Production Technician

European manufacturers expand NVIDIA use in engineering and factories. NVIDIA's Factory Operations Blueprint (FOX) is available in early access on GitHub, giving developers tools to build always-on factory management agents. These agents can coordinate machines, inspect products, investigate problems and manage fleets by connecting models such as Cosmos and Nemotron with engineering software and operational systems. Cosmos in-context learning lets agents adapt from examples and documentation without fine-tuning. Siemens is integrating FOX-based agents into Intelligence Center X to identify root causes and coordinate operations, with estimated cost savings of 20%. Its Process Preparation X agents also use Cosmos Data Factory tools to generate defect images for circuit-board inspection. Siemens used this approach at its Erlangen electronics factory to reduce inspection calibration time by 40%. Engineering applications include NVIDIA's general release of cuDSS at GTC Berlin to accelerate demanding simulation calculations. Mercedes-Benz uses GPU-powered aerodynamic simulations and AI agents for vehicle inspection and windshield assembly verification. Semiconductor simulation benchmarks using Synopsys software show up to 40x speedup on imec's model, while Infineon reports a 12x speedup in 3D breakdown simulation. Manufacturers are also testing production systems virtually. Audi uses an Omniverse-based battery-production digital twin and is working with Dassault Systèmes and Agile Robots to validate mobile humanoid robots and logistics processes. Schaeffler is integrating actuator digital twins with Isaac simulation frameworks. Schneider Electric and AVEVA are developing a digital twin for Nscale to compare power, cooling and equipment configurations before physical changes. Originally reported by blogs.nvidia.com Read the source

Nvidia and Nokia turn telecom networks into AI computing hubs. * by Nathalia ITS - October 22, 2026 Reading time: 4 min On October 21, at GTC Berlin, Nvidia and Nokia advanced AI native networking solutions across telecom infrastructure in Europe and the US, allowing operators to combine 5G connectivity with accelerated computing to deliver AI services closer to businesses and connected devices. Telecom companies are now looking for ways to benefit from growing demand for AI computing. Operators already own data centers, fiber connections, and thousands of network sites, giving them infrastructure that technology companies would otherwise need to build. Rather than using these assets only to carry data, operators are beginning to process it, opening a new market for services that need immediate responses. AI-RAN turns telecom infrastructure into computing power. According to NVIDIA, T-Mobile, Deutsche Telekom, Orange, Elisa, and BT are among the operators developing distributed intelligence platforms that combine traditional mobile services with AI processing. At the center is AI Radio Access Network (AI-RAN) technology, which allows computing equipment to handle mobile traffic and AI tasks on the same platform. For operators, the appeal of an AI-native network is that existing infrastructure can support new services without requiring a separate computing system for every application. "Telecom networks were built to carry the world's conversations. Now, they will also carry the world's intelligence," said Jensen Huang, founder and CEO of NVIDIA. "With Nokia, we're extending accelerated computing into the radio network itself, creating the foundation for AI-native 5G." The technology is already moving beyond laboratory demonstrations. T-Mobile completed what the companies describe as the world's first AI-RAN call on a commercial network in Bellevue, Washington, using Nokia's CUDA-accelerated RAN software and NVIDIA-powered Aerial RAN Computer (ARC) servers with massive multiple-input multiple-output (MIMO) radios, connected to an existing 5G standalone core and a commercial smartphone. Elisa completed what Nvidia calls Europe's first AI-RAN field trial, running voice, mobile data, and video analytics simultaneously over 5G. The trial demonstrated how an AI native networking platform could share computing resources between network operations and other services without requiring separate hardware for each workload, and Elisa expects commercial AI-RAN services in 2027. Understanding this architecture is important because an AI Ran telecom network architecture diagram shows more than the connection between a smartphone and a cell tower. Computing resources can be placed at central data centers, regional facilities, or radio sites, depending on how quickly an application needs information. This is particularly relevant for industrial automation. Deutsche Telekom is working with Agile Robots on systems where robots handle immediate actions locally while more demanding tasks are processed at the network edge. Such arrangements give native connectivity a larger role in coordinating machines without depending entirely on distant cloud servers. The business opportunity behind AI native networking solutions therefore extends beyond faster mobile connections. Operators could use their networks to provide computing capacity directly to companies, although turning those technical capabilities into profitable services remains a challenge. AI native networking benefits for businesses. For businesses, the value of AI native applications depends on whether they can respond quickly, because timing can affect the outcome. Orange is exploring this difference through multilingual voice intelligence services for enterprises and emergency responders. Using NVIDIA's Nemotron Speech models on its own ARC platform, the operator aims to translate conversations closer to users, showing how an AI native networking platform could reduce the need to send every request to distant servers. Meanwhile, BT is preparing for live AI-RAN trials in Birmingham in 2027, with Nokia and Nvidia, and is adopting Nemotron to build agents that inspect field-installation photos and flag faults, to cut repeat site visits, while Deutsche Telekom and Elisa are working toward broader deployments. These trials will help determine whether native connectivity can deliver consistent performance when operators run conventional mobile services and AI workloads together. There is also a question of how these services will be sold. Companies searching for the best network-as-a-service framework for AI networking will need more than access to computing power. For telecom operators, distributed intelligence platforms could provide that combination by keeping computing resources within their own infrastructure. However, this would also require investments capable of managing different workloads. Nokia's financial results show why the industry is paying attention. According to The Wall Street Journal, the company's network infrastructure sales rose 17% year over year in the third quarter, while orders from AI and cloud customers reached €1.5 billion ($1.68 billion). "In AI & Cloud, net sales more than doubled year over year," said Justin Hotard, Nokia CEO. Yet stronger demand brings semiconductor shortages that already affect equipment's availability, and Nokia expects supply constraints to continue. The company is expanding production at its semiconductor facility in San Jose and integrating optical networking specialist Infinera, acquired for $2.3 billion. These investments are increasingly relevant as AI native network infrastructure requires both advanced processors and high-capacity connections between computing locations. Beyond hardware, Nokia and Cisco are integrating Nvidia Nemotron models into network management software. Each platform enables intelligent distribution of operational tasks, allowing AI agents to help identify faults, recommend fixes, and reduce manual intervention. Such capabilities could make AI-native network operations more efficient, but they also raise expectations for network reliability. Failure may eventually affect not only mobile connectivity, but also the AI services businesses rely on. That is where the broader significance of AI native networking solutions becomes clearer. Telecom operators are beginning to compete over where intelligence is processed, who controls the infrastructure, and which companies can deliver it reliably. The next test for AI native networking solutions is not whether AI can run inside a telecom network. Early demonstrations have already shown that it can. Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on its Telecom sections to stay informed and up-to-date with its daily articles.

