A wave of collaboration between South Korean firms and NVIDIA signals a shift towards a comprehensive AI infrastructure, though concerns about dependency on the tech giant remain. Major companies such as SK Hynix, Naver, LG, and Doosan have announced partnership plans with NVIDIA to enhance their capabilities in AI semiconductors, data centers, and robotics.
Major Partnerships in AI Infrastructure
On the 8th of this month, leaders from these firms met with NVIDIA CEO Jensen Huang, who had recently visited Korea. This gathering highlighted their commitment to developing AI technologies, particularly in AI-dedicated semiconductors and digital twin-based AI simulation. NVIDIA, which commands over 90% of the data center GPU market, is using its market position to expand its influence in the AI sector through these partnerships. The collaboration aims to create a full-stack enterprise, covering everything from open-source AI model development to physical AI training platforms.
Among the projects, Naver's initiative to establish a gigawatt-scale AI factory stands out. This facility will utilize advanced GPUs to generate AI tokens, serving as a dedicated data center for AI training and inference. Naver plans to co-design and co-invest in this infrastructure alongside NVIDIA, sharing the associated risks and rewards. The company's ambitions extend beyond the Asia-Pacific region, with a goal of reaching a capacity of 200 megawatts by 2028—equivalent to the output of a nuclear power plant.
Strengthening the Semiconductor Supply Chain
Another key partnership involves SK Group, which has strengthened its collaboration with NVIDIA, a relationship developed through multiple meetings between SK Chairman Chey Tae-won and Huang. SK Hynix will co-develop next-generation memory tailored for AI applications, solidifying its status as NVIDIA’s largest memory partner. The two companies plan to work together on memory for NVIDIA's upcoming AI platforms, ensuring a long-term technological partnership that could transform semiconductor design and manufacturing.
Robotics and Physical AI Advancements
LG Group and Doosan are also advancing in robotics and physical AI, collaborating with NVIDIA to co-develop next-generation humanoid robots. LG aims to enhance its development processes, while Doosan intends to integrate its manufacturing capabilities with NVIDIA’s physical AI platforms. This includes advancing Doosan’s robotics operating systems and exploring energy solutions for NVIDIA's AI factories.
Implications and Future Outlook
Despite the promising nature of these partnerships, analysts are wary of potential over-reliance on NVIDIA’s ecosystem. South Korean companies have championed the concept of ‘sovereign AI’ through their own developments, but aligning these goals with partnerships could lead to dependency on external platforms. Critics have also pointed out the absence of specific investment commitments from NVIDIA, particularly when compared to pledges made in other regions.
As these collaborations unfold, the strategic approach taken by South Korean firms will be crucial. While the partnerships may yield immediate benefits in technology and infrastructure development, the long-term impact on autonomy in AI development remains uncertain. With the global AI landscape evolving rapidly, striking a balance between collaboration and independence will likely shape the future of the sector in South Korea and beyond.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
