SK Telecom is moving to reposition itself as an AI infrastructure platform as it tries to recover its standing inside SK Group. The South Korean carrier's shift comes as domestic telecom growth slows, new ventures have underperformed, and security incidents have weakened its momentum.
According to The Bell, the company is building an AI strategy around data centers, cloud services, a personal assistant, and its own foundation models, with the goal of standing alongside SK Hynix as a key engine of SK Group's next growth cycle. SK Group leadership has also outlined a long-term AI data center investment plan worth about 1,000 trillion won (US$680 billion), with AI data centers positioned as the core of a wider industrial push.
SK Telecom is targeting AI data center construction of up to 15GW and wants to become an Asian AI infrastructure hub. The company said a 1GW AI data center could cost about 70 trillion won, reflecting higher costs for high-performance computing infrastructure and memory. To fund the buildout, the carrier plans to rely on liquid assets, strategic partner investment, long-term customer contracts, and project financing.
The Nvidia partnership has become a major part of that effort. SK Telecom and Nvidia plan to build a GW-scale AI Cloud using Nvidia's DSX platform, and the first AI Factory is scheduled to launch in 2027. The company is also working with Amazon Web Services on the Ulsan AI data center, extending its presence in cloud and infrastructure services.
On the software side, SK Telecom is pushing its A.dot assistant and A.X K1 foundation model into consumer, enterprise, and public-sector markets. Company data show A.X K1 is South Korea's first 519B-class ultra-large AI model, and it has entered the second phase of the government's sovereign AI foundation model program. The company has also reported results in public agencies, including the Ministry of Science and ICT and the Ministry of National Defense, while its investments in Anthropic, Perplexity, Scatter Lab, and Lambda are intended to strengthen links across AI models, search, conversational AI, GPU cloud, and infrastructure.
