What happened
LG Electronics has officially joined the Nvidia Partner Network (NPN) as a preferred partner in the power and cooling category. This designation follows a series of technical validations for LG’s cooling hardware, including its coolant distribution units (CDUs) designed for high-density AI data centers.
LG Electronics announced its inclusion in the Nvidia Partner Network (NPN) as a preferred partner for power and cooling. The NPN is a global program that connects Nvidia with hardware and infrastructure providers to support the deployment of AI computing systems.
The partnership follows a series of product qualifications. In July, LG secured Nvidia validation for a 600-kilowatt CDU, followed by a 1-megawatt unit. Most recently, LG’s 2.6-megawatt CDU qualified as an 'Nvidia DSX Ready' component, meeting the reference design requirements for Nvidia’s DSX platform for large-scale AI data centers.
LG’s cooling portfolio, which the company markets as 'chip to chiller,' includes chillers, coolant distribution units (CDUs), and cold plates that draw heat directly from AI chips.
Source details: koreaherald.com ↗
Why it matters
As AI accelerators generate increasingly high heat loads, traditional air-cooling methods are becoming insufficient, making liquid cooling essential for large-scale AI infrastructure. By securing NPN status, LG gains deeper integration into Nvidia’s ecosystem, allowing its cooling systems to be considered during the initial design and planning phases of new data centers. This partnership is a strategic move for LG to capture market share among North American hyperscalers and colocation operators, positioning its 'chip to chiller' portfolio as a standard for megawatt-scale AI deployments.
The shift toward high-performance AI servers has created a critical need for advanced thermal management. Liquid cooling is increasingly required to handle the heat density of modern AI accelerators that exceed the capabilities of conventional air-cooling systems.
NPN status provides LG with a competitive advantage by allowing its equipment to be integrated into the design phase of data centers built on Nvidia’s architecture. This is expected to facilitate LG's expansion into the North American market, targeting major hyperscalers and colocation providers.
The 'chip to chiller' approach allows LG to offer a comprehensive cooling chain, from the chillers that produce chilled water to the CDUs and cold plates that manage heat at the server and chip levels.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
What did neural scaling law research (e.g. Kaplan et al., 2020) observe?
What to watch next
The primary focus will be on LG’s ability to secure contracts for large-scale AI data center projects in North America. Market observers should monitor whether this partnership leads to widespread adoption of LG’s 2.6-megawatt CDUs, which recently qualified for Nvidia’s DSX platform, and how the company competes against established thermal management providers in the global AI infrastructure supply chain.
The company has explicitly stated its intent to pursue megawatt-scale AI data center projects in North America. Success will depend on the company's ability to scale production and integrate its hardware into the complex supply chains of major data center operators.
Future updates will likely involve the announcement of specific data center projects utilizing LG’s qualified CDUs, which will serve as a for the effectiveness of this partnership in real-world, high-density environments.