LG Electronics has been formally added to Nvidia’s partner network for power and cooling, extending the Korean company’s push into the rapidly growing market for AI data center thermal management.
The move follows Nvidia qualification of LG’s coolant distribution units, including a 2.6-megawatt system designed for high-density AI server deployments. LG said the latest unit meets applicable Nvidia DSX reference design requirements for AI factory infrastructure.
Cooling is becoming a central AI infrastructure problem
Modern AI clusters use large numbers of high-performance GPUs packed into increasingly dense server racks. That concentration produces heat loads that conventional air-cooling systems can struggle to manage efficiently.
LG is positioning liquid cooling as a core part of its data center business. Its “Chip-to-Chiller” portfolio includes chillers, coolant distribution units and cold plates designed to move heat away from AI accelerators and server racks.
From 600 kilowatts to 2.6 megawatts
LG previously received Nvidia validation for a 600-kilowatt coolant distribution unit and later expanded the portfolio with 1-megawatt and 2.6-megawatt systems. The larger products are aimed at the next generation of AI factories, where power density is expected to rise as newer accelerator platforms are deployed.
Nvidia’s partner network brings together companies supplying the hardware and infrastructure required to build AI data centers, including computing, networking, power and cooling systems.
AI spending is creating a new market for industrial technology
The growth of generative AI has pushed demand beyond chips and servers. Power delivery, cooling equipment and energy efficiency have become major constraints for operators planning new data centers.
LG’s inclusion in Nvidia’s official ecosystem reflects that shift. The company is using its experience in commercial HVAC and industrial systems to compete for infrastructure projects that may require megawatt-scale cooling capacity.
For data center operators, the practical question will be whether suppliers can deploy these systems quickly enough to keep pace with AI server installations while controlling electricity and water use. The cooling market is becoming one of the less visible but increasingly important parts of the AI investment cycle.
