SK Telecom and NVIDIA Plan Gigawatt-Scale AI Cloud in South Korea

SK Telecom and NVIDIA outline South Korean AI infrastructure plan

SK Telecom and NVIDIA said they will build a gigawatt-scale NVIDIA DSX AI Cloud in South Korea, with the first AI factory planned to come online in 2027.

The announcement was published by SK Telecom Newsroom and NVIDIA Newsroom, and NVIDIA also distributed the release through GlobeNewswire. According to SK Telecom, the planned AI Cloud is intended to support training, inference, and agentic AI workloads. NVIDIA said the project will use the NVIDIA DSX platform as the architecture blueprint for AI factory deployment in Korea.

First AI factory targeted for 2027

SK Telecom said the first AI factory under the plan is expected in 2027. NVIDIA’s announcement describes the broader project as a gigawatt-scale AI Cloud in Korea, using NVIDIA DSX to provide the design foundation for facilities built to run large-scale AI workloads.

The companies did not provide detailed public information in the cited announcements about the exact site, capital expenditure, procurement schedule, or energy sourcing for the first facility. They also did not disclose customer commitments or pricing for future services. The statements available from SK Telecom and NVIDIA focus on the technical direction and intended national role of the infrastructure rather than commercial operating details.

DSX positioned as an AI factory blueprint

NVIDIA’s release says NVIDIA DSX will serve as the AI factory architecture blueprint. In NVIDIA’s terminology, AI factories are data center facilities designed to produce AI outputs by running compute-intensive model training and inference workloads.

SK Telecom’s version of the announcement says the AI Cloud will be designed for training, inference, and agentic AI workloads. Agentic AI generally refers to systems that can plan and execute multi-step tasks using models, tools, and external data sources, although the companies’ public statements do not specify which agentic applications will run on the planned infrastructure.

The GlobeNewswire version of NVIDIA’s release also states that NVIDIA and SK Group plan joint research on next-generation AI factory architectures. The announcement does not detail the research timetable, staffing, or specific technical milestones.

Strategic context for South Korea

SK Telecom framed the project as part of Korea’s broader AI infrastructure development. NVIDIA’s announcement similarly described the effort as supporting Korea’s AI innovation. Those claims come from the companies’ own releases, so they should be read as the participants’ stated strategic rationale rather than an independent assessment of national impact.

The plan reflects a wider industry shift in which telecommunications providers, cloud operators, semiconductor companies, and governments are investing in facilities purpose-built for AI workloads. In this case, the publicly cited sources identify SK Telecom as the Korean telecommunications partner and NVIDIA as the provider of the DSX platform and AI factory architecture.

What remains unclear

The announcements establish the companies’ intent and the 2027 target for the first AI factory, but several practical questions remain unanswered in the public materials. SK Telecom and NVIDIA did not specify how much of the gigawatt-scale capacity will be available at launch, how quickly capacity may expand, or whether the infrastructure will serve only domestic customers.

They also did not identify the mix of training and inference services planned for customers, nor did they provide sustainability metrics such as projected power usage, cooling design, or renewable energy commitments. Those details will matter because gigawatt-scale AI infrastructure can carry significant implications for electricity demand, facility siting, and operating costs.

For now, the sourced facts are narrower: SK Telecom and NVIDIA say they plan a gigawatt-scale NVIDIA DSX AI Cloud in South Korea; the first AI factory is planned for 2027; the infrastructure is intended for training, inference, and agentic AI workloads; and NVIDIA and SK Group plan joint research into next-generation AI factory architectures.