Jensen Huang Plans to Meet Executives from Samsung, SK Hynix, and Naver
Alina Collins
Nvidia CEO Jensen Huang is preparing a roundtable with Samsung's Jay Y. Lee, SK Group's Chey Tae-won and Naver's Lee Hae-jin, tentatively set for July 24 near Silicon Valley — the second meeting in just over a month, with HBM4 supply terms and AI infrastructure investment in focus ahead of Vera Rubin shipments.
Is this meeting actually confirmed?
Korean media — Yonhap, ETNews and Newsis — report a roundtable tentatively on July 24 near Silicon Valley, but no company has formally announced the time, venue or agenda.
Hyundai Motor's Chung Euisun, OpenAI's Sam Altman, Anthropic's Dario Amodei and Microsoft executives have been mentioned as possible attendees, but none have confirmed.
This means → every agenda detail remains media speculation; investors should wait for official statements.
Why meet twice in a single month?
In June, Huang visited South Korea and met executives from Samsung, SK, Hyundai, LG and Naver, covering HBM4 delivery timelines, SOCAMM memory — a new memory module designed for AI servers — foundry partnerships and next-generation HBM co-development.
Nvidia's Vera Rubin platform — its next-generation AI chip system — is in full production ramp, with commercial shipments planned for this autumn.
This means → the shipping window is closing in; every link in the supply chain needs to lock down HBM4 terms and capacity allocation before mass production. The meeting cadence itself is a signal.
What has each side already committed to on AI infrastructure?
SK Telecom is working with Nvidia to build a gigawatt-scale AI cloud in South Korea using Nvidia's DSX platform, with the first AI factory expected to be operational by 2027.
Naver has announced a gigawatt-scale AI factory plan with Nvidia: 55 MW in the first half of 2027, scaling to 100 MW later that year, 200 MW in 2028, with an ultimate target of 1 GW.
In plain terms = these two Korean companies are not just buying Nvidia chips — they are co-building entire AI data centres with Nvidia, measured in power-plant-scale megawatts.
How deep are Samsung and SK Hynix in the supply chain?
Samsung and SK Hynix both supply high-bandwidth memory (HBM) to Nvidia. Their HBM4 products are expected to go into the Vera Rubin system.
Both companies are also part of OpenAI's Stargate initiative, expanding advanced memory capacity for OpenAI's AI infrastructure.
On the foundry side, Samsung has confirmed it will manufacture a new language processing unit (LPU — a chip purpose-built for large-language-model inference) for Groq; Nvidia lists the Groq 3 LPX as part of the Vera Rubin platform. Anthropic is also reportedly in talks to use Samsung's 2 nm process for AI chips, though neither side has confirmed.
What should the market watch for?
If the meeting goes ahead, it lands just before Vera Rubin's mass commercial shipments.
The key question: whether the parties can reach clearer consensus on HBM4 supply terms and the scale of AI infrastructure investment.
This reflects Nvidia's evolution from "selling chips" to "locking in an entire AI infrastructure supply chain" — with Korean companies deeply embedded across memory, foundry and data-centre buildout simultaneously.
Content is for reference only, not financial advice.