LG joins Nvidia network as AI cooling demand rises

LG Electronics has joined the NVIDIA Partner Network as a Power and Cooling Solution Preferred Partner, placing it within NVIDIA's supplier ecosystem for AI data centre infrastructure.

The partnership comes as demand rises for thermal management systems in high-density AI data centres, where graphics processor-based servers generate heavier heat loads than conventional computing equipment. LG has been expanding its liquid cooling range to meet those requirements.

At the centre of that range is a 2.6-megawatt Coolant Distribution Unit, which has qualified as an NVIDIA DSX Ready CDU. Its 600-kilowatt and 1-megawatt models have also completed infrastructure qualification under NVIDIA's requirements.

According to LG, those qualifications cover cooling performance, failover and reliability. It added that the 2.6-megawatt unit currently has the largest cooling capacity among CDU products listed as DSX Ready partners on the NVIDIA marketplace.

Cooling systems have become a key battleground in the build-out of AI computing sites, as operators seek to manage higher rack densities and control energy use. Liquid cooling has drawn growing interest because it can move heat away from processors more directly than air-based systems alone.

LG said its CDU line uses a direct-to-chip liquid cooling approach, delivering coolant to the CPU, GPU and other heat-generating components. The systems are designed to work alongside existing air-cooling equipment, creating a hybrid arrangement rather than requiring a full replacement of installed infrastructure.

LG's data centre cooling products are intended to be integrated at an early stage of facility planning. That approach gives operators a basis for designing cooling systems and related infrastructure around expected power and thermal demands.

AI infrastructure

The announcement reflects a wider shift by industrial and electronics groups into the data centre supply chain, particularly in areas tied to AI workloads. While chipmakers and server manufacturers have captured much of the attention around AI investment, demand has also risen for suppliers of power systems, cooling equipment and other physical infrastructure needed to run larger computing clusters.

For NVIDIA, preferred partner arrangements help create an ecosystem of approved vendors whose equipment has been tested against its infrastructure requirements. For suppliers such as LG, the designation offers a route into projects led by cloud groups, hyperscalers and other operators building AI-focused facilities.

LG framed the partnership as part of a broader effort to strengthen its position in thermal management for AI data centres. The company is better known globally for consumer electronics and appliances, but it also has a commercial heating, ventilation and air conditioning business that includes chillers and other cooling products.

Its focus on coolant distribution units places it in a segment that is becoming more important as processors consume more power and produce more heat. Operators of AI sites have been weighing a mix of direct-to-chip cooling, immersion systems and upgraded air-cooling designs as they seek to support denser hardware layouts.

Qualification process

NVIDIA's DSX Ready designation indicates that cooling equipment has met a set of technical criteria for use in AI infrastructure deployments associated with the chip designer's systems. LG said its recent qualification extends across multiple CDU sizes, from 600 kilowatts to 2.6 megawatts, giving it a broader product range for different facility requirements.

LG also said it intends to seek NVIDIA qualification for a 4.0-megawatt CDU model. If approved, that would further expand its offering at the top end of data centre cooling demand, where developers are planning larger AI clusters and higher-density installations.

One point of context in LG's account of the partnership was its link between the cooling product range and what it called next-generation AI factories. The term is increasingly used in the sector to describe data centres built primarily for AI training and inference workloads rather than more general enterprise computing tasks.

James Lee, President of the LG ES Company, said joining the NVIDIA Partner Network as a Power and Cooling Preferred Partner was "a pivotal step in our strategy to support the expansion of the global AI ecosystem".

"Together with NVIDIA, LG will help shape the future of AI data centres by delivering advanced cooling solutions that address the industry's most critical thermal challenges," Lee said.