Manufacturing
Objective
Siemens powers AI–driven innovation across every industry and industrial workflow and empowers its customers to design complex products and systems—from chips to AI factories—and adapt manufacturing and production in real time.
To bring AI into the real world and enable faster innovation, continuous optimization, and more resilient, sustainable manufacturing, Siemens is integrating NVIDIA AI infrastructure, simulation libraries, models, frameworks, and blueprints across its full lifecycle of products, including Siemens Teamcenter, Simcenter STAR-CCM+, and the Siemens Industrial Copilot.
Customer
BMW Group
HD Hyundai
Maserati
PepsiCo
Partner
Siemens
Use Case
Simulation / Modeling / Design
By integrating NVIDIA CUDA-X™ and Omniverse™ libraries and AI models, Siemens software delivers:
Increased Efficiencies:
- Photorealistic visualization of massive, complex, and feature-rich datasets common in engineering and manufacturing.
- Design, simulate, and operate products, plants, and factories in full context before anything is built.
- 10%–15% reductions in capital expenditure.
Better Decision-Making:
- Alignment across global teams, with tracking for feedback and changes within Teamcenter, accelerating design reviews and reducing errors and data discrepancies.
Increased Productivity:
- 25% reduction in reactive maintenance time for workers using the Industrial Copilot.
- More productive manufacturing plants with reduced downtimes.
- 20% increase in throughput and faster design cycles.
Accelerating Innovation and Savings from Product Design to Operations
Global manufacturing faces a critical skill and labor shortage, with projections of 85 million vacant positions by 2030. Manufacturers across automotive, aerospace, and marine also face mounting pressures to accelerate innovation, reduce costs, and ensure product quality and sustainability.
At the same time, the productivity of expert teams of industrial designers and engineers is impacted by fragmented workflows and technologies. Using separate applications for rendering, collaboration, and data management, for example, creates silos and data discrepancies, making it difficult to conduct thorough collaborative design and styling reviews. Working with visualization tools that struggle to handle the scale of massive, complex products and systems compound these challenges and result in bottlenecks in design reviews, misalignment between teams, and costly physical prototyping. The inability to visualize products and systems in physically accurate environments limits the ability of teams to identify issues early, impacting both functional and aesthetic outcomes. Overall, these workflow and technology limitations slow down time-to-market and introduce errors that can ripple through the entire product lifecycle.
To address these challenges, Siemens and NVIDIA are collaborating to bring the latest AI, simulation, and digital twin technologies to the world’s design, engineering, and simulation experts and their workflows.
Accelerated Product Development With Digital Twins and Physically Based Simulations
Siemens, NVIDIA, and BMW are collaborating to accelerate automotive aerodynamics simulations using advanced computational fluid dynamics (CFD) software and accelerated AI infrastructure.
Aerodynamic simulations are essential for optimizing vehicle shape, balancing design and efficiency, and ensuring performance targets are met. BMW runs many aerodynamic simulations daily, each simulating up to 100 million cells (tiny 3D blocks representing air and car surfaces). While CFD software has traditionally run on CPUs, GPU-accelerated simulations offer faster performance at lower cost.
Siemens’ Simcenter STAR-CCM+ CFD software now runs natively on both CPUs and GPUs, ensuring consistent results regardless of hardware while enabling GPU acceleration. When running a 458-million-cell steady state aerodynamics case on the latest NVIDIA B200 GPU node, performance is equivalent to over 10,000 CPU cores.
Running CFD simulations of this scale on GPUs also reduces energy consumption and infrastructure costs. Further performance gains are possible with newer generation accelerated systems, including NVIDIA DGX™ B200 and H200, which feature higher memory bandwidth to deliver faster simulations.
Additionally, the integration of NVIDIA Omniverse libraries and generative AI capabilities into Simcenter STAR-CCM+ makes physics-based digital twins more immersive. These integrations, which are part of a broader integration of Omniverse libraries within the Siemens Xcelerator Platform, allow engineers to leverage AI within their CFD workflows and design, build, and test next-generation products, manufacturing processes, and factories virtually before physical construction begins.
For example, Maserati is tapping into Siemens solutions powered by Omniverse application programming interfaces to interactively visualize airflow over car bodies and improve its manufacturing process.
