- SK Telecom and NVIDIA forge a partnership to introduce the first AI factory online as part of the Gigaword-scale AI Cloud initiative.
- The AI Cloud is a GPU-based infrastructure built on NVIDIA DSX, used specifically for large AI workloads.
- The venture introduces sovereign and agentic AI across Korean industries, placing AI infrastructure in telecom networks.
As artificial intelligence systems transition to physical and sovereign agentic AI, the necessity of computing infrastructure has increased. Conventional cloud platforms for generic workloads do not suit the energy and performance demands of the latest AI systems.
To solve this issue, SK Telecom and NVIDIA have announced a gigawatt-scale AI cloud in South Korea to facilitate AI factories. This AI factory manufactures tokens using NVIDIA’s DSX platform. As the foundation, the project shifts towards GPU-based cloud infrastructure for suitable optimization, inference, and huge tasks. The first AI factory will be released online in 2027.
Why Does SK Telecom Integrate AI Cloud?
The AI Cloud launched by SK Telecom is not traditional, but a GPU-based AI Cloud consisting of AI factories. According to SK Telecom, built for AI workloads, these factories transform large datasets into efficient tokens. This optimizes the performance per megawatt.
The primary aspect is NVIDIA DSX, a full-stack AI factory infrastructure combining systems, software, and optimized computing. This project reduces token costs while ensuring efficient proprietary production. Unlike conventional providers, which provide flexibility, the AI Cloud focuses on AI-centric aspects, efficiency, and inference.
What is Behind the First Online AI Factory Launch?
The rollout of SK Telecom’s first AI factory is a cornerstone for South Korea’s ecosystem. It represents a comprehensive stance where artificial intelligence architecture moves closer to national control, especially for sovereign AI. The nation’s economy relies on semiconductors, telecommunications, robotics, and mobility niches where artificial intelligence rapidly transitions.
By integrating native AI factories, SK Telecom plans to achieve a high-end architecture that supports large workloads without depending on external platforms. This shift also places telecom companies at the core of artificial intelligence, using it as a medium for national-scale intelligence systems.
How Does NVIDIA DSX Allow GPU-Based AI Factories?
NVIDIA DSX Powered AI factories optimize the core design across hardware, software, and operations. Its infrastructure is built to use low token cost and high efficiency, key aspects for sustainable AI deployment.
The twin-core components that ensure optimal functioning are DSX Max LPUs, which maximize token performance per megawatt, and DSX OS, which adds an additional layer that supports lifecycle management and resilience. By streamlining the AI factory design, NVIDIA DSX allows SK Telecom to boost deployment and ensure consistent performance across agentic workloads and inference.
The purpose of the AI cloud is to serve organizations and industries nationwide, supporting autonomous, physical, and agentic AI services. These consist of accelerating manufacturing, robotics, enterprise automation, and large-scale artificial intelligence agents. SK Telecom’s prior expertise in data centers and proprietary architecture provides a base for quick acceleration.
The company put forward the future ideas through deployments in semiconductor manufacturing and participation in Korea’s sovereign AI foundation model project. Moreover, SK Telecom’s role as an NVIDIA cloud partner embeds it in a world architecture, allowing access to NVIDIA’s new technologies while delivering native services that cater to regional needs.
The launch of South Korea’s first AI factory by SK Telecom determines a crucial shift in how AI architecture is embedded and perceived. By creating a gigawatt-scale GPU-based AI cloud with NVIDIA DSX, the organization moves beyond conventional cloud providers towards local AI factories suitable for efficiency and autonomy. As artificial intelligence becomes a crucial part of economies, ventures like this depict the necessity of special architecture with energy-efficient design. With this AI cloud stance, SK Telecom places itself in a position to develop the nation’s infrastructure.
