Strategic Overview
SK Group stands as the second pillar of Korea's chaebol-driven technology strategy and occupies a position of extraordinary leverage within the K-Moonshot initiative. Through SK Hynix, the group controls the global chokepoint for AI memory semiconductors. Through SK Telecom, it is building Korea's AI-native telecommunications infrastructure and deploying one of the world's largest enterprise AI agent ecosystems. Through SK Innovation and SK On, it connects to the energy storage and data center power supply chains that underpin sovereign AI compute capacity.
The group's strategic coherence around AI infrastructure, from the silicon layer through the network layer to the application layer, gives SK a differentiated profile among K-Moonshot corporate partners. While Samsung Group offers breadth across consumer electronics, biologics, and robotics, SK Group's value proposition is vertical depth in the AI compute stack. For analysts tracking K-Moonshot's Mission 11: AI Accelerator Chips and the broader semiconductor sector, SK Group is the primary entity to monitor.
SK Hynix: Commanding the AI Memory Supply Chain
SK Hynix has established itself as the dominant supplier of High Bandwidth Memory (HBM), the specialized DRAM technology that has become the critical enabler for AI training and inference accelerators. With an estimated 70% market share in HBM4, SK Hynix's position is formidable. The company's entire 2026 HBM production has been sold out, a demand signal that reflects the insatiable appetite of hyperscale cloud providers and AI laboratories for memory bandwidth.
The M15X fabrication facility, scheduled to come online in May 2026, represents SK Hynix's next major capacity expansion. Located in Cheongju, the facility will significantly increase the company's ability to produce advanced HBM and DRAM products. The investment in M15X is part of a broader capital expenditure programme that positions SK Hynix to maintain its market leadership through the end of the decade, even as Samsung Electronics and Micron Technology intensify their competing HBM programmes.
SK Hynix's technological roadmap includes 16-layer HBM4 at 48GB capacity, utilizing the H3 (HBM3E successor) architecture. The progression from 8-layer to 12-layer to 16-layer stacking represents not merely an incremental improvement but a compounding advance in memory bandwidth per package, which directly determines the scale of AI models that can be trained and served efficiently. Each generational improvement in HBM technology enables larger parameter counts, longer context windows, and more sophisticated multimodal capabilities in frontier AI systems.
The geopolitical dimensions of SK Hynix's position are significant. As detailed in the HBM Dominance Deep Dive, Korea's control of the HBM supply chain gives it asymmetric leverage in the global AI infrastructure race. Neither the United States nor China can currently produce HBM at scale domestically. This supply chain position is a core strategic asset underpinning the K-Moonshot initiative's credibility and gives Korea bargaining power in the Korea-US chip alliance negotiations.
SK Telecom: AI-Native Telecommunications
SK Telecom unveiled its AI Native strategy at Mobile World Congress (MWC) 2026, signaling a fundamental reorientation of the company from a traditional telecommunications provider to an AI infrastructure and services company. The centrepiece of this strategy is a 519-billion-parameter AI model, one of the largest models developed by a Korean entity, designed to power a new generation of intelligent services across the SK ecosystem.
The company has committed to a trillion-won AI investment programme encompassing AI model development, data center infrastructure, and enterprise AI deployment. SK Telecom's data center expansion targets exceed 1GW of total capacity, a figure that places it among Asia's largest AI compute infrastructure operators. This capacity is essential for the sovereign AI computing objectives embedded in Mission 7: General-Purpose Physical AI Models and Computing Platforms, which requires that Korea maintain domestic compute infrastructure sufficient to train and deploy frontier AI systems without dependence on foreign cloud providers.
SK Telecom has deployed over 2,000 AI agents across enterprise and consumer applications, establishing an early-mover position in agentic AI deployment. These agents handle tasks ranging from customer service automation to network optimization, creating a real-world deployment environment that generates continuous training data and identifies practical bottlenecks in AI agent reliability, safety, and performance.
OpenAI Strategic Partnership
SK Telecom's strategic partnership with OpenAI represents one of the most significant international AI collaborations involving a Korean company. The partnership provides SK Telecom with access to OpenAI's frontier model capabilities while offering OpenAI a channel into the Korean market and, through SK's infrastructure, a foothold in Asian AI deployment. For K-Moonshot strategists, this partnership raises important questions about the balance between international collaboration and the AI sovereignty objectives that underpin several national missions.
The partnership structure allows SK Telecom to fine-tune OpenAI models on Korean-language data and domain-specific enterprise datasets, creating differentiated AI services that leverage both OpenAI's foundational capabilities and SK's local market knowledge. This approach contrasts with the fully domestic model development strategies pursued by Naver (HyperCLOVA X) and Kakao (Kanana), illustrating the diversity of AI strategies within the K-Moonshot corporate ecosystem.
