Samsung, SK, Hyundai expand AI partnerships with Big Tech firms to $950 bil.

President Lee Jae Myung raises a beer during a dinner with global artificial intelligence industry leaders in San Francisco, Friday (local time). From left are SK Group Chairman Chey Tae-won, presidential chief of staff for policy Kim Yong-beom, Microsoft Azure Hardware System President Rani Borkar, Nvidia CEO Jensen Huang, Lee, Broadcom CEO Hock Tan, Samsung Electronics Executive Chairman Lee Jae-yong and Hyundai Motor Group Executive Chair Chung Euisun. Yonhap

Samsung, SK Group, Hyundai Motor Group and other Korean tech conglomerates expanded their artificial intelligence (AI) and semiconductor partnerships with U.S. Big Tech companies to $950 billion during President Lee Jae Myung's visit to San Francisco over the weekend.

According to presidential chief of staff for policy Kim Yong-beom, Korean companies and Big Tech firms agreed to pursue a combined $950 billion worth of semiconductor and AI partnerships, on the occasion of the San Francisco AI Summit on Friday (local time).

Among them, Samsung Electronics signed a $200 billion memorandum of understanding (MOU) with Broadcom, under which Samsung's memory and foundry businesses will supply AI-specific high-bandwidth memory (HBM) chips and logic semiconductor manufacturing services for Broadcom's AI accelerators through 2030.

Broadcom is one of the world's leading fabless companies specializing in application-specific integrated circuit chips, designing custom processors tailored to the needs of individual hyperscalers and cloud service providers.

As such customers increasingly seek AI accelerators designed to their own specifications, demand is also rising for memory chips customized to support those processors.

Samsung Electronics Executive Chairman Lee Jae-yong, right, and Samsung Electronics Foundry President Han Jin-man, second from right, poses with Broadcom CEO Hock Tan, left, with President Lee Jae Myung overseeing their signing of a chip partnership memorandum of understanding at the San Francisco AI Summit, Friday (local time). Yonhap

Samsung began mass production and shipments of the industry's first HBM4 featuring a 4-nanometer base die capable of supporting customized demands in February. In May, it became the first in the industry to provide samples of the upgraded HBM4E to customers.

Samsung said it plans to leverage those technologies to supply HBM and other advanced memory solutions to maximize Broadcom’s AI accelerators.

Samsung foundry will apply its advanced sub-2-nanometer process technology to Broadcom's key product lineup, including networking semiconductor solutions supporting next-generation high-speed data communications, as well as offering packaging services using 2-nanometer process.

Samsung said the two sides will closely cooperate in every stage of chip making process — from chip design and fabrication process optimization to advanced packaging and mass production.

"In the era of AI, semiconductor solutions that tightly integrate memory, logic chips and advanced packaging are becoming increasingly important," Samsung Electronics Vice Chairman and Head of Device Solutions Jun Young-hyun said. "By expanding our collaboration with Broadcom, we will accelerate the development of future AI infrastructure and deliver greater value to our customers."

Samsung SDS, meanwhile, inked a strategic partnership with Anthropic to deploy Claude across domestic businesses and jointly expand enterprise AI services in the country.

Under the agreement, the companies will collaborate to pursue new business opportunities in Korea, train engineers to deploy and operate Anthropic's Claude models and support Samsung affiliates' transition to AI-native operations through Claude Enterprise.

Samsung SDS said the partnership will include joint marketing, proof-of-concept projects and business development initiatives aimed at expanding enterprise AI adoption across industries. It will also launch a training program, Applied AI Engineers, to train specialists to support the deployment, integration, operation and technical support of Claude models for enterprise customers.

SK Group Chairman Chey Tae-won, right, talks to Nvidia CEO Jensen Huang during the San Francisco AI Summit on Friday (local time). Yonhap

SK Group expands ties with Nvidia, Microsoft

SK Group expanded its AI infrastructure partnerships with Nvidia, Microsoft, Anthropic and Amazon Web Services (AWS) on the sidelines of the event.

It signed a letter of intent with Nvidia for a comprehensive partnership worth more than $500 billion to build AI infrastructure, spanning from AI data center to AI memory supplies.

Under the agreement, SK Telecom will receive Nvidia's next-generation graphics processing units (GPUs) and pursue investment cooperation to build Korea's largest AI data center with a planned capacity of 2 gigawatts.

SK Telecom plans to use Nvdia’s AI data center architecture DSX and Vera Rubin computing system to build and operate the data center and others beginning in 2027.

SK Telecom also signed an MOU with Anthropic, agreeing on the basic framework for cooperation to build gigawatt-scale AI data centers in Korea, as well as expanding existing ties with AWS.

SK hynix expanded its long-term AI memory partnership with Nvidia. The two companies plan to jointly develop and optimize next-generation AI memory, including HBM, to meet growing infrastructure demand.

SK hynix and Microsoft also agreed to pursue a long-term memory supply partnership. The two companies agreed to explore ways to provide Microsoft with server memory optimized for AI workloads over the medium to long term.

“Competitiveness in the AI era depends not only on how well AI is utilized, but also on how much intelligence can be generated," SK Group Chairman Chey Tae-won said. "Building on SK Telecom's AI infrastructure capabilities and SK hynix's AI memory technologies, we will work with Nvidia to build a world-class AI factory and help Korea evolve from a consumer of AI into a global hub driving AI innovation."

Hyundai Motor Group Executive Chair Chung Euisun, left, poses with Nvidia CEO Jensen Huang during the former's visit to the Nvidia headquarters in Santa Clara, Calif., Friday (local time). Courtesy of Nvidia

Hyundai Motor bets on physical AI

Hyundai Motor Group used the San Francisco event to declare its ambition to become a physical AI specialist.

During the event, Hyundai Motor Group Executive Chair Chung Euisun said the group is transforming itself from a traditional automaker into a provider of physical AI solutions spanning autonomous driving, robotics and AI factories.

The group's vision begins with making vehicles, robots and other individual devices more intelligent, and then extends to AI factories that connect entire manufacturing processes through AI. Its ultimate goal is creating a city-scale integrated intelligence.

To achieve that vision, Chung said Hyundai Motor Group will combine its strengths in manufacturing, mobility and robotics with the software, AI infrastructure and algorithm capabilities of global Big Tech companies such as Nvidia, Waymo and Google DeepMind.

With Nvidia, Hyundai Motor will jointly develop an AI Robot Reference Platform and establish a Robot Application Center in Korea.

Hyundai Motor Group will also continue its autonomous driving foundry partnership with Waymo using IONIQ 5 vehicles produced at Hyundai Motor Group Metaplant America in Georgia.

Naver is partnering with Nvidia and Canadian investment firm Brookfield to build a 200-megawatt AI factory with a capacity for about 100,000 GPUs at its GAK Sejong hyperscale data center after securing a combined $10 billion in investment and financing from the two companies.

Nvidia plans to invest $1 billion in the project, while Brookfield has signed a nonbinding agreement to provide up to $9 billion in financing. Naver will raise the remaining funds needed for the project.

The company said it plans to eventually expand the facility's capacity to 1 gigawatt. It will deploy Nvidia's Vera Rubin and Blackwell platforms.

“Nvidia's strategic investment and an infrastructure agreement with Brookfield marked the start of full-scale execution of the company's AI factory strategy,” Naver’s founder and Chairman Lee Hae-jin said.