Nvidia Partners with SK Hynix, Naver, Doosan to Boost South Korea's AI Infrastructure
Nvidia is partnering with South Korean companies SK hynix, Naver, and Doosan Group to expand the country's AI infrastructure. This includes advancing next-generation memory chips for AI data centers, building new gigawatt-scale AI clouds, and utilizing physical AI technology for robotics. The collaborations aim to ensure a stable supply of advanced memory and accelerate the expansion of global AI infrastructure.
Key Takeaways
- SK hynix entered a multiyear technology partnership with Nvidia to advance next-generation memory chips crucial for AI data centers and ensure stable supply.
- SK Telecom, a sister company of SK hynix, will build a gigawatt-scale AI cloud in South Korea powered by Nvidia's AI chips, with its first data center slated for early next year.
- Nvidia is cooperating with Naver to expand capacity at its Gak Sejong data center and plans future gigawatt-scale AI factory builds, contingent on capacity procurement and power availability.
- Doosan Group will utilize Nvidia's physical AI technology for its robotics development and integrate its energy solutions into Nvidia's data center platforms.
Why It Matters
These partnerships deepen Nvidia's integration into South Korea's advanced technology ecosystem, a critical hub for memory chip production and other high-tech manufacturing. The collaborations address key bottlenecks in AI infrastructure development, specifically the demand for advanced memory and scalable AI cloud capacity. Watch for the operational dates of SK Telecom's and Naver's new AI data centers, which will indicate the pace of this expanded infrastructure rollout.
Additional Context
Nvidia's agreements in South Korea are part of a broader strategy to establish a comprehensive AI factory ecosystem within the country. According to Stockwirex (June 2026), Nvidia CEO Jensen Huang explicitly described the partnerships as anchoring Nvidia's AI infrastructure supply chain across memory, cloud, energy, and data center layers. The deals, which include SK Telecom beyond the primary article's mention, were reportedly structured as a coordinated package during Huang's June 2026 visit to Seoul. Stockwirex further reports that SK hynix, already providing over 80% of its HBM revenue to Nvidia in 2024, extended its relationship into HBM4 and CXL memory, critical for Nvidia's post-Blackwell accelerator platforms with mass production targeting 2026-2027. Naver's collaboration involves deploying Blackwell GPUs and Spectrum-X networking at its Sejong data center, aiming for mid-to-late 2026 operation, while SK Telecom's gigawatt-scale AI cloud facility is targeting operational status in 2027 (Yahoo Finance, June 2026). Doosan's role extends to materials supply for chip manufacturing, energy infrastructure with micro-grids and fuel cells for data centers, and physical AI applications (Stockwirex, June 2026). This strategic clustering of partnerships in South Korea aligns with US export controls, which have redirected AI compute investment towards US-aligned jurisdictions, positioning South Korea as a critical node in a "trusted semiconductor supply chain" due to its dominant HBM production and robust infrastructure (Stockwirex, June 2026).
Read full article at siliconangle.com