LG Electronics has secured NVIDIA certification for its 600kW Coolant Distribution Unit (CDU), validating the solution’s performance, reliability and compatibility with high-density artificial intelligence infrastructure.

The company said the unit was assessed against more than 100 technical criteria covering cooling performance, operational reliability and failover capabilities for AI factories. The certification positions LG among a select group of suppliers verified to support NVIDIA-based AI infrastructure.

LG’s CDU uses Direct-to-Chip liquid cooling to deliver coolant directly to heat-generating components, including graphics processing units and central processing units. The technology works alongside conventional air-cooling systems to manage the increased thermal loads generated by AI servers.

The unit offers temperature control precision of ±0.25 degrees Celsius and incorporates virtual sensors, real-time monitoring and leak detection. When integrated with LG’s Data Center Cooling Control Manager, it supports predictive control, centralised monitoring and connectivity with data centre management platforms.

LG evaluates its cooling systems at its Chip-to-Chiller validation facility in Pyeongtaek, where it tests temperature fluctuations, system responses and operational stability across the complete cooling chain. The company’s portfolio includes chillers, CDUs and cold plates for end-to-end thermal management.

LG plans to pursue NVIDIA certification for its larger CDU models, including 1MW, 2.5MW and 4MW units. The company has recently secured data centre cooling projects in North America and at the SM+ Data Center in Jakarta, Indonesia.

“As AI data centres continue to scale, the need for highly reliable liquid cooling solutions has never been greater,” said James Lee, President, LG ES Company. He added that LG is expanding its AI infrastructure business through integrated cooling solutions designed to support physical AI applications.