SK Telecom (NYSE: SKM) has announced advancements in its digital twin initiatives for semiconductor manufacturing, applying the technology to SK hynix fabrication facilities using NVIDIA’s Omniverse libraries. The development marks a step toward the creation of more intelligent and autonomous manufacturing environments in the semiconductor industry.
The company said it successfully completed a proof of concept (PoC) last year for deploying digital twin technology in an SK hynix semiconductor fab. Building on those results, SK Telecom plans to commercialize the solution in stages, supporting SK hynix’s broader goal of establishing an “Autonomous Fab” by 2030.
Digital twins are virtual replicas of physical facilities, equipment, and processes that allow manufacturers to simulate operations and evaluate potential changes before implementing them in real-world environments. The technology is increasingly viewed as a key component of physical AI, enabling organizations to improve efficiency, reduce operational risks, and make more informed decisions through data-driven simulations.
According to SK Telecom, semiconductor fabrication plants represent one of the most demanding environments for digital twin implementation due to their extensive infrastructure, intricate equipment configurations, and enormous volumes of operational data.
To address these challenges, the company has developed what it calls “Agentic Digital Twin Modeling” technology using the NVIDIA Agent Toolkit. The solution is designed to automate and intelligently process manufacturing data, including equipment specifications and spatial information, required to build and maintain digital twin environments.
SK Telecom said the technology improves processes such as data conversion, scene optimization, and system performance management, helping streamline the creation and operation of digital twins in large-scale industrial settings.
The company is also integrating NVIDIA Omniverse libraries into its platform to enhance the performance of OpenUSD-based three-dimensional digital environments. The integration aims to improve scene-loading speeds, execution efficiency, and GPU and memory utilization when handling large and complex manufacturing datasets.
By optimizing these capabilities, SK Telecom expects to provide a scalable and stable digital twin environment capable of supporting semiconductor fabs, where thousands of interconnected systems and equipment generate vast amounts of operational data.
Mike Geyer, Head of Industrial Digital Twins at NVIDIA (NASDAQ: NVDA), said semiconductor fabs are among the most challenging manufacturing environments because of their combination of large-scale 3D data, complex equipment structures, and demanding optimization requirements. He noted that SK Telecom has demonstrated strong technical capabilities in applying and validating NVIDIA Omniverse libraries and the NVIDIA Agent Toolkit in real-world industrial settings.
Cho Ik-hwan, Head of Physical AI at SK Telecom, said the collaboration has demonstrated that manufacturing digital twins can evolve beyond visualization tools into physical AI platforms capable of understanding and optimizing large-scale manufacturing data. He added that the company intends to expand its collaboration with NVIDIA and extend physical AI technologies into additional industrial sectors beyond semiconductors.
The initiative aligns with SK Telecom’s broader strategy of expanding its enterprise and public-sector AI business. The company said it will continue strengthening its portfolio of AI solutions, spanning infrastructure, models, and services, as demand for industrial AI applications grows across multiple sectors.
