NVIDIA has expanded its supply-chain commitments to align manufacturing capacity with growing demand for accelerated computing and AIartificial intelligenceAI stands for artificial intelligence infrastructure platforms. The company is adding substrate capacity, increasing advanced packaging output and deepening collaboration with contract manufacturers and foundries across its current and next-generation product lines.
On the substrate side, NVIDIA is extending its relationship with Kinsus and Samsung Electro-Mechanics. Kinsus plans to raise dedicated advanced ABFAjinomoto build-up filmAn ABF is a build-up film used for high-density circuit structures in substrates. package substrate capacity for NVIDIA by more than 3.5x from its Youth campus in Taoyuan, Taiwan. Production is expected to start in the second half of 2027, with the additional capacity targeted for the third quarter of 2028.
Samsung Electro-Mechanics has amended its existing substrate supply agreement to improve operational capabilities at its factories in support of higher production for NVIDIA’s accelerated computing platforms. The collaboration continues to cover substrate manufacturing in Sejong and Busan, South Korea, alongside compute and high-speed networking technologies.
In advanced packaging, NVIDIA is expanding capacity with Amkor and SPIL. The agreements with Amkor support manufacturing expansion in Arizona and South Korea and total nearly US$3 billion (approximately €2.7 billion). They include prepayments intended to help Amkor scale capacity for NVIDIA’s long-term production needs. SPIL is building out advanced packaging operations in Taiwan under a multiyear agreement supported by multibillion-dollar funding from NVIDIA, with dedicated capacity expected to increase by mid-2029.
The SPIL project adds new production lines with turnkey support for bumping, interposer wafer fabrication, assembly and testing. These steps are aimed at semiconductor products with demanding requirements for performance, bandwidth, power efficiency and system integration.
NVIDIA also said the third quarter marked the first volume shipment of Vera Rubin NVL72 and Vera CPUcentral processing unitMain processor that executes program instructions and performs data operations.s for agentic AIartificial intelligenceAI stands for artificial intelligence. Those products use TSMCTaiwan Semiconductor Manufacturing CompanyTSMC is named as a founding partner of the EPIC Center.’s 5 nm, 3 nm and CoWoS advanced packaging technologies. Foxconn is also deepening its collaboration with NVIDIA to support next-generation GPUgraphics processing unitA GPU is a parallel graphics processor widely used for AI acceleration. and networking platforms through its global production network.
Further manufacturing commitments announced earlier this year included additional substrate, testing and optical-engine expansion across partners such as IBIDEN, King Yuan Electronics, Samsung Electro-Mechanics, SPIL and Unimicron. Taken together, the agreements point to a broad capacity buildout across the supply chain as NVIDIA prepares for higher-volume AIartificial intelligenceAI stands for artificial intelligence infrastructure production.