Command palette. 14 hours ago European industrial leaders are expanding the use of NVIDIA CUDA-X, Omniverse, Cosmos and Nemotron across engineering, simulation, manufacturing and factory operations, signaling a broad shift toward AI-driven industrial workflows. The moves, highlighted around GTC Berlin, position NVIDIA's open models, acceleration libraries and digital-twin tools as core infrastructure for companies seeking higher productivity, faster innovation and more efficient production. A major element of the announcement is the early-access availability on GitHub of the NVIDIA Factory Operations Blueprint, known as FOX. The blueprint gives developers a path to build always-on factory manager agents that can coordinate work across machines and support specialist tasks such as visual inspection, root cause analysis, report generation and fleet management. These agents can connect to frontier models such as Cosmos and Nemotron, as well as to foundational industrial systems including computer-aided design, product lifecycle management, simulation software, and IT and operational technology platforms. By correlating visual findings with machine logs, they can monitor operations in real time, investigate issues and coordinate responses. Cosmos in-context learning allows agents to adapt using examples and documentation without fine-tuning. Siemens is a central partner in this ecosystem, integrating FOX-based agents into its Intelligence Center X agentic enterprise system to identify root causes and coordinate factory operations, with estimated cost savings of 20 percent. Siemens' Process Preparation X agents also use NVIDIA Cosmos Data Factory tools to generate defect images for circuit-board inspection. At the company's Erlangen electronics factory, this approach has reduced inspection-process calibration time by 40 percent. The broader announcement covers accelerated engineering from early design through production planning. NVIDIA CUDA-X libraries are speeding simulation workloads, giving engineers faster feedback on structural stress, fluid flow, electromagnetic behavior and semiconductor performance. Mercedes-Benz is using Siemens' Simcenter STAR-CCM+ on NVIDIA GPUs to compare vehicle designs through faster aerodynamic simulations, while also applying NVIDIA Cosmos 3 and the Metropolis VSS blueprint to build agents for vehicle inspection and windshield assembly verification. Cadence is benchmarking automotive fluid-dynamics workloads, and Autodesk is bringing NVIDIA PhysicsNeMo and Warp frameworks into its simulation tools to help designers predict performance and iterate more quickly. In semiconductors, imec and Infineon are using Synopsys software with NVIDIA accelerated computing to speed simulation workloads, reporting benchmark results of up to 40x for S-Litho and 12x for 3D breakdown simulation. In aerospace, Leonardo is working with Neural Concept to combine aircraft geometry and aerodynamic simulation results using Omniverse. Car manufacturers are also using Dassault Systèmes' LEO and AURA virtual companions, powered by Nemotron and integrated with CATIA and SIMULIA AI agents, to design and virtually test vehicle components before building physical prototypes. A major theme is virtual validation of factories, robots and AI infrastructure. NVIDIA Omniverse and Isaac libraries are enabling digital twins that let manufacturers test production systems and robot behavior before changing the factory floor. Audi is using an Omniverse-based battery-production digital twin to test concepts before equipment is available, and is working with Dassault Systèmes and Agile Robots to simulate mobile humanoid robots and logistics processes. Schaeffler is integrating digital twins of humanoid actuators with Isaac simulation frameworks, while Wandelbots is using NOVA with Isaac Sim to validate assembly workflows digitally. The pattern extends to industrial facilities and AI data centers. Falkor, a Kongsberg company, is combining agents, Omniverse and Microsoft Azure virtual machines with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs to support spatial intelligence for inspection and maintenance in process and energy industries. Schneider Electric, together with AVEVA, is developing an AI factory digital twin with Nscale to compare power, cooling and equipment configurations before deployment.