South Korea does not want to remain the country that simply manufactures the hardware powering someone else’s artificial intelligence.
President Lee Jae Myung made that clear during a high-level AI summit in San Francisco, where some of the most influential names in technology gathered alongside South Korea’s biggest corporate leaders.
The meeting included Nvidia CEO Jensen Huang, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei and Broadcom CEO Hock Tan. Lee’s message to them was straightforward: South Korea wants deeper cooperation, bigger investments and a central position in the next phase of global AI development.
Lee Brings Global AI Leaders Into One Room
The San Francisco AI Summit was more than a ceremonial stop during Lee’s US visit.
Before attending the main event, the South Korean president held separate meetings with the leaders of Nvidia, OpenAI, Anthropic and Broadcom. Those conversations focused on how the companies could take a greater role in South Korea’s AI investment and infrastructure projects.
Around 150 executives, researchers, investors and startup founders reportedly attended the summit. South Korean business figures were also heavily represented, including Samsung Electronics Executive Chairman Jay Y. Lee, SK Group Chairman Chey Tae-won, Hyundai Motor Group Executive Chair Euisun Chung and Naver founder Lee Hae-jin.
That mix matters. South Korea already has the chipmakers, industrial groups and manufacturing capacity. US technology companies bring the models, software platforms and global cloud demand. Lee is trying to connect those parts before another country does.
The San Francisco AI Declaration Sets the Direction
During his keynote address, Lee introduced what his administration called the San Francisco AI Declaration.
The declaration outlines a plan for broader technological cooperation between South Korea and the United States. It also places South Korea’s semiconductor industry at the centre of the country’s AI ambitions.
Lee presented South Korea as a future anchor in the global AI supply chain rather than a secondary supplier. The country wants to help technology companies secure the chips, memory and infrastructure required to train and operate increasingly demanding AI systems.
That is not a small distinction.
Plenty of countries have published national AI plans. Far fewer can back those plans with large-scale semiconductor production, automotive manufacturing, robotics expertise, data centre investments and globally recognised technology brands.
South Korea can.
Samsung and SK Hynix Give South Korea Real Leverage
The AI race keeps returning to the same physical problem: computing capacity.
Developers may build more capable models, but those models still need advanced processors, high-bandwidth memory and enormous data centre systems. Demand continues to grow faster than some parts of the supply chain can respond.
Samsung Electronics and SK Hynix give South Korea unusual leverage in that environment. Together, the companies are major suppliers of the memory products used in AI servers and accelerator systems.
South Korean officials have argued that the production capabilities of these companies could help global technology firms build AI infrastructure more reliably. Recent announcements have also linked Korean and US companies to semiconductor partnerships potentially worth hundreds of billions of dollars over several years.
The numbers are huge. They should also be treated carefully until individual contracts, timelines and investment commitments become clearer.
Still, the direction is obvious. AI companies need chips, and South Korea wants to become harder to bypass.
Nvidia, OpenAI and Anthropic Need Different Things From Korea
Not every company at the summit arrived with the same priorities.
Nvidia depends on a vast manufacturing ecosystem to support demand for its AI accelerators. Memory supply, advanced packaging and data centre construction all shape how quickly those systems can reach customers.
OpenAI and Anthropic need access to larger pools of computing power as they train and deploy more capable models. Both companies also benefit when governments and industrial groups create new markets for enterprise and public-sector AI services.
Broadcom sits in another part of the infrastructure stack, supplying networking and custom semiconductor technology used inside modern data centres.
South Korea can engage with all of them, but through different doors. Samsung and SK Hynix cover semiconductors and memory. Naver brings domestic AI and cloud capabilities. Hyundai adds robotics, autonomous systems and physical AI. The government can support the energy, research and regulatory infrastructure surrounding those projects.
This is less like one partnership and more like a collection of overlapping deals.
Physical AI Could Become South Korea’s Strongest Opening
Generative AI usually dominates the headlines, yet South Korea may have a better opportunity in physical AI.
The term covers systems that allow machines, robots, vehicles and industrial equipment to understand their surroundings and act with less human intervention. It is where foundation models begin moving out of browsers and into factories, warehouses and streets.
South Korea already has major companies operating across automobiles, electronics, shipbuilding, batteries, robotics and heavy industry. That industrial base could turn the country into a large testing ground for AI systems that interact with the physical world.
Hyundai’s participation in the summit was therefore not incidental. Neither was Samsung’s.
A country filled with advanced factories can do more than purchase AI software. It can test whether that software survives contact with actual machines, production schedules and safety requirements.
South Korea Wants to Move From AI Follower to Supply-Chain Leader
Lee has framed the rise of AI as a chance for South Korea to stop following larger technology powers and begin shaping the market itself.
That will take more than diplomatic meetings.
The country still faces pressure around energy supply, access to specialised talent, data centre capacity and competition from the United States, China, Japan and Europe. Domestic AI developers also need room to grow alongside international companies rather than becoming customers of foreign platforms by default.
The San Francisco summit does not solve those problems.
It does, however, reveal South Korea’s strategy. The government wants to use the country’s semiconductor strength as the starting point, then expand into models, cloud infrastructure, robotics and industrial applications.
Chips open the conversation. The larger goal is influence over the entire AI stack.
The Summit Turns AI Diplomacy Into Industrial Policy
Technology summits often produce polished declarations and vague promises. This one carried more weight because the companies involved already depend on one another.
US AI firms need Korean semiconductor capacity. South Korean manufacturers need long-term customers for expensive new facilities. Domestic conglomerates want access to leading models and computing platforms. The government wants investment, jobs and a stronger geopolitical position.
Those interests will not always align neatly. They do not have to.
South Korea’s AI push looks practical precisely because it begins with existing commercial pressure rather than a distant vision. Memory shortages, rising computing costs and demand for new data centres are already forcing companies to form closer partnerships.
Lee is trying to turn that pressure into a national advantage.
What Comes Next for South Korea’s AI Strategy
The next stage will be less glamorous than the summit photographs.
Investors will watch for signed supply agreements, new research centres, data centre projects and direct commitments from companies such as Nvidia, OpenAI and Anthropic. Details around funding, energy requirements and construction schedules will matter more than the headline value attached to broad cooperation plans.
South Korea also needs to show that international partnerships can strengthen its domestic AI ecosystem. A strategy built entirely around supplying chips to foreign companies would bring revenue, but it would not automatically make the country an AI leader.
The opportunity sits in the middle: supply the infrastructure, develop local models, build industrial applications and persuade global companies to conduct more of their research and deployment work inside South Korea.
That is the new era Lee is talking about. Not just faster chips leaving Korean factories, but more of the AI economy taking shape around them.
Sources
- Reuters – South Korea President Lee Looking to Open New Era of AI With Global Tech Companies
- Yonhap News Agency – Lee Meets OpenAI and Nvidia CEOs
- The Korea Times – Lee Declares Vision to Make Korea an AI Supply-Chain Leader
- The Korea Herald – Korean and US Tech Companies Pursue Chip Partnerships
- SBS News – South Korea and Global Tech Firms Outline Semiconductor Cooperation