it Free

Emulate3D exists to fool a controller. Engineers wire a PLC to a physics-based 3D model of the machinery. The controller sends commands, the model sends sensor signals back, and the logic gets debugged before any steel reaches the floor. On Nov. 16, Rockwell Automation said that model can now be a whole factory.

Rockwell has built NVIDIA Omniverse, NVIDIA RTX rendering and the OpenUSD scene framework into Emulate3D to support factory-scale visualization and virtual commissioning. The details come from the company's announcement and from Automation World, which reported it the same day. The pitch is earlier testing: model and tune production environments sooner in design, catch problems, and lower commissioning risk before equipment ships. Matheus Bulho, Rockwell's senior vice president of Software & Control, framed it around AI. He said the combination lets customers simulate larger, more complex systems and gives AI systems a place to learn and check decisions before deployment.

For the controls lead or plant project manager scoping a new line, that changes the planning question. If the tool can hold an entire plant, someone has to decide how much of the plant deserves a model, and who on the team will build it.

Three NVIDIA layers on a controls-testing core

Rockwell acquired Emulate3D in 2019, according to ARC Advisory Group, which has followed the product since. ARC describes its traditional use as simulating machines, material-handling systems and production environments to test automation logic. In a 2024 Automation World interview, Michael Sarvo, Rockwell's digital design business development manager for North America, drew a useful line. Simulation predicts theoretical performance early, while design choices are still open. The digital twin used for virtual commissioning is dimensionally accurate, physics-based and connected to the actual control system through its inputs and outputs.

Craig Resnick's ARC analysis splits the new additions by job. OpenUSD provides a common framework for describing and exchanging 3D assets and scene data. RTX adds accelerated rendering. Omniverse supplies the base for visualization and simulation workflows that can scale. Together, ARC says, they let teams evaluate production systems, controls, layouts and operating strategies in one model, where Emulate3D projects previously worked machine by machine.

Virtual commissioning connects simulated equipment to real or emulated control systems before installation. Teams use it to validate control logic, equipment interactions, sequences and layouts. Problems found this way are generally less disruptive and less costly to fix than problems found on the plant floor.

Rockwell also says the expanded platform produces deterministic outcomes. In plain terms, the same inputs should give the same result on every run. That repeatability is what lets an engineer compare one proposed change, AI-generated or not, against another with any confidence.

The release also fits into a larger toolchain. ARC notes that FactoryTalk Conductor, which Rockwell introduced in 2026, coordinates AI agents and engineering workflows across Studio 5000 Logix Designer, Logix Echo, Emulate3D and FactoryTalk Optix. That puts the factory model in the same workflow as the PLC program and the emulated controller. ARC also points out that Rockwell had previously announced plans to bring Omniverse into Emulate3D, so this week's news is that plan shipping as product.

What earlier Emulate3D projects paid back

Rockwell's case for virtual commissioning predates the NVIDIA work. In the same 2024 interview, Sarvo described RidgeTech Automation commissioning material-handling systems for two 100,000-square-foot vertical farms. The equipment included automatic storage and retrieval units more than 40 feet tall, machines with no room for error at first startup. RidgeTech proved out the logic in Emulate3D beforehand and ran the operator interface alongside the twin to settle client concerns before shipment. That shortened the initial commissioning window.

GA Pet Food Partners went a step further. It wired an Emulate3D twin of its new automation into its manufacturing execution system to test and tune the MES before go-live, and estimated the project would otherwise have run three to four months longer.

In this week's release, GE Appliances provides the customer view. Lora Green, its senior manager for Advanced Manufacturing Simulation, said virtual testing becomes more valuable as manufacturing systems grow more complex. In her account, better simulation lets her team evaluate ideas earlier and go into deployment with more confidence.

Both of Sarvo's examples were system-sized: one storage and conveyance package, and one automation upgrade tied to an MES. A factory-scale model suggests the next gains sit in the interactions between systems. For operators whose startup trouble tends to appear where separately supplied systems meet, a plant-wide model is where the extra fidelity would earn its keep. For a single-machine retrofit, the scope Emulate3D already covered may be enough.