Siemens and Maserati
Real-Time Collaboration From Anywhere
To accelerate innovation and cycle times in product development, Siemens developed the Teamcenter Digital Reality Viewer. By integrating NVIDIA Omniverse libraries, users can take advantage of real-time ray tracing and GPU-accelerated rendering directly within Teamcenter, enabling seamless, photorealistic, bill-of-materials (BOM)-driven visualization of multi-CAD data managed in Teamcenter. This cloud-based rendering-as-a-service is embedded directly within the Teamcenter web client, providing instant, browser-based access to photorealistic digital twins for all users.
Teamcenter Digital Reality Viewer is hosted in the secure Siemens Xcelerator Cloud Service environment and powered by a centralized, cloud-based GPU cluster, meaning users can access its high-performance visualization capabilities on demand without the burden of managing the underlying infrastructure and software.
Teamcenter’s collaboration tools allow stakeholders to review, annotate, and iterate on designs together in real time. All feedback and changes are tracked in Teamcenter for full traceability. Users can also validate product appearance in multiple virtual environments, simulate real-world lighting, and conduct detailed visual analysis of assemblies, materials, and finishes.
Factory-Scale Design and Simulation
The Digital Twin Composer from Siemens enables teams to design, simulate, and operate products, plants, and factories in full context before anything is built. Digital Twin Composer brings together 2D and 3D digital twin data from across the Siemens Xcelerator portfolio and fuses it with real-time operational information in a managed, secure, photorealistic scene powered by NVIDIA Omniverse libraries, creating one living model that spans the entire lifecycle of a product, process, or facility.
PepsiCo is an early adopter, working with Siemens and NVIDIA to convert selected U.S. manufacturing and warehouse facilities into high‑fidelity 3D digital twins that allow them to simulate their end‑to‑end plant operations and supply chains.
Using Digital Twin Composer and computer vision, PepsiCo teams can recreate every machine, conveyor, pallet route, and operator path with physics‑level accuracy, and take advantage of AI agents that can simulate and refine system changes and identify up to 90% of potential issues before any physical modifications are made. This has already delivered an estimated 20% increase in throughput on initial deployments, driven faster design cycles with nearly 100% design validation, and produced 10%–15% reductions in capex by uncovering hidden capacity and allowing teams to validate investments virtually before committing to physical upgrades.
“NVIDIA Omniverse integration enables Siemens to increase product efficiencies and productivity, allowing our customers to save time on product development and design. By approximating real-world behavior in a digital twin, our customers can draw conclusions and optimize product design faster, allowing faster time to market.”
Enzo Krka
Senior Product Manager, Siemens
Industrial Copilot Assistants to Empower Shop Floor Workers
The Industrial Copilot is an AI assistant that runs on the factory floor to help workers be more efficient, productive, and capable. The copilot integrates with both static information repositories (manuals, maintenance documents) and dynamic data sources (machine controllers, ERP systems), creating a comprehensive knowledge ecosystem. Workers can leverage the copilot to help optimize production, troubleshoot machine faults, and flag manufacturing issues through video feeds.
Workers interact with the copilot through natural language or speech for ease of use on the shop floor. The copilot also supports multimodal inputs including text, images, and time-series data. A maintenance engineer confronted with a broken machine can take a photo of the problematic component, and the copilot can analyze the image, identify issues, and offer step-by-step instructions to resolve the issue.
The Siemens Industrial Copilot is deployed on premises to ensure that sensitive production data never leaves the factory site. Cloud models are replaced with optimized edge versions, small language models (SLMs) are employed for specialized tasks, and NVIDIA NIM™ microservices enable optimized inference. By balancing hardware constraints with performance needs, copilots operate on industrial PCs at the edge running NVIDIA hardware and software.
The copilot was built with a fusion of components within the AI Blueprint for video search and summarization (VSS) including VLMs, LLMs, and NeMo™ microservices. In early pilots, Siemens has reported a 30% increase in productivity, with the potential to reach up to 50%.
Learn more from Matthias Loskyll, head of virtual control and industrial AI at Siemens Factory Automation, who joined the NVIDIA AI Podcast to discuss how Siemens’ work with NVIDIA is reshaping manufacturing, as the industry hits a turning point.
A New Standard for Efficiency and Innovation
Siemens is redefining engineering and manufacturing by integrating AI, digital twin technology, and accelerated AI infrastructure. With the integration of NVIDIA libraries, models, and frameworks across platforms including Siemens Teamcenter, Simcenter STAR-CCM+, and Siemens Industrial Copilot, Siemens empowers manufacturers to accelerate product development, optimize operations, and enhance workforce skills, setting new industry standards for efficiency and innovation.
Start developing with NVIDIA Omniverse libraries and OpenUSD.