Business·Industry

NVIDIA-Driven AI Boom Extends Beyond Memory to Circuit Boards… Samsung and SK Get a Boost

Vera Rubin Shipments in Full Swing… HBM4 Demand Expected to Grow Samsung and SK Benefit from Memory Boom, Including Server DRAM SamsungElectroMechanics Considers Supporting Additional Investment in NVIDIA Collaboration Extends Beyond Circuit Boards to MLCCs… Accelerating Expansion of AI Supply Chains

JAEMIN SONG
2026-11-18 17:15:36
[Edaily Reporter JAEMIN SONG ] As NVIDIA, a leader in the artificial intelligence (AI) semiconductor market, begins full-scale shipments of its next-generation AI platform “Vera Rubin,” expectations are growing that the domestic semiconductor and components industries will benefit. SamsungElectronics and SK hynix are expected to continue their boom, driven by expanding demand for High-Bandwidth Memory (HBM). SamsungElectroMechanics is also expanding its scope of cooperation with NVIDIA to include multilayer ceramic capacitors (MLCCs), following its collaboration on advanced packaging substrates.

NVIDIA logo sign (Photo: EPA·Yonhap News)

According to industry sources on the 18th, NVIDIA’s revenue for the third quarter of fiscal year 2027 reached $115.8 billion, a 103% increase from the same period last year. Data center revenue rose 110% to $107.6 billion. The company also projected fourth-quarter revenue of $130 billion, exceeding market expectations.

Vera Rubin Shipments in Full Swing… Samsung and SK Memory Boom Continues

In this earnings announcement, NVIDIA officially confirmed the expansion of mass production and the start of shipments for the “Vera Rubin NVL72.” As the transition to next-generation products gains momentum, the expansion of HBM4 supply by domestic memory manufacturers is also expected to accelerate.

Vera Rubin is equipped with HBM4, the sixth-generation HBM. A single NVL72 unit, consisting of 72 GPUs, contains approximately 20.7 terabytes (TB) of HBM capacity. While the capacity is the same as that of the previous generation, Blackwell Ultra, the memory bandwidth has increased by up to 2.8 times, making a stable supply of high-performance HBM4 critical. SK hynix is working to maintain its market leadership based on its existing supply infrastructure, while SamsungElectronics is focusing on securing market share by expanding its HBM4 supply.

As shipments of NVIDIA’s next-generation products increase, HBM4 supply volumes are expected to rise accordingly. Additionally, as the volume of data processed by AI servers grows, demand for server DRAM and enterprise solid-state drives (SSDs) is also expanding. With advanced DRAM production capacity concentrated on HBM production, the supply of standard DRAM is becoming tight, which is expected to contribute to improved profitability for SamsungElectronics and SK hynix.

Supply Chain Expands to Include Circuit Boards and Packaging… Strengthened Collaboration with SamsungElectroMechanics

NVIDIA’s production expansion is also driving demand for key components such as advanced printed circuit boards and MLCCs. This is because as the performance and power consumption of AI semiconductors increase, so does the importance of the precision of the printed circuit boards that connect the chips and the MLCCs that ensure a stable power supply.

SamsungElectroMechanics announced today that it is strengthening its collaboration with NVIDIA to supply advanced packaging substrates for next-generation graphics processing units (GPUs). Building on the strategic partnership announced last October, the company plans to increase production capacity at its Sejong and Busan facilities and expand the supply of high-performance substrates for data centers.

In addition to its existing support for capacity expansion, NVIDIA is also considering providing additional funding to expand production infrastructure in Sejong and Busan. Based on this, SamsungElectroMechanics plans to expand its advanced facilities and establish an additional mass production system starting in 2029. NVIDIA’s support for investment in production facilities is a measure aimed at securing a stable supply of key components necessary for increasing the production of next-generation AI semiconductors.

The scope of the partnership is also expanding from printed circuit boards to MLCCs. According to SamsungElectroMechanics, the required capacitance of MLCCs per GPU and ASIC for AI servers is trending toward doubling every year. As power consumption by high-performance AI semiconductors increases, demand is also rising for high-capacity, high-reliability MLCCs that maintain stable voltage. NVIDIA is also working with Taiwan’s Kinseus to expand printed circuit board production capacity.

Deborah Schouquist, Senior Vice President of Operations at NVIDIA, stated, “Close collaboration across the entire global supply chain is essential to meet the long-term demand for AI infrastructure.”

Economy

Corporation

IT·Science

Economy

HyundaiMotor Group Completes Restructuring Focused on Commercial Robots… Atlas Welcomes Its First Semiconductor Customer

Boston Dynamics’ humanoid robot Atlas (right) and quadruped robot Spot (left). (Photo courtesy of HyundaiMotor Group) HyundaiMotor Group has completed a business restructuring centered on commerci…
2026-11-18 17:17:41

Corporation

Ensol Bio Sets Its Sights on a "One-Shot" KOSDAQ Technology Evaluation in the First Half of Next Year… "P2K, H1K, E1K" Trio Takes Center Stage

Ensol Biosciences, an innovative new drug developer listed on KONEX, has officially announced a “one-shot” strategy to pass the technology evaluation—a key hurdle for a technology-based special listin…
2026-11-18 17:10:02

IT·Science

NHN Officially Launches New Puzzle Game ‘My Favorite Child: Puzzle Star’ in Over 170 Countries Worldwide

NHN(181710)announced on the 18th that it has officially launched the match-3 puzzle game “My Favorite Child: Puzzle Star” in over 170 countries worldwide. (Photo courtesy of NHN) “My Favorite Ch…
2026-11-18 14:57:27