LS Electric chairman Koo Ja-kyun outlines a two-way vision linking AI and the power grid

Power & GridBrief

Summary

LS Electric chairman Koo Ja-kyun gave the keynote address on next-generation power grids at the 2026 International Conference on Electricity Distribution (CIRED) workshop, held on November 2 at the JW Marriott Dongdaemun Square in Seoul, the company said, as reported by 디일렉.

Koo argued that keeping AI data centers running reliably requires a more dependable grid, and that AI itself should be used to run that grid more efficiently. He framed the two ideas as a mutually reinforcing loop. Under Grids for AI, the network supplies the power AI infrastructure needs. Under AI for Grids, AI tools manage the network.

He identified direct current (DC) and intelligent platforms as the core technologies for the next-generation grid. The available text ends at that point. It gives no product details, investment figures, customers or deployment timelines.

Why it matters

A major Korean electrical equipment maker is publicly tying its grid strategy to AI data center load. That points to where distribution-level suppliers see demand forming. The emphasis on DC architecture matters to operators weighing power distribution designs for dense AI halls. For global hyperscalers and colocation players building in Korea and the wider Asia-Pacific, it signals that local switchgear and grid vendors are positioning around DC and software-managed networks.

Expert Lens

Keynote visions rarely change procurement on their own. A push toward DC distribution, however, would touch rack power design, protection equipment and safety standards that are still maturing for data center use. Adoption is likely to hinge on whether utilities and building codes in Korea accept DC architectures at scale, which may slow early deployments. AI-driven grid management could help ease interconnection constraints, but only if grid operators trust it with real-time control. The next signal to watch is whether LS Electric announces a named DC power or grid-software contract tied to a specific AI data center project.

Sources

This brief was drafted with AI assistance from public sources; every number is attributed to the sources above. See the AI-assisted Disclosure and Methodology.