'We will do our best to help Nvidia succeed'

Jeon Young-hyun, Samsung Electronics co-CEO and head of the Device Solutions division, walks to a meeting with Nvidia CEO Jensen Huang at the Shilla Hotel in Jung-gu, Seoul, on Monday. [Yonhap]
Jeon Young-hyun, Samsung Electronics co-CEO and head of the Device Solutions division, walks to a meeting with Nvidia CEO Jensen Huang at the Shilla Hotel in Jung-gu, Seoul, on Monday. [Yonhap]

Jeon Young-hyun, Samsung Electronics co-CEO and head of the Device Solutions division, met with Nvidia CEO Jensen Huang on Monday to discuss supply of HBM4E and HBM5 — the seventh- and eighth-generation high-bandwidth memory chips.

Speaking to reporters after the meeting at the Shilla Hotel in Jung-gu, Seoul, Jeon said the talks with Huang and senior Nvidia executives went exceptionally well. "I have been working with CEO Huang for a long time, but today felt like our best conversation yet," he said.

Jeon said the two sides covered a broad range of future cooperation. "We talked extensively about long-term collaboration starting next year — including HBM4E, foundry business and HBM5," he said, adding that Samsung must ensure sufficient supply of HBM4 and SOCAMM, a low-power memory module for servers, beginning this year.

Nvidia's delegation included Madison Huang, Jensen Huang's eldest daughter and a senior director at the company, as well as Jeff Fisher, senior vice president. Samsung was represented by Jeon and Kim Jae-jun, vice president of the memory business division.

Asked about expanding foundry cooperation, Jeon said Samsung and Nvidia are already collaborating on autonomous driving chips and Nvidia's Groq accelerator chip at the 4-nanometer and 8-nanometer nodes. "We are also discussing collaboration on the next generation," he said.

Samsung is supplying HBM4 — its sixth-generation product delivering speeds exceeding 11.7 gigabits per second, among the highest in the industry — for Nvidia's next-generation AI accelerator Vera Rubin, which has entered full mass production. Samsung is also providing LPDDR5X-based SOCAMM2 for the Vera CPU and PCIe Gen6-based PM1763 storage solutions.

Samsung recently completed the world's first sample shipment of HBM4E, its seventh-generation product, and has delivered samples to Nvidia.

Asked about Huang's earlier remark that SK Hynix "has been and will continue to be Nvidia's largest memory partner," Jeon said Samsung would let its results speak for themselves. "We will work hard at what we do and show you through our results later," he said.

On whether Samsung plans to sign a long-term memory supply agreement with Nvidia, Jeon said the company's focus is on being the best partner it can be. "We will do our utmost to help Nvidia succeed," he said.


20ki@heraldcorp.com