SK Telecom Rebuilds Core Infrastructure Around AI Agents
SK Telecom rebuilds entire infrastructure around AI agents, upgrading to trillion-parameter LLM and deploying autonomous systems across billing, network ops, and customer service.
SK Telecom is restructuring its entire technology stack around AI agents, moving beyond surface-level integrations to rebuild core systems from the ground up. The Korean telecom giant's strategy involves rewriting internal systems, expanding data center capacity to gigawatt scale, and upgrading its sovereign LLM to over one trillion parameters.
This represents a fundamental shift in how telecom operators approach enterprise AI. Rather than bolting AI features onto existing infrastructure, SK Telecom is treating AI as the foundational operating layer across billing, customer service, and network operations.
Infrastructure Overhaul Strategy
The centerpiece of SK Telecom's transformation is a complete rebuild of integrated IT systems. The company is redesigning core business functions to be AI-native from the start:
- Sales systems — rebuilt for real-time personalization based on usage patterns
- Line management — automated provisioning and optimization
- Billing systems — dynamic pricing models driven by AI analysis
- Zero Trust security — AI-based monitoring and access controls
This approach differs from typical enterprise AI deployments that layer intelligence on top of legacy systems. SK Telecom is rewriting the foundational code that runs its business operations.
Autonomous Network Operations
The company's autonomous agents strategy extends deep into network infrastructure. AI systems now handle wireless quality management, traffic control, and equipment operations without human intervention.
Key technical implementations include AI-RAN technology for latency reduction and automated fault detection across the network core. These agents operate at the infrastructure level, making real-time decisions about resource allocation and performance optimization.
Customer-Facing Agent Integration
SK Telecom is deploying an integrated AI agent across customer touchpoints. The system connects experiences between T world (main portal) and T Direct Shop (e-commerce platform).
The agent architecture includes:
- Usage pattern analysis — continuous monitoring of customer behavior
- Cross-channel recommendations — personalized suggestions across all platforms
- AI Contact Center — real-time assistance for support representatives
- Retail integration — in-store recommendations based on digital history
Hyperscale Infrastructure Buildout
SK Telecom is constructing hyperscale AI data centers targeting over 1 gigawatt capacity across Korea. This positions the infrastructure to support both domestic AI workloads and attract international investment.
The company already operates the Haein GPU cluster and has deployed Petasus AI Cloud for GPU-as-a-service offerings. The virtualization solution launched last year and is now expanding globally.
Sovereign AI Model Development
SK Telecom's foundation model currently operates at 519 billion parameters—the largest in Korea. The planned upgrade to over one trillion parameters will add multimodal capabilities for processing image, voice, and video data.
Unlike generic LLMs, this model is being optimized for telecom-specific use cases. The company is developing manufacturing-focused AI packages with SK hynix that analyze process data in real-time to reduce defect rates.
Internal AI Transformation
The organizational restructure includes systematic tracking and governance of AI deployment across departments. SK Telecom has implemented:
- AX Dashboard — monitoring AI usage across teams and individuals
- AI Board — executive oversight of transformation initiatives
- AI playground — no-code agent development environment for employees
Over 2,000 AI agents are currently deployed across marketing, legal, and public relations functions. This internal adoption provides real-world testing for customer-facing implementations.
Manufacturing and Enterprise Applications
Beyond telecom services, SK Telecom is packaging its AI infrastructure for enterprise clients. The manufacturing-focused solution with SK hynix demonstrates how telecom operators can expand into industrial AI applications.
The offering includes infrastructure, model access, and complete solutions—positioning SK Telecom as an AI platform provider rather than just a connectivity company.
Bottom Line
SK Telecom's approach represents a blueprint for enterprise AI transformation that goes beyond adding intelligent features. By rebuilding core systems around autonomous agents and treating AI as the operating model rather than an add-on, the company is demonstrating how established enterprises can restructure for an AI-native future.
The success of this transformation will depend on execution at scale and the ability to maintain service quality during the transition. For other enterprises watching, the key insight is treating AI as foundational infrastructure rather than supplementary tooling.