Prabhu RamSK Telecom has hived off its AI infrastructure business from its AI services business with the launch of SK Hyper.
SK Hyper is a dedicated subsidiary responsible for the company’s AI data center business. It brings together the AI infrastructure value chain, from land acquisition and power infrastructure to customer acquisition and operations. The company is backed by KRW 750 billion in committed investment through 2030 and targets 15GW of AI data center capacity by 2035.
What’s Driving Growth
The primary constraint in AI deployment has shifted from model capability to compute, power, and data center capacity.
Land with reliable grid access, transmission infrastructure, and cooling capacity is becoming as strategic as spectrum once was for telecom operators. SK Telecom’s response is structural. Rather than expanding an internal business unit, it has created a standalone company with dedicated capital, leadership, and operational authority. This enables faster execution while allowing AI infrastructure investments to scale independently of the priorities of its core telecom business.
Why It Matters
SK Telecom’s move reflects a broader industry shift.
Operators including NTT, Singtel (through Nxera), and KT are investing in dedicated AI infrastructure businesses, building on existing strengths in networks, real estate, enterprise relationships, and power management. Telecom operators are increasingly positioning themselves as AI infrastructure providers rather than connectivity providers alone.
The opportunity is significant. Data center capacity across Asia-Pacific is expected to expand rapidly over the coming years. Yet the biggest constraint is not construction. It is power. Land can be acquired and facilities built on schedule, but securing sufficient electricity is becoming the defining challenge for AI infrastructure expansion.
The Big Picture
The AI race is entering its infrastructure phase.
Competitive advantage will increasingly depend not only on who develops the best AI models, but on who can secure power, deploy compute capacity, and operate AI infrastructure at scale. SK Telecom’s creation of SK Hyper reflects a broader industry reality: AI leadership will be defined as much by infrastructure as by innovation.
