SK hynix (000660.KS), the world's largest maker of high-bandwidth memory (HBM), and Nvidia (NVDA) announced a multiyear technology partnership on June 7, 2026, to co-develop next-generation memory for Nvidia's AI "factory" platforms. The two companies framed the agreement as a long-term commitment to align memory development with Nvidia's product roadmap (SK hynix Newsroom; NVIDIA Newsroom).
For a portfolio manager, the first question is blunt: does this lock in or change SK hynix's memory allocation inside Nvidia's next-generation Rubin ramp, or is it a framework with no new money attached? Based on the disclosures, it is the latter. Neither company disclosed any financial terms, purchase volumes, pricing or capacity commitments (The Korea Herald; SK hynix Newsroom). What the deal does is formalize, in name, a supply relationship that already dominates Nvidia's memory pipeline.
What was actually agreed
Under the partnership, SK hynix will develop memory for Nvidia's next-generation systems — including Vera Rubin AI supercomputers, Vera CPUs, RTX Spark-powered PCs and Jetson Thor robotics platforms — extending beyond data-center accelerators into what Nvidia calls personal AI and physical AI (SK hynix Newsroom). Separately, the companies will apply Nvidia's own software to chipmaking: CUDA-X libraries and PhysicsNeMo for semiconductor simulation and TCAD workflows, and Nvidia Omniverse, OpenUSD and cuOpt to build factory digital twins aimed at autonomous fab operations (SK hynix Newsroom).
SK Group Chairman Chey Tae-won said the two companies "have been building toward this for years, and this partnership reflects the depth of that collaboration" (SK hynix Newsroom). Nvidia founder and CEO Jensen Huang called SK hynix "an extraordinary partner" that has played "a central role in delivering advanced memory technologies" (SK hynix Newsroom).
Sizing it against what SK hynix already has
The announcement matters less for what it adds than for what it confirms. SK hynix held roughly 62% of HBM shipments as of Q2 2025, the largest share of any supplier (Astute Group, citing market-share data). On the next-generation HBM4 that the Rubin platform will use, TrendForce reported in January 2026 that "Nvidia is expected to allocate roughly two-thirds of its HBM4 demand for the Vera Rubin platform to SK hynix, with the company's share understood to be close to 70%" — above earlier expectations of just over 50% (TrendForce). In other words, the commercial relationship the press release formalizes is one where SK hynix is already the default supplier; the agreement reduces roadmap and supply-cycle risk rather than redrawing market share.
That continuity has history. SK hynix was the first company to supply HBM3 DRAM to Nvidia, beginning mass production in 2022 and serving as Nvidia's effectively sole HBM3 supplier until Samsung began entering the supply chain in late 2023 (SK hynix Newsroom, 2022; TrendForce). The June 7 pact reads as the codification of a four-year head start.
What the market may be watching
Because no volumes or prices were disclosed, the partnership itself does not change near-term numbers. The data points that would confirm whether the codified relationship translates into incremental memory revenue are concrete: SK hynix's next quarterly earnings, the timing and yield of HBM4 mass production, and the launch cadence of Nvidia's Vera Rubin systems. Until those land, the agreement is best read as a strategic signal — a public lock-in of a relationship the supply data already showed — rather than a fresh financial catalyst.
This article is for informational purposes only and does not constitute investment advice. All figures are attributed to the cited sources as of publication; readers should verify current data before making any decisions.