ABB's robot-cell version of the bet

Rockwell isn't the only automation supplier building its models on NVIDIA. Design World's Puja Mitra reported in March that ABB Robotics was integrating Omniverse libraries into RobotStudio, its robot programming and simulation software, under the name RobotStudio HyperReality. ABB focuses on synthetic data for training physical AI models. Its virtual controller runs the same firmware as the physical robot, which tightens the match between simulated and real behavior. ABB also published performance claims for the tool.

ABB's published claims for RobotStudio HyperReality

Up to 80%
Reduction in setup and commissioning time ABB says the tool can deliver
Up to 40%
Cost reduction ABB attributes to needing fewer physical prototypes
50%
Shorter time to market ABB cites for complex products such as consumer electronics
8-15 mm to 0.5 mm
Positioning error before and after ABB's Absolute Accuracy technology, which keeps virtual and physical robot paths aligned

ABB Robotics, via Design World

Foxconn is piloting HyperReality on consumer electronics assembly. Many device variants and small, delicate metal parts make that work hard to set up, so Foxconn trains robots in a virtual environment on synthetic data covering different production scenarios before they reach the line. In March, ABB said select customers were testing the software ahead of a broader release to RobotStudio users in the second half of 2026.

The two launches approach the same problem from opposite ends: ABB from the robot cell outward, Rockwell from the control system and plant layout. Siemens announced plans in 2022 to connect its Xcelerator platform to Omniverse, Robotics 24/7 reported, so three large automation suppliers now feed models into the same NVIDIA layer. That could matter for plants running mixed equipment if OpenUSD assets actually move cleanly between vendor tools. Buyers should test that in a pilot before relying on it.

There's a limit to the comparison. ABB's percentages are ceilings for ABB's robot applications, and they say nothing directly about a factory-scale Emulate3D project. They show what kind of evidence to ask for. They are not a number to drop into a Rockwell business case.

Model scope and staffing decide payback

Bigger models are also where Rockwell's own guidance urges restraint. In 2024, Sarvo warned that digital twins put return on investment at risk when engineers go too deep. In his account, teams that build a twin of everything without a clear purpose end up with a model that is overly complex and costly to engineer, and that may still miss the insight it was built for. Software that can hold a full factory makes that temptation easier to act on.

Rockwell's own digital-design lead has argued that the software is the small part of a digital twin's cost; the engineers who can build and run one are the big part. Sarvo said twin modeling works when it makes up most of a person's job, as opposed to an occasional assignment, and that this takes real organizational commitment. For a plant adding factory-scale models to its project plans, that staffing decision belongs in the budget alongside the license.

Fidelity has a second limit. Sarvo's linear motor example makes it concrete: a commissioning twin needs the motor to respond to control signals and report its motion, while a study of the same motor's thermal behavior needs a different model entirely. A plant-scale commissioning model tells a team whether the logic and the handoffs work. It won't answer every engineering question about the equipment inside it.

Before scoping a factory-scale twin, list the decisions it has to settle, such as control logic, sequences, equipment interactions or layout, and model only to the fidelity those decisions need. Then name the person who will own the model after go-live.

The signal to watch is the first customer that puts a number on a plant-scale Emulate3D project, the way GA Pet Food Partners did with its three-to-four-month estimate for one automation upgrade. Until then, each plant has to run its own math: engineering hours spent on the model against startup days saved.

Featured companies

Your experts belong here

Every story in MarketScale Industrial IoT starts with a company putting its controls engineers, plant-floor specialists, and integration partners on the record. Buyers are already reading this topic. The only question is whose experts they find.

Plant and controls buyers research deep before contact, and your engineers get to shape that research.

Book DemoSee how it works15 minutes, straight to a calendar.

About the author

MarketScale Newsroom
MarketScale NewsroomEditorial Team, MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

B2B Weekly

The week in Industrial IoT, and sixteen other industries, every Monday.

Ten stories, one-line takes, five minutes. Free.

