Summary
NVIDIA says partners are planning capacity expansions across substrates, advanced packaging, semiconductor testing and optical engines. The targets include a 2.5× expansion in substrate capacity over three years and new production capacity scheduled from 2027.
NVIDIA says it is working with six manufacturing partners on capacity expansions for components and processes used in its AI computing systems. The plans cover semiconductor substrates, advanced packaging, chip testing and optical engines, with new production capacity targeted from 2027 and further expansion through 2028.
The company’s announcement, made on October 21, 2026, sets out a 2.5× expansion in substrate capacity over the next three years. It also includes plans for NVIDIA’s total semiconductor testing capacity to double from calendar 2026 to 2027. These are manufacturing targets and projections, rather than reports of capacity already brought online.
Capacity plans span four manufacturing stages
Substrates are layers that provide electrical connections between semiconductor dies and the rest of a package. NVIDIA says IBIDEN, Samsung Electro-Mechanics and Unimicron Technology Corporation will contribute to its substrate expansion.
At IBIDEN’s Ono Plant in Japan, an expansion supported by a multibillion-dollar capacity reservation advance from NVIDIA targets capacity reaching three times current levels. Additional capacity is expected to begin ramping in the first quarter of 2028. Samsung Electro-Mechanics is joining NVIDIA’s substrate supply chain with existing capacity in South Korea, while also upgrading facilities and building factories in Sejong and Busan. Unimicron’s KF substrate expansion in Taiwan is expected to start adding capacity in the second quarter of 2027 and reach full capacity by mid-2028.
For advanced packaging, NVIDIA and Amkor plan to increase interposer wafer capacity by more than 300% by the end of 2028. An interposer is a layer used to connect components within some chip packages. The planned expansion in South Korea, alongside new Amkor manufacturing facilities there, is intended to support higher volumes for NVIDIA’s current and next-generation platforms.
Testing is another planned growth area. NVIDIA expects its overall semiconductor testing capacity to double between calendar 2026 and 2027, supported by buildouts in the United States, Taiwan and Singapore. A $1.1 billion capacity reservation advance to King Yuan Electronics (KYEC) will support new US testing facilities and equipment, with high-volume production targeted for April 2027. Expanded testing operations in Singapore are expected to begin high-volume production in February 2027.
Optical engines target faster growth
NVIDIA and Siliconware Precision Industries (SPIL) plan to expand production in Taiwan of optical engines used in co-packaged optics. These components support the movement of data using optical signals in computing systems. The partners are targeting a 15-fold increase in optical-engine capacity by mid-2027, with another doubling aimed for the end of 2028.
The investment across these stages matters because AI computing systems depend on more than the processors themselves: substrates, packaging, testing and high-speed connections are all part of getting systems manufactured at scale. Expanding capacity at several points in that chain could help NVIDIA and its suppliers support higher production volumes and reduce reliance on a narrower set of manufacturing capabilities. The announcement describes the partners’ plans and intended capacity; whether and when each target is reached will depend on execution.