As Nvidia CEO Jensen Huang returned to Korea for the first time in seven months and began a packed three-night, four-day schedule, attention is focused on the Korean companies and AI-based core technologies he has mentioned. In particular, SK Telecom (017670) saw its shares surge after being named a key cooperation partner based on "digital twin technology" at "GTC Taipei," held in Taiwan just before Huang's visit to Korea.
Digital twin technology transforms the structures, environments, and operational processes of actual sites into digital assets and replicates them identically in a virtual space, thereby innovating operating systems. It prevents costly trial-and-error that could occur when changing process layouts or logistics routes without virtual testing. It can also significantly improve cost efficiency in business by reducing unpredictable contingencies and downtime.
In response, not only SKT (030200) but also LG Uplus (032640) and KT have entered the development of digital twin technology. Some forecast that, beyond telecom network management, businesses utilizing digital twins in various fields such as manufacturing, mobility, and logistics will become a new source of revenue for mobile carriers.
SKT, Joining Hands with Nvidia, Runs Simulations of 'hynix Semiconductor' Manufacturing Processes
What Is a Digital Twin
A technology that replicates real-world spaces, such as factories, identically in a virtual environment to verify in advance the effects of various scenarios. It can simulate process changes and equipment placement ahead of time to reduce trial-and-error and support data-based decision-making, drawing attention as a core physical AI technology.
In a keynote speech at the GTC Taipei conference held on the 1st of this month, Nvidia introduced SKT as a key cooperation partner in the field of manufacturing physical AI and unveiled SKT's digital twin technology in a video.
This was because SKT applied a digital twin to an SK hynix chip fab using Nvidia Omniverse and optimized it for a complex, large-scale manufacturing environment. Omniverse is Nvidia's collaboration platform for digital twins and 3D simulation.
In addition, SKT developed "agentic digital twin modeling" technology that uses Nvidia's "Agent Toolkit" to automate and intelligently process various data, such as equipment and spatial structures at manufacturing sites, in a way suited to digital twin environments. Through this, SKT improved the efficiency of data conversion, scene optimization, and performance improvement tasks that occur during the construction and operation of digital twins at manufacturing sites.
The company is also pursuing platform enhancement to integrate the Nvidia Omniverse library and improve loading speeds, execution performance, and GPU and memory usage efficiency for large-scale 3D scenes. This is expected to enable the realization of a stable and scalable digital twin environment even in complex, large-data manufacturing environments such as chip fabs.
An SKT official explained to The Seoul Economic Daily, "We have currently formed a regular meeting body between Nvidia and SKT for technology development and business cooperation," adding, "We are conducting not only executive-level meetings but also working-level meetings to solve problems and jointly develop in the process of developing the Omniverse and SKT digital twin platforms."
SKT has already completed a proof of concept (PoC) applying a digital twin to SK hynix's chip fab last year. Based on this, SK hynix plans to continue phased commercialization as part of its "Autonomous Factory 2030" construction goal.
LG Uplus Expands Autonomous Operation with Digital Twin Replicating Telecom Facilities, Also Validates DCIM
LG Uplus has been using technology combining digital twins and AI from this year in the process of managing telecom facility buildings where various communication equipment is placed. It replicated the actual telecom facility environment in a virtual space so that equipment placement and operating status can be checked and inspected in advance on screen.
Based on this, an AI agent constantly analyzes environmental changes such as power, temperature, and humidity within the facility to detect anomalies and automatically performs necessary measures. In addition, the company deployed the AI autonomous driving robot "U-BOT" inside the facility to roam around directly, collect equipment status and environmental information, and reflect it in the digital twin model.
In collaboration with LG Electronics, the company is also pursuing validation of data center cooling efficiency technology (DCIM) combining AI and digital twin technology starting in January. LG Electronics develops digital twin-based cooling and HVAC equipment control algorithms and takes on the role of deriving optimal operating conditions, using the cooling equipment at the actual Pyeongchon 2 Center as the validation target.
In this process, LG Uplus provides the actual data center operating environment and data, and leads the overall PoC environment composition and validation. The goal is to reduce cooling energy consumption by up to 10% and to expand the application of validation results to major data centers such as the Paju AIDC, scheduled for completion next year.
An LG Uplus official explained, "DCIM is currently being advanced beyond a simple monitoring system toward an operating platform that standardizes power, cooling, asset, and operational data and can expand into digital twins, predictive maintenance, and AI Work agents."
KT: "Building Simulation Environment for Production and Logistics Site Operation Management"
KT is also developing digital twin technology as a foundational technology for realizing physical AI. A KT official explained to The Seoul Economic Daily, "We are building a simulation environment that reflects the spatial structures and operating status of actual manufacturing, logistics, and service sites in a virtual environment, where robots and AI agents can learn and verify various situations in advance."
In particular, the company is reviewing the application of technology that uses digital twin-based simulation to analyze the operating status of production and logistics sites in real time and to verify in advance the movement routes, task dispatching, and bottleneck sections performed by multiple heterogeneous robots. The goal is to derive optimal operating plans before actual field application and to improve operational efficiency and stability.
KT plans to continue expanding its capabilities to realize an integrated physical AI environment combining robots, equipment, and operating systems by linking digital twins with the K-RaaS (KT Robot as a Service) platform.
