South Korean telecom giant SK Telecom (SKT) plans to partner with Nvidia to build a gigawatt-scale artificial intelligence (AI) cloud in its home country, creating dedicated infrastructure to support the country’s growing demand for AI.
In this effort, SKT will deploy Nvidia’s DSX platform to build a large-scale AI factory dedicated to manufacturing the computing tokens needed for advanced AI workloads. The first of these facilities is expected to be operational in 2027.
Unlike general-purpose cloud environments, SKT’s AI cloud is customized for graphics processing unit (GPU) accelerated computing. It is designed to power training, inference, and agent AI workloads with a focus on providing sovereign AI services, physical AI services, and enterprise AI services to industries across Korea. SKT plans to eventually expand these services to the broader Asian market.
Nvidia’s DSX platform is a full-stack reference architecture consisting of software, hardware, and operations that generate AI tokens. This includes DSX MaxLPS software, which delivers the highest token performance per megawatt, and the DSX OS operating layer, which manages the AI Factory’s lifecycle, health automation, and multi-tenant capabilities.
“Telecommunications networks are becoming national AI infrastructure,” said Nividia CEO Jensen Huang. “They connect people, businesses, devices, and machines, and now they can become the backbone of the new AI cloud. With Nvidia DSX, SK Telecom can build Korea’s AI cloud at scale, bringing agent, enterprise, and physical AI to the businesses and industries that power Korea and the world.”
This partnership goes beyond deploying AI infrastructure. SK Group and NVIDIA also announced plans to begin joint research on next-generation AI factory architecture. This “silicon-to-grid” innovation will focus on accelerated computing, memory technology, and data center operations to drive more resilient and scalable AI services.
SK Group Chairman Choi Tae-won said the partnership will secure the company’s full-stack AI infrastructure capabilities, from chip-level hardware to broader data center operations. “We intend to work with Nvidia to address GPU, memory, and energy challenges and become the leading AI cloud company shaping Asia’s AI ecosystem,” he added.
SKT also officially becomes an Nvidia cloud partner, joining the tech giant’s global ecosystem of AI infrastructure and software providers.
Build a full-stack AI powerhouse
The deal with Nvidia is the latest in a series of moves by SKT to transform from a traditional connectivity provider to a full-stack global AI company.
Earlier this year, a consortium led by SKT built the AX K1 foundational model, which excels at Korean-based tasks as well as math and coding. The 519 billion parameter model is currently being used in the South Korean government’s sovereign AI project.
Beyond Nvidia, SKT is building its own custom AI infrastructure stack to improve power efficiency. In April 2026, the company partnered with semiconductor company Arm and AI accelerator startup Rebellions to jointly develop custom data center hardware and software for inference workloads.
Additionally, SKT is making significant investments in the broader AI ecosystem. In addition to establishing a dedicated AI investment unit to back emerging startups, the operator has poured significant capital into generative AI, most notably making a $100 million investment in Claude developer Anthropic in 2023.
The deepening relationship between the carrier and NVIDIA was made clear at last week’s GTC Taipei event. There, the company demonstrated how it is using the NVIDIA Omniverse library to apply a digital twin to the SK Hynix semiconductor factory, optimizing the complex manufacturing environment and laying the foundation for the industrial AI applications that the new gigawatt-scale cloud aims to support.
