What's happening
Nvidia and SK Group announced on July 25, 2026, that they have signed letters of intent for a strategic relationship valued at more than $500 billion. The agreement spans multiple SK subsidiaries: a long-term memory supply deal with SK hynix, first unveiled in June 2026, and a commitment by SK Telecom to build a 2-gigawatt AI data center utilizing Nvidia's DSX platform and Vera Rubin systems, with SK hynix HBM4 memory serving as the underlying memory architecture. The first AI data center facility under the partnership is scheduled to enter service in 2027, with expansions of AI infrastructure and factories also commencing that year.
An SK Group executive framed the partnership in terms of national industrial ambition: "By leveraging SK hynix's AI memory and SK Telecom's AI infrastructure capabilities, SK will collaborate with NVIDIA to build a world-class AI factory, helping Korea transcend its role as a leading adopter of AI and become a global hub that drives AI innovation." The deal represents one of the largest publicly announced AI infrastructure commitments to date, linking Nvidia's hardware and software ecosystem directly to two of South Korea's largest technology conglomerates across the memory, telecommunications, and data center verticals.
Why it matters for markets
For Nvidia, the agreement locks in a long-term demand signal across two of the most capital-intensive segments of the AI buildout: high-bandwidth memory supply and large-scale data center deployment. Nvidia currently carries a market capitalization of approximately $5.01 trillion and reported revenue of $253.49 billion, reflecting its already dominant position in AI accelerator hardware. A partnership of this stated scale — exceeding $500 billion in combined commitment — extends that revenue runway further into the next hardware generation, specifically through the Vera Rubin platform and HBM4 memory integration, both of which represent next-generation product cycles rather than existing deployed infrastructure.
The 2-gigawatt data center commitment from SK Telecom is notable in terms of infrastructure scale. Data center power capacity at that level represents a substantial physical buildout, and the exclusive use of Nvidia's DSX platform and Vera Rubin systems within that facility deepens the hardware lock-in that has characterized Nvidia's competitive positioning. The HBM4 memory specification also signals that the partnership is oriented toward future product generations, not current inventory, which has implications for the timeline of revenue recognition — with the 2027 service entry date suggesting financial impact will materialize over a multi-year horizon rather than in the near term.
For the broader AI infrastructure supply chain, the deal underscores the continued concentration of large-scale AI capital expenditure around Nvidia's hardware ecosystem. The pairing of SK hynix's memory roadmap with Nvidia's accelerator platforms formalizes a supply-chain relationship that had been developing since at least June 2026, when the memory supply agreement was first unveiled, and now extends it into data center construction and factory-level AI deployment.
Sectors and assets to watch
The primary ticker directly implicated is NVDA (Nvidia Corporation), whose DSX platform and Vera Rubin systems are specified as the hardware foundation for the SK Telecom data center. Nvidia's product portfolio — including its AI accelerators and the CUDA software platform — sits at the center of the partnership's technical architecture. SK hynix, the memory subsidiary of SK Group, is the designated HBM4 supplier under the long-term agreement; as a Korea-listed entity (KRX: 000660), it is not a U.S.-traded primary subject of this story but is a direct operational counterparty to Nvidia in the deal's memory supply component. SK Telecom (NYSE: SKM) is the data center construction and operations counterparty.
More broadly, the agreement is relevant to the AI data center infrastructure sector, including companies involved in power delivery, cooling, and networking for hyperscale facilities, given the 2-gigawatt capacity specification of the planned SK Telecom facility. The high-bandwidth memory segment — where SK hynix competes with Samsung and Micron Technology (MU) — also warrants monitoring, as the formal designation of HBM4 as the memory standard for this deployment may influence procurement decisions and competitive dynamics within that market.
What to watch next
Key developments to monitor include the formal conversion of the letters of intent into binding contracts, the specific timeline and phasing of the SK Telecom data center groundbreaking and construction milestones ahead of the 2027 service entry target, and any further specification of the AI factory expansion plans that are also slated to begin in 2027. Investors and analysts will also be watching for disclosure of how the $500 billion valuation is structured across the partnership's components — whether it reflects hardware procurement commitments, infrastructure investment, or a combination — as that breakdown will determine how and when the financial impact flows through Nvidia's revenue recognition. Updates to SK hynix's HBM4 production ramp and any amendments to the June 2026 memory supply agreement terms will also be relevant to assessing the partnership's execution trajectory.