The projects are scheduled to kick off in October 2026. Samsung will act as the sole global vendor for KT’s project and the main vendor for SK Telecom’s initiative. Both efforts will utilize 5G Standalone (5G SA) private networks deployed at specific industrial sites. The core objective is to demonstrate how AI-native connectivity can enhance safety, efficiency, and operational autonomy in high-stakes industrial environments, moving beyond traditional network management to active, real-time decision support for robotic systems.
Technological Architecture: Network in a Server
Central to these deployments is Samsung’s Network in a Server (NIS) platform. Unlike conventional RAN architectures that treat the network as a passive pipe for data, the NIS platform integrates virtualized RAN (vRAN), an AI Core, and AI applications onto a single edge platform. This consolidation allows for the processing of network intelligence at the edge, close to where the data is generated, which is critical for applications requiring ultra-low latency and high reliability.
For the KT project at HD Hyundai Samho, the NIS platform is supported by Samsung’s CognitiV Network Operations Suite (NOS). This suite facilitates the management of the AI-driven network functions. The infrastructure is designed to provide the reliable, low-latency connectivity necessary for real-time robotic control. Furthermore, the solution can incorporate accelerated GPU compute resources as needed, enabling the network to handle the intensive processing demands of physical AI tasks such as welding and painting operations. At the shipyard, KT will test several physical AI applications, including these industrial robots and autonomous units designed for telecom facility operations.
Petrochemical Safety and Autonomous Monitoring
Parallel to the shipyard trials, SK Telecom will conduct a physical AI trial at the SK Incheon Petrochem facility. This project focuses on enhancing safety in hazardous environments through autonomous monitoring. A key component is an autonomous patrol robot that transmits high-definition video in real time while moving through the facility. AI algorithms then analyze this video feed to identify potential hazards, facilitating integrated monitoring and rapid response capabilities.
Samsung’s NIS solution will also support this deployment, ensuring that the video data is transmitted with the fidelity and speed required for accurate AI analysis. The goal here is to explore how AI-native connectivity can transform safety protocols in petrochemical environments, where human presence is often limited due to risk, and where rapid detection of anomalies is paramount.
“The collaboration aims to accelerate the adoption of cutting-edge technologies, including AI RAN, physical AI, and autonomous operations, while laying the groundwork for the transition to 6G,” said Samsung.
Pathway to 6G and Commercial Implications
While these projects are built on 5G Standalone technology, they are viewed by industry participants as a critical precursor to 6G standards. By embedding AI directly into the network architecture, these trials address the limitations of current networks in supporting complex, real-time autonomous systems. The commercial implications for telecom operators are significant; by becoming the primary vendors for these high-profile government-backed initiatives, Samsung positions itself at the forefront of the next generation of network infrastructure sales.
For developers and industrial users, the success of these trials will determine the viability of AI-RAN as a standard for industrial automation. If the NIS platform can consistently deliver the promised ultra-low latency and reliable connectivity in diverse environments—ranging from the noisy, metal-rich environment of a shipyard to the hazardous chemical plants of a petrochemical facility—it could unlock new possibilities for autonomous operations across various sectors.
As the projects launch in October 2026, the focus will shift from theoretical capabilities to observed performance. The integration of GPU compute for AI tasks and the real-time analysis of video feeds will be closely monitored to assess whether AI-RAN can truly deliver on its promise of transforming industrial networks into intelligent, responsive systems. The outcomes of these trials will provide essential data for the continued evolution toward 6G, where AI is expected to be a native component of network design rather than an add-on feature.
