LG Electronics has strengthened its position in the artificial intelligence infrastructure market after joining the NVIDIA Partner Network (NPN) as a Power and Cooling Solution Preferred Partner, bringing its liquid cooling technology into sharper focus as demand for AI computing capacity grows.
The partnership centres on one of the less visible but increasingly important requirements of AI infrastructure: managing the heat generated by high-performance computing equipment.
AI data centres require substantial computing power to run increasingly demanding workloads, placing greater pressure on cooling systems. LG is expanding its portfolio of liquid cooling solutions to help data centre operators manage these thermal demands while improving efficiency and reliability.
The company’s approach involves integrating cooling systems into facility planning from the early stages, allowing developers to account for power requirements, heat management and supporting infrastructure before construction is complete.
Among its latest developments is a 2.6-megawatt Coolant Distribution Unit (CDU), which has qualified as an NVIDIA DSX Ready CDU. According to LG, it currently has the largest cooling capacity among NVIDIA’s DSX Ready CDU partners.
LG has also completed infrastructure qualification for its 600kW and 1MW CDU models. The qualification process covers technical requirements relating to cooling performance, system failover capabilities and reliability, supporting the use of the equipment in high-density AI environments.
The company’s cooling systems use direct-to-chip liquid cooling, which delivers coolant directly to heat-generating components such as central processing units (CPUs) and graphics processing units (GPUs). The technology works alongside existing air-cooling infrastructure, creating a hybrid system intended to manage heat more efficiently.
The growing importance of such systems comes as technology companies and data centre operators expand infrastructure to support AI services. Cooling has become a key consideration in facility design as operators seek to manage rising thermal loads without compromising equipment performance.
LG plans to showcase its qualified CDU models at Data Centre World Asia 2026, scheduled for September 29 and 30 at Marina Bay Sands in Singapore. The company also plans to pursue NVIDIA qualification for its 4.0MW CDU model by the end of 2026, further expanding its offering for large-scale AI facilities.
“Joining the NVIDIA Partner Network as a Power and Cooling Preferred Partner is a pivotal step in our strategy to support the expansion of the global AI ecosystem,” said James Lee, president of LG’s ES Company.
He said the collaboration would support the development of cooling solutions for some of the industry’s most demanding thermal management challenges.
For LG, the partnership provides another route into the expanding AI infrastructure market, where cooling technology is becoming an increasingly important part of building and operating next-generation data centres.
Never miss a story 💜
Breaking news, real stories and money tips for Kenyan women, as they happen.