Transforming Manufacturing with AI-Powered 3D Digital Twins and Remote Monitoring – with Rad Desiraju of Microsoft and Mike Geyer of NVIDIA

This interview analysis is sponsored by Microsoft and NVIDIA. It was written, edited, and published in alignment with our Emerj sponsored content guidelines. Learn more about our thought leadership and content creation services on our Emerj Media Services page.

Manufacturers worldwide are under increasing pressure to enhance operational efficiency and agility in response to evolving market demands.

According to the U.S. National Institute of Standards and Technology (NIST), manufacturers are increasingly relying on operational dashboards to monitor key performance indicators (KPIs) in real time, enabling proactive maintenance, throughput optimization, and quality control. Dashboards are central to smart manufacturing initiatives, providing visibility across disparate systems and improving responsiveness at the shop floor level.

McKinsey estimates that digital transformation in manufacturing can improve labor productivity by 15% to 30% This productivity boost may be effected through increased automation of manual tasks, enhanced operational transparency, and AI-supported decision-making, among other outcomes.

Yet, the shift from manual to digitalized systems presents new challenges. Transitioning to simulation dashboards requires standardized data inputs, scalable infrastructure, and interoperable systems that can support real-time edge computing. 

Compounding these challenges is the fact that many manufacturers face persistent data barriers. Research from data integrity leader Precisely, in partnership with the Center for Applied AI and Business Analytics at Drexel University’s LeBow College of Business (Drexel LeBow), found that only 12% of organizations report having data of sufficient quality and accessibility to support effective AI implementation.

Meanwhile, 64% cited data quality as their top data integrity challenge — up significantly from 50% in 2023. These widespread gaps in data requirements versus integrity and accessibility underscore the difficulty manufacturers currently face in driving accelerated and successful AI integration.

Emerj Editorial Director Matthew DeMello recently hosted a conversation with Rad Desiraju, Corporate Vice President of Manufacturing at Microsoft, and Mike Geyer, Director of AI for Industrial at NVIDIA, on the ‘AI in Business’ podcast to explore how manufacturers can overcome these hurdles and harness the full potential of digital twins.

Their discussion covered topics such as data interoperability, infrastructure scaling, and the practical benefits of simulation dashboards. Both Rad and Mike emphasized the importance of integrated platforms and GPU-accelerated edge computing in driving operational efficiency, safety, and predictive capabilities in manufacturing.

This article examines two critical insights for manufacturing leaders from their conversation:

• Advancing manufacturing intelligence through 3D digital twins: Transitioning from traditional monitoring dashboards to generative AI-driven 3D digital twins enables real-time simulation analysis, improving decision-making for enhanced throughput, safety, and operational agility.
• Building scalable, interoperable infrastructure: Driving successful AI deployments by standardizing diverse data sources, adopting containerized edge computing, and leveraging GPU-accelerated platforms to reduce latency, improve data interoperability, and accelerate time-to-value.

Listen to the full episode below:

Guest: Rad Desiraju, Director of Worldwide Industry Advisory, Microsoft

Expertise: Manufacturing, Digital Transformation, Industry Advisory

Brief Recognition: Rad serves as Microsoft’s global industry advisor for manufacturing, helping enterprise clients adopt next-generation solutions, such as AI-powered digital twins and edge AI. He’s a featured presenter at events such as SEMICON West and SEMI’s “Digital Twin” workshops that focus on interoperability and standards. Rad is a recognized thought leader who actively advises manufacturers on operational modernization through collaboration with industry partners.

Guest: Mike Geyer, Head of Digital Twins, NVIDIA

Expertise: Digital Twins, Industrial AI, Robotics, Platform Strategy

Brief Recognition: Mike leads NVIDIA’s industrial AI initiatives, driving the development and adoption of digital twins and simulation technologies and libraries for manufacturing. With prior roles at Caterpillar and Autodesk, he brings decades of domain experience. His recent LinkedIn posts highlight NVIDIA’s collaboration with major manufacturers to transform with industrial and physical AI.

Advancing Manufacturing Intelligence Through 3D Digital Twins

To set the stage for the evolution of manufacturing dashboards over recent decades, Rad Desiraju outlines three clear phases. The way he describes this history serves as a spectrum for where many manufacturers stand across the global economy in terms of the sophistication of their dashboard deployments:

• Diagnostic dashboards that show what happened
• Operational dashboards that show what is happening
• Simulation dashboards that enable manufacturers to explore potential outcomes under various conditions.

Rad emphasizes that while many manufacturers have adopted monitoring dashboards to visualize factory data, most remain stuck at the first two stages, unable to simulate real-time “what-if” scenarios. 

