LG’s Microsoft and Nvidia AI Push Is About Infrastructure, Not Chips

LG Chairman Koo Kwang-mo is bringing several affiliates into a coordinated AI infrastructure strategy. His first meeting with Microsoft CEO Satya Nadella expanded their companies’ cooperation, while Nvidia’s DSX Ready program qualified LG cooling and battery-storage products. The distinction is commercial: a partnership creates opportunities and a qualification establishes product eligibility, but neither is an announced equipment order.91718

What changed in the Microsoft relationship?

Koo and Nadella met on Monday, September 21, 2026, at Microsoft’s headquarters in Redmond, Washington. Executives from LG Electronics, LG Uplus, LG CNS and LG Sciencepark also attended. The companies agreed to expand cooperation in three areas: AI data-center infrastructure, workplace AI transformation and physical AI.933

The meeting builds on earlier work rather than starting the relationship from scratch. LG Electronics and Microsoft announced an AI data-center cooling alliance in January 2025, and LG and LS affiliates presented a broader infrastructure package at a Microsoft technology showcase in Redmond in December 2025.1921 Those events establish a progression toward the September agreement; the cited material does not establish the details of every intervening negotiation.

For data centers, LG and Microsoft plan to explore opportunities involving LG cooling, power and IT systems. Longer-term possibilities include prefabricated data centers assembled from factory-built modules containing equipment such as power systems, cooling and server racks.533 These are prospective areas of cooperation, not identified Microsoft deployments. The provided reporting does not disclose awarded projects, locations, equipment volumes or contract values.1132

The partnership also runs in the other direction: LG plans to adopt Microsoft AI and cloud tools for office work, including Copilot and Microsoft 365 E7: The Frontier Suite. For physical AI, LG plans to use Microsoft cloud infrastructure to develop and train models for manufacturing, logistics and mobility. Its proposed industrial-data “flywheel” would use operational data to improve models and apply the results back in the field; that is a development plan, not evidence that such a system is already deployed at scale.81516

Where does Nvidia fit?

In the September 22 announcement cycle, LG Electronics reported DSX Ready approval for a 2.5-megawatt coolant distribution unit, while LG Energy Solution said its battery energy-storage system qualified as Nvidia DSX Ready BESS.151718 The approvals concern different needs in an AI data center: liquid cooling for computing equipment and battery storage for power infrastructure.1718

Nvidia describes DSX Ready as a qualification program for products meeting applicable requirements of its AI factory reference designs. Qualification gives customers a way to identify compatible components; it does not mean Nvidia has purchased those products or that sales to other large technology customers are guaranteed.18

The qualifications follow earlier relationship-building. Koo’s meeting with Nvidia CEO Jensen Huang was followed by June discussions about AI infrastructure and robotics, and LG said the companies signed a memorandum of understanding on future business initiatives in August. LG Electronics has also described work with Nvidia on robotics and a manufacturing-data flywheel.242643

Why the distinction matters

LG’s strategy is to supply and develop the systems around AI computing—not to announce an LG processor: cooling equipment, power and battery storage, data-center capabilities, IT services and industrial AI applications. Bringing those affiliates together could make LG a more comprehensive infrastructure partner. For now, the verified milestones are an expanded Microsoft cooperation framework and Nvidia product qualifications. Actual supply contracts would require separate evidence of orders or deployments.9171833