NVIDIA SK hynix Partnership Expands AI Memory Supply
NVIDIA SK hynix partnership to co-develop AI memory and expand supply for AI factories, easing HBM supply risk and shifting trader flows.
KEY TAKEAWAYS
- NVIDIA and SK hynix had announced a multiyear partnership to co-develop next-generation memory and expand supply.
- SK Telecom and Naver had announced DSX deployments and gigawatt-scale AI factory plans in South Korea.
- Jensen Huang had said memory shortages would persist for several years, framing constraints as structural.
HIGH POTENTIAL TRADES SENT DIRECTLY TO YOUR INBOX
Add your email to receive our free daily newsletter. No spam, unsubscribe anytime.
NVIDIA Corporation (NVDA) and SK hynix announced on June 7, 2026, a multiyear technology partnership to co-develop next-generation memory and expand supply for NVIDIA’s AI-factory roadmap. Alongside this, NVIDIA struck deals with SK Telecom and Naver to deploy its DSX AI factory architecture and build gigawatt-scale AI capacity in South Korea.
Korea AI Infrastructure Deals
NVIDIA said the partnership with SK hynix aims to advance next-generation memory for “AI factories,” its term for full-stack AI infrastructure that trains, deploys, and operates models at scale. The collaboration focuses on co-developing memory aligned with NVIDIA’s AI infrastructure roadmap and expanding supply to meet the accelerating global buildout of AI data centers.
SK Telecom will add NVIDIA-powered AI cloud capacity based on the DSX AI factory architecture to accelerate AI startups, robotics, and industrial physical AI applications. This local cloud infrastructure supports model development and edge-to-cloud robotics use cases, positioning Korea as a hub for AI innovation.
Naver plans to use NVIDIA technology to build gigawatt-scale AI factories to meet rising global demand for AI services and physical AI. Gigawatt-scale refers to power draws around one gigawatt per cluster, placing these deployments among the largest AI compute facilities worldwide and highlighting the scale these partnerships target.
Memory Supply Outlook and Strategic Impact
NVIDIA CEO Jensen Huang, speaking in Seoul, warned that supply chains across wafers, packaging, and silicon photonics remain strained and that these constraints will persist for several years. He described the capacity gap as structural, underscoring the need for coordinated efforts.
NVIDIA positions its DSX platform as a comprehensive playbook for building and operating AI factories. DSX combines modular software libraries, application programming interfaces (APIs), reference designs, and accelerated compute platforms to reduce token cost and shorten time to production. The company has validated an enterprise AI-factory design around its Blackwell GPUs, BlueField data processing units (DPUs), Spectrum-X networking, and NVIDIA AI Enterprise software, packaging these as a repeatable blueprint for large-scale deployments.
The SK hynix partnership explicitly aims to expand memory supply for the global AI factory buildout, addressing capacity constraints in high-bandwidth memory (HBM) for GPUs and AI accelerators. The SK Telecom and Naver agreements focus on deploying DSX-based cloud and data-center scale infrastructure closer to users, shaping local supply and deployment pathways.
The reviewed releases do not disclose financial terms, unit volume commitments, or contract values tied to these partnerships. The arrangements are structured as technology and infrastructure collaborations rather than mergers or equity transactions, with no regulatory filings referenced.
"Expands supply for the accelerating global AI factory buildout," the NVIDIA press release stated.
