South Korean industrial conglomerates are expanding their partnerships with US chipmaker Nvidia (NVDA) to embed artificial intelligence (AI) across advanced pharmaceuticals, manufacturing, and robotics. The alliances move the country beyond simple semiconductor fabrication into automated industrial operations.

The deals come as South Korea faces severe domestic energy constraints and a hyper-competitive global tech landscape. The country operates an isolated power grid and imports the vast majority of its energy. It relies heavily on nuclear power, which accounts for 31% of domestic electricity generation according to the International Energy Agency, providing the stable baseload power needed to fuel new high-density AI data hubs. Lock-ins with US computational infrastructure are now critical to secure long-term industrial efficiency.

SK Biopharma's AI research

SK Biopharmaceuticals (326030.KS) plans to use Nvidia graphics processing units (GPUs) to transform its entire drug discovery process into an AI-driven operating system, CEO Lee Dong-hoon announced on June 22. The drugmaker will deploy autonomous computational agents to manage molecule design, target identification, and operational analysis, according to Chosun Daily.

The strategy mirrors the Western pharmaceutical model. Nvidia launched an $1bn joint research venture with Eli Lilly in January 2024 to run 24-hour automated lab testing cycles. SK Biopharmaceuticals will feed its own data into a similar continuous learning loop. This includes records from 2,000 compound synthesis units and 2.3mn pages of US Food and Drug Administration (FDA) approval files accumulated during the development of its epilepsy treatment Cenobamate.

Corporate links are tightening. SK Group (034730.KS) Chairman Chey Tae-won met Nvidia CEO Jensen Huang in South Korea recently alongside their daughters, who both hold senior business development roles at the respective firms. Lee was also the only South Korean biotech executive invited to an exclusive Nvidia industry reception earlier this year.

As a first step, SK Biopharmaceuticals signed a KRW4 trillion ($2.6bn) collaborative research contract with Chinese generative AI specialist Insilico Medicine. The deal is the first out of the SK Open Innovation Centre. Insilico will deploy its Pharma.AI platform in the early discovery phase to cut candidate isolation times by 50%. SK Biopharmaceuticals will retain all molecular designs and prediction data to internalise its AI capabilities. The firm will funnel these discoveries through its new LinX innovation hub in New Jersey to commercialise Asian biotech assets in the US market.

Co-Engineering networks, robotics

SK Hynix (000660.KS) has formalised a multi-year technology pact with Nvidia to co-develop advanced hardware. The memory maker is in the process of aligning its product roadmap options with Silicon Valley to help match high-bandwidth memory (HBM) modules together with Nvidia data centre platforms, The Korea Times reports. The company uses Nvidia software libraries to run physics-based chip simulations. This cuts qualification times for next-generation chips and accelerates factory deployment.

Concurrently, SK Telecom (017670.KS) is working with Nvidia to construct gigawatt-scale AI cloud infrastructure and "AI factories" in South Korea. The telecom operator is moving away from general-purpose network services to host dedicated computing hubs optimised for automated workloads. The first specialised facility will go online in 2027 to provide local logistics and tech companies with high-token efficiency per megawatt.

A delegation of 30 senior executives from LG CNS (064400.KS), LG Electronics (066570.KS), and LG AI Research also travelled to California to finalise automated factory and robotics agreements with Nvidia, according to The Korea Times. The summit followed a meeting in Seoul between LG Group (003550.KS.) Chairman Koo Kwang-mo and Nvidia leadership to co-develop reference robots and future mobility platforms. LG will combine Nvidia software with its own home appliances and automotive components to build simulated industrial workflows and test humanoid warehouse systems.

Securing supply lines

Internet giant Naver is itself partnering with Nvidia to deploy localised sovereign AI infrastructure within South Korea, The Korea Times reports. The firm is initiating a 55-megawatt expansion at its hyperscale data facility in Sejong, with long-term plans to reach gigawatt capacity. The infrastructure will run regional foundation models and world-simulation platforms. This guarantees data residency and compliance for local corporate and government clients.

In parallel, Doosan Group is collaborating with Nvidia to engineer specialised physical infrastructure for heavy data processing. The company is designing bespoke cooling methodologies and modular power distribution loops. These systems are built to handle the extreme power densities and thermal loads generated by large graphics processing clusters under continuous industrial workloads.

Broader technology and component supply collaborations remain highly active across South Korea's remaining industrial heavyweights. Samsung Electronics (005930.KS) is working with Nvidia on next-generation foundry processes and advanced chip packaging solutions to secure HBM supply lines. At the same time, Hyundai Motor Group is integrating Nvidia high-performance computing architecture directly into its vehicles to power autonomous driving models and connected transport systems.