Full Breakdown
NVIDIA and SK Hynix Sign Multi-Year AI Memory Pact
6/8/2026, 9:15:27 PM
Partnership Overview
On June 8, 2026, NVIDIA and SK Hynix announced a multiyear partnership to co-develop memory for NVIDIA’s platforms. The deal covers HBM4 and DRAM for Vera Rubin supercomputers, Vera CPUs, RTX Spark PCs, and Jetson Thor robotics, and embeds NVIDIA software tools into SK Hynix’s chip design and fab processes.
Background & Context
AI model training and inference now demand terabytes of high-bandwidth memory per server, creating a bottleneck for data-centers. NVIDIA’s Vera platforms require HBM4, but supply constraints have prompted the company to lock in long-term capacity with its largest memory supplier.
Key Figures & Groups
The partnership is led by Jensen Huang, founder and CEO of NVIDIA, and Chey Tae-won, chairman of SK Group, who oversees SK Hynix. SK Hynix CEO Kwak Noh-Jung heads the semiconductor design and fab teams that will integrate NVIDIA’s CUDA-X, PhysicsNeMo, and Omniverse tools.
Strategic Implications
The alliance aims to keep memory supply ahead of AI demand, shorten chip-design cycles via AI-accelerated simulation, and enable autonomous fab operations through digital twins. SK Hynix, a US$1.12 trillion-valued firm, joins Samsung and Micron as an HBM4 supplier, with Vera servers targeting up to 1.2 TB/s bandwidth. The deal deepens South Korea’s role in the global AI supply chain and may reshape competition with Samsung and Micron.
Official Statements & Responses
NVIDIA said the partnership will “co-develop the next generation of memory for AI factories” and align SK Hynix capacity with its AI roadmap, supporting frontier AI. SK Hynix highlighted diversification into AI infrastructure and physical AI markets while applying AI to semiconductor design and fab automation.
Criticism & Opposition
Analysts note the lack of disclosed financial terms or volume commitments, leaving supply certainty unclear. Prioritizing AI-factory memory may heighten price pressure on DRAM and DDR5. The announcement coincided with a sell-off in Korean tech stocks.
Conflicting Reports & Gaps
Sources differ on whether SK Hynix receives exclusive rights; NVIDIA still sources HBM4 from Samsung and Micron. No detailed timeline beyond Q3 2026 or exact deal length is provided.
Verbatim Quotes
- “AI factories are the engines of the next industrial revolution, and advanced memory is essential to their performance.” — Jensen Huang, NVIDIA CEO
- “SK Hynix has been Nvidia’s largest memory partner. SK Hynix will continue to be Nvidia’s largest memory partner,” — Jensen Huang, NVIDIA CEO
- “BANDF AD SK Group chairman Chey Tae-Won said: “SK Hynix and Nvidia have been building toward this for years, and this partnership reflects the depth of that collaboration.” — Chey Tae-won, Chairman of SK Group
- “We already procure and we buy from SK Hynix already billions and billions of US dollars each year, and it’s going to grow substantially,” — Jensen Huang, NVIDIA CEO
What’s Next
NVIDIA and SK Hynix will extend joint R&D on Vera platforms, expand digital-twin fab tools, and support SK Telecom’s gigawatt AI cloud slated for 2027.
