SK Telecom CEO shifts strategy to efficiency, AI

SK Telecom CEO Jung Jai-hun announced a shift in company strategy from growth based on scale to a focus on efficiency and AI at his first town hall meeting on December 16.

Jung, who became CEO in October, said the company will now use return on invested capital instead of EBITDA as its main performance metric to emphasize profitability.

He urged staff to move away from established practices, and said management would take responsibility for any failures during this transition.

Jung addressed recent hacking incidents and service disruptions, saying that improving service quality, security, and operational safety would be a priority.

He identified AI as the company’s main growth area, and shared plans to expand AI data center operations, develop new AI solutions, and create proprietary AI models for global competition.

🔗 Source: The Korea Herald

🧠 Food for thought

Implications, context, and why it matters.

SK Telecom’s ROIC pivot lacks targets amid AI buildout

SK Telecom shifted its metric from return on invested capital (ROIC) to EBITDA. It gave no baseline ROIC, targets, or whether ROIC exceeds weighted average cost of capital (WACC). That leaves investors guessing as Q3 2025 revenue fell 12.2% with a net loss of South Korean won (KRW) 206.6 billion 1.
Ulsan will scale from 103 MW to 1 GW 1. The firm committed KRW 5 trillion to AI over five years 2 and is advancing an Ulsan artificial intelligence data center (AIDC) pegged at KRW 7 trillion 1. Investors need disclosed hurdle rates and payback to judge returns.

Vendors and GPU lessors see openings in SK Telecom’s data center build

Power and cooling suppliers can expect RFPs as Ulsan expands to 1 GW 1. SK Telecom signed a procurement pact with Schneider Electric for switchgear and uninterruptible power supply (UPS) systems 3. Energy storage with SK On lithium-ion batteries fits Ulsan and a Vietnam liquefied natural gas (LNG) site with SK Innovation 3 1.
GPU leasing firms (companies that rent chips for AI workloads) and data center construction integrators (firms that design then build facilities end-to-end) can target the Manufacturing AI Cloud plan to buy over 2,000 NVIDIA RTX PRO 6000 Blackwell GPUs 4. It was named top bidder for a government lease to deploy 1,000 Nvidia B200 GPUs 5. SK Telecom is adding out-rack and clustering packages 4. It is also releasing in-rack options to speed AIDC builds 4.