Industrial IoT: are you visible to AI?

Before they reach out, Industrial IoT buyers ask AI engines which vendors to trust. Explore how your experts, customers, and partners can become useful content for buyers and AI search.

Free Trial

You just read one Industrial IoT expert. Your company is full of them.

This article was produced through MarketScale. The same platform turns your controls engineers, plant-floor specialists, and integration partners into the articles, video, and social content Industrial IoT buyers are searching for. Start a free trial and see it with your own people. For qualifying companies, no credit card, no demo required.

NPS +73 · 1,000+ creators · 38+ countries

What your free trial includes

Hands-on access to the MarketScale platform
Media requests to your crowd, remote recording, AI writing tools
No demo required. No credit card.
For qualifying companies. Company confirmation required.

More Industrial IoT Insights

Hyungji Elite will tag 4 million 2027-season garments with Beontag RFID

Hyungji Elite will tag 4 million 2027-season garments with Beontag RFID

South Korean uniform maker Hyungji Elite will tag 4 million pieces of one brand's 2027-season merchandise with Beontag RFID tags. The near-term gain is automated inspections and inventory audits. Better demand forecasting comes later, and only if item-level reads carry through production, logistics and sales.

  • 01At 4 million pieces for one brand's season, Hyungji Elite is sizing its first RFID step at production volume, with every garment carrying its own identity from the factory onward.
  • 02Manual stock checks fail in two directions at once: missing items cost sales, phantom gaps trigger overstock. An RFID business case should measure both, not just count speed.

Nov 17, 2026

Tutor Intelligence raises $75M for robot work priced from $14 an hour

Tutor Intelligence raises $75M for robot work priced from $14 an hour

Tutor Intelligence raised a $75 million Series B led by Andreessen Horowitz, bringing its total funding to $117 million. Its Sonny and Cassie robots are billed by the hour or by the pick, and the company's pricing page lists palletizing at $14 to $18 an hour. With no upfront purchase, that model could let sites pay for robots from operating budgets instead of capital requests.

  • 01For stacking and unloading pallets, Tutor charges $14 to $18 per robot hour, with an 80-hour monthly minimum and a flat $10,000 implementation fee, according to Inc. Customers make no upfront purchase and sign no long-term contract.
  • 02Tutor's robots are working in almost four times as many facilities as a year ago, the company said. Nearly a third of organizations that received Tutor robots in the past 12 months were existing customers installing more robots.
  • 03Nearly a third of the organizations that got robots in the past year were existing customers adding to their fleets—a repeat rate worth watching alongside Tutor’s broader growth goals, which Dealroom notes are still targets rather than proof at scale.

Nov 17, 2026

Why Germany ranks third in robot density despite buying fewer robots

Why Germany ranks third in robot density despite buying fewer robots

Global robot density averaged a record 152 robots per 10,000 manufacturing workers, led by South Korea at 1,066. The ranking counts robots already in service. So Germany ranks third despite an 8% drop in purchases, while China ranks 19th after buying 59% of new robots.

  • 01Robot density counts robots already in service per 10,000 manufacturing workers, so Germany can rank third at 469 in a year its robot purchases fell 8%.
  • 02At the global average of 152, a 500-person plant works out to roughly 7.6 robots in service; at South Korea's 1,066, roughly 53. Density is one yardstick, but the article says plants should track new installations alongside totals to judge how current their fleets are.
  • 03Installations are forecast to reach 655,000 in 2026 and 806,000 by 2029, and new Korean automotive investments taking effect in 2027 and 2028 could push demand above its recent flat installation trend.

Nov 16, 2026

Explore More Industrial IoT Insights

Read more expert perspectives from across Industrial IoT.

Browse Industrial IoT Hub

About the Expert

MarketScale Newsroom
MarketScale Newsroom

Editorial Team

MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

For B2B teams

Your experts could be publishing here

Stories like this one run on content MarketScale captures from real practitioners. See how your team's expertise becomes coverage in Industrial IoT and beyond.

Book a Demo

Or call us. No forms required. We pick up. 214-945-2512