kt cloud Unveils Roadmap to Supply 1GW of AIDC by 2031

At kt cloud summit 2026, held on September 15, kt cloud revealed its infrastructure strategy to support customers' AX (AI Transformation). Over 1,500 attendees, including public institutions and corporate representatives, shared real-world application cases of AI Data Centers (AIDC) and cloud technologies.

kt cloud presented a Composite AI vision centered on kt cloud PLATFORM and announced a roadmap to supply over 1GW of AIDC infrastructure across approximately 20 sites by 2031. This is part of a plan to expand base infrastructure to reliably support large-scale AI computing.

Power Infrastructure Collaboration and Technical Strategy

To address the scaling of AI data centers, kt cloud signed a strategic partnership with LS Electric. The two companies will establish a collaborative framework for the timely supply of key power facilities aligned with construction schedules and jointly research power solutions optimized for Modular Data Centers (MDC). This lays the foundation for rapidly and reliably building large-scale AIDCs.

Additionally, kt cloud is exploring the use of MCP (Model Context Protocol) to resolve the issue of switching between multiple tools when identifying root causes of failures in DevOps environments. The goal is to enhance operational efficiency by enabling AI to access and synthesize information from various systems such as Kubernetes, GitHub, ArgoCD, and Slack via MCP.

AI Service Performance and Cost Optimization

The event also showcased technical cases that balance AI service performance and cost. For RAG (Retrieval-Augmented Generation), methods were presented to improve answer quality and response speed by removing unnecessary context and re-ranking. Furthermore, infrastructure design strategies utilizing HBM, DRAM, and SSD were discussed to resolve memory bottlenecks during AI inference. kt cloud validated performance and cost efficiency through three internalization cases where models were selected and deployed according to specific business characteristics, such as AI Portal, security monitoring, and failure monitoring.