Company Offers an Expanded Range of Large-Capacity Liquid Cooling Solutions
for High-Density AI Infrastructure
News Summary
SEOUL, Sep. 22, 2026 — LG Electronics (LG) has qualified its 2.6-megawatt (MW) Coolant Distribution Unit (CDU) as NVIDIA DSX Ready CDU. The product meets the applicable NVIDIA DSX reference design requirements for AI factory infrastructure. The NVIDIA DSX AI Factory Platform unifies AI factory design and operations across compute, networking, power, cooling, facilities and software. Qualified to meet NVIDIA functional requirements and listed on the NVIDIA DSX AI Factory Marketplace, the 2.6MW CDU expands LG’s ability to address the rapidly growing thermal management requirements of AI infrastructure.
Supporting the Growing Demands of High-Density AI Infrastructure
With next-generation AI accelerators, including NVIDIA Blackwell architecture, continuing to advance, AI server energy consumption and heat emissions are rapidly increasing. Due to server rack power density now reaching hundreds of kilowatts, demand is shifting from conventional cooling configurations deployed at the individual rack level to MW-class centralized, high-capacity CDUs capable of managing thermal loads across multiple high-density server racks.
LG has completed the infrastructure qualification of the 600kW, 1MW and 2.6MW models. The 2.6MW model achieved the additional qualification requirements to be DSX Ready. The NVIDIA DSX AI Factory Platform unifies AI factory design and operations across compute, networking, power, cooling, facilities and software.
With the qualification of its 2.6MW model, categorized as a “Large-Scale” solution by NVIDIA, LG is further strengthening its position as a global manufacturer capable of supporting the liquid cooling CDU needs of AI factory infrastructure.
Managing Substantial Thermal Loads
A key component of a direct-to-chip (DTC) liquid cooling system, a CDU circulates coolant through cold plates to remove heat from CPUs and GPUs and transfer it to the data center’s cooling infrastructure. LG’s 2.6MW CDU is designed to efficiently manage the substantial thermal loads generated by large-scale AI server clusters. This high-capacity solution helps data center operators make more efficient use of available space while supporting operational continuity.
As AI infrastructure becomes increasingly power- and compute-intensive, the adoption of liquid cooling is quickly accelerating. According to global market research firm TrendForce, the market penetration of liquid cooling for AI chips is expected to rise from 14 percent in 2024 to approximately 60 percent by 2027.
Strengthening “Chip-to-Chiller” Capabilities for AI Data Centers
The qualification of the 2.6MW CDU highlights LG’s comprehensive “Chip-to-Chiller” portfolio, which addresses thermal management across the entire AI data center cooling ecosystem.
The company’s portfolio encompasses high-efficiency chillers, Computer Room Air Handlers (CRAHs) and liquid cooling solutions designed to dissipate the heat generated by high-performance AI servers. By combining these technologies, LG is able to provide customized thermal management solutions that cover the whole cooling process, from absorbing heat at the AI chip level to ultimately rejecting that heat from the data center facility.
“NVIDIA’s qualification of our 2.6MW CDU for AI factory infrastructure reinforces our position as a first mover in the next-generation liquid cooling market,” said James Lee, president of the LG Eco Solution Company. “We will continue to supply AI data center operators with highly reliable, in-house developed cooling solutions from our comprehensive Chip-to-Chiller portfolio.”
# # #
