
SK Telecom reported an 89% year-on-year jump in revenue from its AI data centre business during the first quarter of 2026, highlighting the company’s aggressive push into artificial intelligence infrastructure. The strong growth reflects rising demand for AI computing services and GPU-based cloud solutions as enterprises expand AI adoption globally.
The company’s AI data centre segment generated approximately 131.4 billion won (around US$89 million) in quarterly revenue. SK Telecom attributed the sharp increase to higher utilization rates across its AI facilities and growing demand for GPU-as-a-service offerings, which allow businesses to access high-performance AI computing resources without building their own infrastructure.
The growth aligns with SK Telecom’s broader “AI Native” strategy, under which the company plans to develop hyperscale AI infrastructure across South Korea. The telecom giant has outlined ambitions to build up to 1 gigawatt of AI data centre capacity, positioning itself as a major infrastructure provider in the emerging AI economy.
SK Telecom is also expanding its AI ecosystem through partnerships and global collaborations. The company is reportedly working with OpenAI on a planned AI data centre project in South Korea and is participating in the Global Telco AI Alliance alongside major telecom firms including Deutsche Telekom and SoftBank. The alliance aims to develop telecom-specific AI models and shared AI standards using pooled industry data.
Beyond traditional telecom operations, SK Telecom is increasingly positioning itself as an AI infrastructure and cloud computing company. The firm has been investing in technologies such as GPU clusters, AI cloud virtualization platforms, and sovereign AI models tailored for enterprise and government use cases. Industry analysts view the company’s latest results as evidence that telecom providers are evolving into major players in the AI infrastructure race, where computing power and data centre capacity are becoming critical competitive advantages.