That’s because meaningful simulation — the kind that helps leaders make proactive decisions — requires access to reliable, standardized data and infrastructure that can interpret it with high fidelity:

“One example of a development platform is what we’re working on with Omniverse, which allows you to use OpenUSD to bring together data from different sources and really combine this collaborative environment in 3D. 

We’re at a point where technology, acceleration, data platforms, and analytics are coming together, making the things we’ve been talking about and dreaming about for the last few decades a reality. It feels like that pace of change is accelerating really quickly.”

– Rad Desiraju, Director of WW Industry Advisory at Microsoft

Mike Geyer expands on the concept, noting that most product manufacturers have already been using 3D product design since the 1990s. The next step is applying the same level of dimensionality and analysis to the factory floor.

According to Mike, simulation dashboards make it possible to model entire production environments dynamically: adjusting product mix, rerouting supply lines, optimizing material staging, and more — all without physical trial and error:

“The back-end power is something that’s also advancing really quickly as compute moves to the cloud and becomes scalable. Manufacturing facilities are not 2D. When you walk around, you might hit your head on something that you wouldn’t see in a 2D dashboard.  

These are three-, four-, or five-story tall facilities with complex dependencies, not only how things move horizontally around the floor, but up and down. You need to be able to simulate the physical world as it exists if you’re going to train physical AI. That’s where the GPU acceleration and the open development platforms, and the intelligence that Microsoft can bring in — stitching all these things together — is what’s helping that status quo change so quickly.”

– Mike Geyer, Head of Digital Twins at NVIDIA

The ability to test and optimize factory performance virtually, as Mike describes, is transforming manufacturing decision-making. Instead of producing based on forecasted demand, companies can now pursue just-in-time production strategies driven by real-time simulation.

Mike further highlights how AI and automation — particularly the rise of humanoid robots and AMRs (Autonomous Mobile Robots) — are enabling these changes by helping manufacturers model not only product flows but also human workflows. Safety is a key outcome, especially in high-risk environments where robots can assist with dangerous or repetitive tasks.

Building Scalable, Interoperable Infrastructure

A core insight from both speakers is that implementing AI-powered simulation and digital twin systems is not simply a software upgrade — it’s an infrastructure challenge that spans the cloud, the edge, and the physical factory environment.

Rad outlines three distinct computing environments required for successful deployment: one to train AI models, one to simulate digital twins, and one to run inference at the edge — often on physical robots or PLC systems. Each environment has unique compute and latency requirements and must be orchestrated seamlessly to produce real-time insights:

“How we try to address this problem is we try to look into three vertical archetypes. The first one is building a modern data architecture, or what gets called a “data lake,” – having that is absolutely critical.

It brings together all the structured data together and sets the foundation for you to have a conversation with that data and ask questions about your data in natural language. That’s the first pillar that we built; it’s called structured data extraction. The second one is called Document intelligence. The third is making them interoperable.”

– Rad Desiraju, Director of Worldwide Industry Advisory at Microsoft

Edge computing, in particular, plays a vital role. Both speakers emphasize that for digital twins to be effective, manufacturers must reduce latency and process data as close to the source as possible. That means adopting containerized infrastructure and GPU acceleration — both in the cloud and on-premises — to manage compute-intensive workloads, such as 3D simulation and sensor data fusion.

Mike notes that a key differentiator of the NVIDIA-Microsoft partnership is the ability to right-size GPU performance to each specific workload. This flexibility helps manufacturers avoid overprovisioning, reduce costs, and shorten time-to-value — especially in industries where just a few minutes of downtime can cost millions.

“One of the things we’ve been really working on a lot with our partners, like Microsoft, is how we can make that compute scalable through these open development platforms that allow our development ecosystem to build digital twins with their own tools that are augmented by some of our parallel compute and accelerated GPU capabilities.”

– Mike Geyer, Head of Digital Twins at NVIDIA

Mike goes on to describe how status quo practices across manufacturing, such as updating warehouses every few years, are only possible with the right infrastructure.

In turn, Rad reinforces that any conversations about these levels of digital transformation and the infrastructure necessary to make them happen must begin with a clear business outcome in mind. Having clear, organization-critical objectives as the focus of any AI adoption is crucial for a wide range of manufacturing use cases, including improving safety, boosting throughput, and reducing operating costs.

From there, teams can align data standards, compute needs, and platform architecture accordingly. Rad emphasizes, “The first rule is: yes, digital twins are beautiful — but the important thing is to ask the question: What is the business value you’re trying to solve for?”