Samsung Electro-Mechanics and NVIDIA have entered into a strategic business partnership aimed at strengthening the global AI semiconductor supply chain and accelerating the expansion of next-generation AI infrastructure. Under this agreement, the two companies will jointly invest in expanding dedicated manufacturing facilities for Ajinomoto Build-up Film (ABF) package substrates, which are critical components used in high-performance AI semiconductors.

Leveraging this expanded production capacity, Samsung Electro-Mechanics plans to supply NVIDIA with advanced ABF package substrates beginning in 2026. The partnership represents a proactive move by NVIDIA to secure a stable supply of advanced package substrates essential for the mass production of its AI chipsets, directly addressing and helping to alleviate supply chain bottlenecks in the rapidly expanding global AI infrastructure market. For Samsung Electro-Mechanics, the agreement offers a significant opportunity to accelerate its expansion into the high-value-added AI semiconductor package substrate market, a sector characterized by exceptionally high technological barriers to entry, while strengthening its competitive position globally.

Dedicated Production Facilities in Sejong and Busan

To ensure the timely and reliable delivery of these critical components, Samsung Electro-Mechanics will establish dedicated production lines at its key manufacturing sites in Sejong and Busan, South Korea. This localized manufacturing strategy is designed to support a multi-year supply agreement covering advanced package substrates for both NVIDIA's compute chipsets and high-speed networking switches. These networking switches play a vital role in enabling ultra-fast chip-to-chip connectivity within modern data centers.

The supply agreement, which commences in 2026, outlines a clear growth trajectory for the partnership:

  • Starting in 2026, the volume of package substrates supplied by Samsung Electro-Mechanics to NVIDIA is expected to increase at a double-digit annual rate.
  • By 2028, the production capacity allocated to NVIDIA is projected to account for a significant portion of Samsung Electro-Mechanics' total annual package substrate production capacity.

This structured ramp-up highlights the long-term nature of the collaboration and the scale of manufacturing commitment required to support next-generation AI platforms.

Advanced Packaging Technologies and Technical Challenges

As AI processors demand higher performance, package substrates must become larger and incorporate more complex technologies, which significantly increases manufacturing complexity. To maintain high manufacturing yields and ensure long-term reliability, several advanced engineering hurdles must be overcome:

  • Ultra-fine circuit patterning is essential to minimize signal loss during high-speed data transmission.
  • Advanced warpage control and substrate flatness technologies are critical to preventing defects during assembly.
  • Next-generation AI chips require sophisticated power management solutions to suppress electrical noise generated during ultra-high-speed computing and to maximize overall power efficiency.

To meet these demanding specifications, advanced packaging processes embed passive components, including multilayer ceramic capacitors (MLCCs) and silicon capacitors, directly within the substrate structure itself. Samsung Electro-Mechanics is uniquely positioned to address these challenges by combining its advanced substrate microfabrication capabilities with its world-class expertise in MLCC and silicon capacitor technologies. By integrating these core capabilities, the company aims to deliver high-performance, high-reliability package substrate solutions optimized for the evolving requirements of AI computing infrastructure.

Executive Leadership Perspectives

The leadership of both organizations emphasized the strategic importance of this collaboration for the future of AI infrastructure.

Duck-Hyun Chang, President and CEO of Samsung Electro-Mechanics, expressed the company's commitment to the partnership:

"It is a great honor to join NVIDIA, a global leader driving AI innovation, as a key infrastructure partner. By combining our advanced FC-BGA manufacturing capabilities with our differentiated expertise in high-reliability MLCC technologies, Samsung Electro-Mechanics will fully leverage its strengths as a comprehensive solutions provider to ensure that NVIDIA's AI platforms are delivered to the market on time with the highest levels of performance and quality."

Debora Shoquist, Executive Vice President of Operations at NVIDIA, highlighted the necessity of expanding the global supply chain:

"The pace of the AI infrastructure buildout requires us to expand manufacturing capacity on a global scale. Our work with Samsung Electro-Mechanics broadens our network of substrate partners and adds manufacturing capacity to help us scale production and deliver the accelerated computing platforms our customers need."

Through this joint investment and long-term supply agreement, both companies are positioning themselves to meet the intense demands of the global AI buildout, establishing a more resilient supply chain for the advanced processors that power modern data centers.