Nvidia and SK Group announce $500B AI infrastructure deal
Partnership spans massive data center buildout and next-generation memory development, with first facility targeting 2027 launch.
Nvidia and SK Group forge massive AI infrastructure partnership
Nvidia and South Korea's SK Group have announced a partnership exceeding $500 billion focused on AI data center infrastructure and advanced memory technology, according to details first reported by Reuters.
The initiative centers on a long-term collaboration with SK Hynix to secure next-generation memory supply for Nvidia's AI systems. The companies will jointly develop high-bandwidth memory (HBM) specifically designed for AI training workloads, AI agents, and physical AI applications.
Major data center deployment planned
As part of the broader agreement, SK Telecom plans to construct a 2-gigawatt AI data center powered by Nvidia's Vera Rubin chips and SK Hynix's HBM4 high-bandwidth memory. The first facility in this deployment is scheduled to come online in 2027.
The scale represents one of the largest announced AI infrastructure commitments to date, reflecting the massive capital requirements for training and deploying advanced AI systems at scale.
Expansion of existing South Korean AI infrastructure
Separately, Nvidia announced it will partner with Naver and Brookfield to expand Naver's existing AI data center operations in South Korea. This expansion builds on the country's growing position as a hub for AI infrastructure investment.
Why it matters
This partnership addresses two critical bottlenecks in AI development: compute infrastructure and specialized memory. High-bandwidth memory has become a key constraint for AI systems, with demand far outstripping supply. By securing dedicated HBM supply from SK Hynix while simultaneously building massive data centers, Nvidia is vertically integrating critical components of the AI stack. The $500+ billion commitment also signals that leading technology companies view AI infrastructure as requiring investments at a scale previously associated with national infrastructure projects. For South Korea, the deal strengthens its position in both memory manufacturing and AI services, potentially creating a regional competitive advantage as AI workloads continue to grow globally.
Memory supply chain implications
The partnership with SK Hynix represents a strategic move to secure access to advanced memory technology. HBM4, the next generation of high-bandwidth memory, will be critical for powering increasingly complex AI models that require rapid data transfer between processors and memory.
SK Hynix is one of only three major manufacturers of HBM globally, making supply agreements strategically important for AI chip makers like Nvidia.
The details were first reported by Reuters, with reporting by Stephen Nellis in San Francisco and Heekyong Yang and Jack Kim in Seoul.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call
