· Industry· 日本語で読む
European manufacturers expand use of AI and digital twins, report says
AI and simulation are reportedly seeing wider use in factory operations and product development across European manufacturing. Examples included Siemens’ factory operations agent and the generation of synthetic images for inspection. The reported reduction rates are based on company estimates or reports, and their applicability requires verification.
What happened
roboticsandautomationnews.com reported on November 22, 2026, on the use of AI, simulation and digital twins by European manufacturers including Siemens, Mercedes-Benz, Audi and Volkswagen. According to the article, NVIDIA presented industrial use cases at GTC Berlin and early access to Factory Operations Blueprint (FOX), a software platform for factory operations. Siemens is integrating an agent developed using FOX into Intelligence Center X, the article said.
Why it matters
Siemens’ estimate of potential cost savings of 20% and the 40% reduction in inspection calibration time at its Erlangen plant relate to different challenges: factory operations support and inspection preparation. Manufacturers considering adoption need to verify not only the metrics but also the processes covered and the conditions used to calculate them before comparing results.
Analysis
The reported examples stand out because AI is being applied to more than one process. Siemens’ use of FOX supports factory operations and root-cause analysis, and the company estimates potential cost savings of 20%. Meanwhile, synthetic defect images at the Erlangen plant were reportedly used to calibrate board inspection. For Mercedes-Benz, the article reported aerodynamic simulations using Siemens Simcenter STAR-CCM+ running on NVIDIA GPUs, as well as visual inspection of windshield assembly. The article covers different applications—factory operations, inspection and vehicle design—but does not show that they are connected through a shared data platform or operational system.
The examples indicate that AI for manufacturing is being introduced in distinct areas: monitoring conditions on the factory floor and analyzing anomalies, generating and evaluating inspection data, and simulation during the design stage. However, FOX was presented as a software platform for developers, and Siemens’ integration was described as support for factory management and root-cause analysis. It should not be construed as a system that autonomously controls equipment or as evidence of an implementation that runs an entire factory as a single operation. The 20% figure is Siemens’ estimate of potential savings, while the 40% figure is a reported reduction in calibration time. Because they measure different things and cover different targets, it is not appropriate to compare them as if they were the same measure of effectiveness.
In a related article, we covered Kakao Mobility reportedly signs an MOU with NVIDIA to develop Level 4 autonomous driving technology, a plan involving NVIDIA technology in another field. That report concerned autonomous driving technology, a different subject from the manufacturing-floor and design-process examples here. The only common thread is that NVIDIA’s computing and software platforms are being discussed for applications across different industries; this does not demonstrate that the technology configurations or deployment results are the same.
Key questions going forward include the scope in which cost savings have been verified after FOX adoption, the conditions behind the 20% estimate, and how well inspection using synthetic images detects real defects. The article’s descriptions alone do not allow the scope or evaluation metrics of Mercedes-Benz’s simulations and visual inspections to be compared. Whether the individual examples, spanning product development and factory operations, lead to sustained data use or coordination between processes requires separate verification.
Sources
Translated from the Japanese original: /news/oushuusei-zou-gyou-ai-dejitarutsuin-no-katsuyou-kakudai-to-h-b52654. The Japanese version is authoritative if they differ.