SK On and SK Innovation: Energy for AI Infrastructure
SK On, SK Group's battery manufacturing subsidiary, has expanded into energy storage systems (ESS) and LFP (lithium iron phosphate) batteries specifically designed for AI data center power management. As AI data center power consumption grows exponentially, the need for reliable, cost-effective energy storage becomes a critical infrastructure challenge. SK On's pivot toward data center ESS solutions positions the company at the intersection of energy and AI infrastructure.
SK Innovation, the group's energy holding company, brings expertise in refining, petrochemicals, and increasingly in green energy and hydrogen. The convergence of SK's energy capabilities with its AI infrastructure ambitions creates a vertically integrated approach to the power challenge that constrains AI compute scaling globally. Data centers targeting the 1GW+ capacity that SK Telecom envisions require power supply arrangements that few companies can orchestrate independently.
This energy-AI nexus connects SK Group to K-Moonshot's Future Energy sector and indirectly to Mission 4: Korean Fusion Demonstration Reactor. While SK Group is not directly involved in fusion research, the long-term energy demands of Korea's AI infrastructure build-out provide a compelling demand-side argument for the fusion programme's strategic rationale.
Quantum Computing Interests
SK Telecom has maintained a strategic interest in quantum computing, including an investment in IonQ and the development of quantum-secure communications protocols. While SK's quantum activities are smaller in scale than its AI and semiconductor operations, they contribute to Mission 12: Error-Correcting Quantum Computers and position the group for the eventual convergence of quantum computing and AI workloads.
SK Telecom's quantum-secure communications research is particularly relevant as AI systems increasingly handle sensitive data across telecommunications networks. The intersection of quantum cryptography and AI data security represents a niche but strategically important area where SK Telecom's telecommunications expertise provides a natural advantage.
AI Talent and Research Ecosystem
SK Group's AI talent strategy operates through multiple channels. SK Telecom's AI research centre, T-Brain, has been a significant employer of Korean AI researchers. SK Hynix's research operations in Icheon and its collaboration with university laboratories feed the Mission 10: World-Class AI Scientists pipeline. The group's scholarship and fellowship programmes at Korean universities, including KAIST and Seoul National University, contribute to the broader AI talent pipeline that K-Moonshot seeks to expand.
SK Group's international research presence, including SK Hynix's design centres in the United States and SK Telecom's AI research collaborations, provides a channel for Korean researchers to access global knowledge networks while maintaining ties to the domestic innovation ecosystem. This international connectivity is an asset for K-Moonshot's talent objectives, provided it does not accelerate brain drain dynamics that the programme is designed to reverse.
Risk Factors and Challenges
SK Hynix's dominant HBM position, while a significant strength, carries concentration risk. A technological disruption in memory architecture, a shift in AI hardware design that reduces HBM dependence, or aggressive capacity expansion by competitors could erode SK Hynix's market share and pricing power. The semiconductor industry's cyclical nature means that the current AI-driven demand boom may eventually moderate, creating margin pressure.
SK Telecom's trillion-won AI investment programme represents a significant capital commitment with uncertain returns. The telecommunications industry globally is grappling with the challenge of monetizing AI capabilities at a rate that justifies the infrastructure investment. SK Telecom's OpenAI partnership, while strategically valuable, also creates dependency on a foreign AI provider whose own strategic priorities may not always align with Korean sovereignty objectives.
The US export control regime poses a persistent risk to SK Hynix's operations, particularly regarding its DRAM manufacturing facility in Dalian, China. While SK Hynix has received temporary exemptions, the long-term trajectory of US-China technology decoupling creates strategic uncertainty that affects capacity planning, supply chain design, and customer relationships.
SK Group's complex corporate structure, with multiple listed entities and cross-shareholding arrangements, introduces governance complexity. Chairman Chey Tae-won's leadership has provided strategic direction, but succession planning and governance reform remain areas of investor scrutiny that could affect the group's K-Moonshot commitments.
Outlook and K-Moonshot Significance
SK Group's K-Moonshot significance derives primarily from its control of the AI memory supply chain through SK Hynix and its construction of Korea's AI-native telecommunications infrastructure through SK Telecom. These two capabilities address foundational requirements of the national AI strategy: domestic hardware supply and sovereign compute infrastructure.
The group's AI investment trajectory, including the trillion-won commitment, 1GW+ data center capacity, 2,000+ AI agents, and the 519-billion-parameter model, demonstrates a commitment to AI transformation that extends beyond incremental product improvement. SK Group is repositioning itself as an AI infrastructure conglomerate, a strategic pivot that aligns closely with K-Moonshot's vision of establishing Korea as a global AI power.
For institutional observers tracking the K-Moonshot Corporate Partnership, SK Group's performance metrics in HBM market share, data center deployment timelines, and AI agent adoption rates provide quantitative benchmarks for assessing the programme's industrial mobilization effectiveness. SK Hynix's ability to maintain its HBM leadership through 2027-2028 will be a critical early indicator of whether Korea can sustain the hardware foundation upon which the entire K-Moonshot edifice depends.