Exclusive: LG’s Koo to Meet Nvidia’s Huang Again Two Months After Seoul Talks

Meeting set for next month at Nvidia’s U.S. headquarters

Return visit follows Jensen Huang’s invitation in June

Working groups from both sides met in the interim

Physical AI and smart factories among topics for discussion

Talks may produce concrete cooperation plans

Koo Kwang-mo, chairman of LG Group, will meet Nvidia Chief Executive Officer Jensen Huang in Silicon Valley next week. The meeting comes two months after they met on June 8 at LG Group’s headquarters in Seoul’s Yeouido district. The two companies have since held working-level talks between South Korea and the US, raising expectations for cooperation in robotics, artificial intelligence infrastructure and smart manufacturing.

Another meeting two months later

Koo is scheduled to meet Huang at Nvidia’s headquarters next week, according to industry officials on Aug. 6. It will be his first trip to Silicon Valley in four months, since April. Koo and Huang held their first official meeting in June at LG Group’s headquarters in Yeouido, Seoul. After about an hour of private talks, they announced plans to expand cooperation in robotics, AI infrastructure and autonomous driving. Koo later told reporters Huang had invited him to California for broader discussions on collaboration.

The two companies have continued their exchanges since that first meeting. On June 22, about 30 LG Group executives visited Nvidia’s headquarters, led by LG Electronics President Ryu Jae-cheol, LG CNS President Hyun Shin-gyoon and LG Science Park President Chung Soo-heon.

With the two sides meeting again so soon, industry officials say more concrete cooperation measures could emerge. On LG Electronics’ second-quarter earnings call on July 30, Chief Financial Officer Kim Chang-tae said a strategic partnership with Nvidia at the AI platform ecosystem level was under consideration.

Robotics, data centers among areas where cooperation is taking shape

Robotics is the area where cooperation has advanced fastest. In June, the companies agreed to jointly develop a reference robot based on Nvidia’s open robot foundation model, Isaac GR00T. The tie-up aligns LG Electronics’ push to expand beyond home appliances into robotics with Nvidia’s need for customers for its robot foundation model and robotics chips. If the reference robot gains widespread adoption, LG Group components such as cameras, actuators and sensors could also become industry standards.

The companies are also working closely in smart manufacturing. The goal is to combine 770 terabytes of manufacturing and production data accumulated by LG Electronics over a decade with Nvidia Omniverse, the chipmaker’s digital twin platform, to build automated factories.

Attention is also focused on whether LG Group’s cooling technology will be used in Nvidia’s AI data centers. LG Electronics said on July 27 that its 600-kilowatt coolant distribution unit had completed validation for Nvidia’s AI infrastructure. A CDU supplies coolant to graphics processing units and central processing units inside servers and helps control heat. It is a key component in the heating, ventilation and air-conditioning industry.

The talks could also cover supplies of Nvidia’s advanced GPUs, which LG Group needs for development of its proprietary EXAONE AI model and for a broader AI push. LG Group is said to be planning to bring in 10,000 Blackwell GPUs from Nvidia. Additional volumes and delivery timing could also come up for discussion.

Kim In-yeop, Silicon Valley correspondent, Korea Economic Daily, inside@hankyung.com