LG expands AI data centre chiller production in US

LG Electronics is expanding chiller production for AI data centres with a new factory in Virginia and additional capacity in South Korea, extending its manufacturing footprint for the equipment to three sites.

The company plans to build the new air-cooled chiller facility in Windsor, Virginia, while increasing output at its existing plants in Changwon and Pyeongtaek. Total investment across the three sites will amount to KRW 150 billion.

Chillers are a key part of data centre cooling systems, and rising computing density has increased the need to manage heat. Demand has been particularly strong in facilities built for artificial intelligence workloads, which require more power and generate greater thermal loads than many conventional data centre operations.

The Virginia plant will have 30,000 square metres of gross floor area on a 170,000-square-metre site. Production is due to begin in the first half of 2027, with the factory making air-cooled chillers for AI data centres in North America.

The US location places LG close to a large concentration of data centre operators. Virginia has become one of the sector's most important hubs, making local supply chains and response times increasingly important for equipment vendors.

North America focus

Figures from Omdia cited by LG show global data centre capacity is expected to rise from 226GW in 2026 to 420GW in 2030, with 60 per cent of the additional capacity concentrated in North America.

That outlook helps explain the decision to add manufacturing in the United States rather than rely solely on exports from Asia. Local production can shorten delivery times and may reduce some supply chain risks for customers building or expanding facilities quickly.

At its South Korean operations, LG is expanding air-cooled chiller production lines in Pyeongtaek and Changwon. It is also introducing a conveyor-based production system for higher-volume manufacturing, with products moving through the process on a single conveyor.

The change is intended to improve operational efficiency and support a faster response to rising order volumes. The expansion in Korea will complement, rather than replace, output from the new US facility.

Cooling portfolio

Beyond expanding manufacturing, LG is presenting the investment as part of a broader approach to AI data centre cooling. The company describes the strategy as "Chip-to-Chiller", linking chillers with liquid cooling products including Coolant Distribution Units and cold plates that remove heat directly from chips.

Liquid cooling has drawn growing attention across the industry as operators seek more effective ways to manage dense AI server clusters. Air cooling remains common in many facilities, but direct chip-level cooling is becoming more relevant for racks running high-intensity AI workloads.

LG said its product range includes a 2.6MW CDU that qualified as an NVIDIA DSX Ready CDU. It has also joined the NVIDIA Partner Network as a Power and Cooling Solution Preferred Partner.

Those links to NVIDIA place LG within a wider ecosystem forming around infrastructure for AI computing. Technology suppliers, power equipment providers and thermal management companies have all sought closer ties with chipmakers and server vendors as demand for specialised data centre systems grows.

Broader strategy

In Asia, LG is applying a "One LG" approach that combines heating, ventilation and air conditioning products with battery systems and data centre design, build and operation work from other affiliates. The approach is being used in a hyperscale AI data centre project for Sinar Mas Group in Jakarta.

LG has also upgraded its Business Innovation Centre in Singapore to allow partners, engineers and integrators to test and assess solutions. The facility is part of the company's effort to deepen ties with regional customers and channel partners involved in data centre projects.

James Lee, President of the LG ES Company, set out the company's position on the investment. "We will continue to strengthen our core capabilities in AI data center cooling solutions, from differentiated product technologies to production capacity," said Lee. "With these capabilities, we will ensure a stable supply and actively address rapidly growing market demand."