Procore expands NVIDIA collaboration on construction digital twins

Procore Technologies used its Groundbreak 2026 conference in Orlando on October 21 to announce the next phase of its collaboration with NVIDIA. The new digital twin capabilities, now in closed beta, let teams switch from Procore's Model Manager to Realism Mode, which runs on NVIDIA Omniverse libraries. According to Procore's announcement, teams can view projects in a highly realistic 3D environment and see how real-world conditions such as extreme weather could affect construction before they happen on the jobsite. The stated payoff is simple: spot the problem early enough to adjust before it becomes costly rework, delays or schedule disruption.

Plenty of software can render a building. For a project executive or a scheduler, what counts is the data the render is connected to. Procore links this one to procurement and schedule data, and to imagery from DroneDeploy, the reality-capture company it acquired.

Three uses, three data feeds

Procore groups the new capabilities into three uses. Each one draws on a different stream of project information.

• Design against reality: DroneDeploy's aerial and ground-based captures bring actual jobsite conditions into the twin, so teams can compare progress with the design and spot discrepancies before they compound.
• Schedule ripple: 3D models linked to procurement and schedule information show how a delayed material delivery or equipment installation could affect downstream work, leaving time to resequence and protect critical milestones.
• Pre-construction scenarios: teams can simulate extreme weather, sun exposure and the movement of crews and equipment to find conflicts before they cause delays.

The second item matters most to whoever runs project controls.

Here is why. The other two uses answer questions about the physical site: does the work match the design, and what happens if the weather turns or two crews want the same patch of ground. The schedule ripple use is the only one that ties the model to commitments with dates attached. When a delivery slips, the value lies in seeing which downstream activities move with it while there is still time to resequence.

Procore says the collaboration pushes forward its system of intelligence, which is built so construction teams can see, understand and act on project information that is connected. DroneDeploy widens the view by combining aerial and ground-based imagery with project data, showing what is really happening on the jobsite. The Omniverse-powered digital twin capabilities help teams understand how conditions and project changes could affect construction. Procore AI then converts that intelligence into action, so teams can identify risks, make informed decisions and resolve potential issues before they turn into costly delays.

Read as a chain, that framing puts the twin in the middle. It takes field imagery in on one side and hands risk signals to Procore AI on the other, which suggests the simulation is only as useful as the captures and records feeding it.

Aimed at the data center build cycle

Procore frames the launch against an infrastructure build cycle it describes as among the largest in history, driven by trillions of dollars going into AI, data centers and energy systems. It cites McKinsey's projection that companies will invest nearly $7 trillion in data center infrastructure alone by 2030. Procore argues that limits on construction capacity and the workforce could slow delivery of those facilities, so builders will need to plan better, not only build faster.

The company says what "better" means: more efficient building, less waste, better use of resources, and more planning and predictability on complex projects. That list leans toward the planning office rather than the field. If labor is the constraint, the gains Procore is selling come from wasting less of the capacity a contractor already has.

NVIDIA described it in similar terms. Vladimir Troy, NVIDIA's vice president of AI Infrastructure, said thousands of decisions across power, cooling, compute, networking and physical infrastructure go into building an AI factory, and that Omniverse lets teams simulate those systems and find problems before they become costly field changes. He added that by linking simulation to project data, the two companies aim to cut risk and speed the path from design to operation.

Steve Davis, Procore's president of product and technology, said he expects the most resilient megaprojects will be built digitally first. As he described it, the collaboration lets teams visualize projects, simulate real-world conditions and test variables "before a single shovel hits the ground."

The early customer voice comes from Switch, a data center designer, builder and operator that uses both Procore and NVIDIA. Corine Winfield, Switch's vice president of data operations, said that seeing a project in 3D gives its teams greater clarity into construction progress, helps them make better-informed decisions earlier, and keeps complex initiatives on schedule.

That profile suggests where the tool fits best: equipment-dense buildings on compressed schedules, where a single late delivery could hold up several trades. For contractors whose backlog leans toward data centers or energy work, the beta is worth tracking now. For a builder on simpler, less equipment-heavy jobs, the weather and sun-exposure simulations may matter more than the procurement link.

Where Procore fits in NVIDIA's construction push

NVIDIA's construction work goes beyond Procore, and it has been building for years. For Construction Pros listed among its top construction technology stories of April 2021 that Bentley Systems was bringing infrastructure digital twins to NVIDIA Omniverse, developing applications for photorealistic, real-time visualization and simulation of massive-scale industrial and civil infrastructure projects.

The effort has since reached the machines. Construction Equipment Guide correspondent Lucy Perry reported in August 2026 that Caterpillar is partnering with NVIDIA to advance integration and automation into machinery, with the intent to accelerate project execution and reduce costs. The same report noted that the construction labor shortage is not going away and that the workers the industry counts on are aging.

NVIDIA has also been part of the industry conversation on what comes next. Sean Young, its director of AECO, geospatial and AI solutions industry marketing, joined leaders from Autodesk, Hensel Phelps and Associated Builders and Contractors on a November 2025 Construction Executive webinar about the jobsite of the future, covering drones, AI, autonomous vehicles, robotics and data.

Procore's version comes from a different direction. Bentley's work centered on modeling infrastructure, and Caterpillar's on the equipment itself. Procore sits on the project management record: schedules, procurement and now drone captures. Connecting Omniverse simulation to that record suggests NVIDIA's construction work could reach the everyday project data that general contractors run on.

That same dependency is the limit. The twin can only flag a late delivery if the delivery date in the system is current. On a team that tracks procurement in spreadsheets outside Procore, the twin would still render the 3D model, but its schedule ripple analysis would have far less to work from.

Commissioning, Maximo and the handover

The integration remains in closed beta. Investing.com's account of the announcement adds that Model Manager and Commissioning are also moving through beta, along with a planned integration with IBM Maximo.

For owners, the Commissioning work and planned Maximo integration could mean the twin stays useful after handover. That would line up with Troy's "design to operation" framing, carrying the model past the punch list and into the building's working life. The thing to watch is what happens when the beta opens up. If early users can show they protected milestones because a delay showed up in the model first, owners could start asking for it in their requirements.