Nvidia Invests Over $1 Billion in Kinsus to Expand ABF Substrate Output
Nvidia will invest more than $1 billion in cash in Taiwanese substrate maker Kinsus, funding equipment purchases to expand its Taoyuan Youshi plant and lift advanced ABF substrate capacity by over 3.5x, with new output expected in H2 2027 and full ramp by Q3 2028. Nvidia also amended its substrate supply agreement with Samsung Electro-Mechanics to raise existing plant throughput and reaffirmed its next-gen GPU and networking partnership with Foxconn.
Related Entities
Full Report
IT之家 November 18 news, NVIDIA announced yesterday local time that on the basis of the cooperation announced in October this year, it will further deepen its relationship with supply chain partners to support artificial intelligence infrastructure construction.

According to the latest announcement, NVIDIA will invest over $1 billion in cash (IT之家 note: approximately 6.68 billion RMB at current exchange rates) in Taiwan, China semiconductor packaging substrate manufacturer Kinsus, purchasing equipment to support the expansion plan of Kinsus's Taoyuan YouShi plant, with the goal of expanding the plant's advanced ABF packaging substrate capacity by more than 3.5 times.
The new capacity at Kinsus's YouShi plant is expected to come online in the second half of 2027 and reach full production in Q3 2028. If capacity is still insufficient, NVIDIA plans to continue participating in Kinsus's subsequent capacity expansion plans, with the goal of securing stable capacity at least before 2030.
NVIDIA also amended its existing substrate supply agreement with Samsung Electro-Mechanics (SEM) to improve the operational capacity of existing plants, thereby supporting increased production of NVIDIA's accelerated computing platforms.
In addition, Hon Hai Chairman Young Liu stated: "Hon Hai will continue to deepen cooperation with NVIDIA in next-generation GPUs and networking platforms. With its globally integrated production network covering Asia, North America, and Europe, Hon Hai is fully capable of supporting the mass production of next-generation artificial intelligence infrastructure."