NVIDIA adds Kinsus deal targeting more than 3.5x substrate capacity
NVIDIA's new Kinsus agreement targets more than 3.5 times Kinsus's current dedicated ABF substrate capacity, with production expected in 2027 and added capacity targeted for 2028.
NVIDIA announced a new multiyear agreement with Kinsus Interconnect Technology targeting more than 3.5 times Kinsus’s current advanced ABF package-substrate capacity dedicated to the chipmaker, the company said.
Kinsus expects production at its Youth campus in Taoyuan, Taiwan, to begin in the second half of 2027. The additional capacity is targeted for the third quarter of 2028. Both dates and the capacity multiple are forward-looking company targets, not delivered output. NVIDIA did not disclose the current capacity in units, so the announcement does not support an absolute-volume estimate.
NVIDIA said the additional substrate capacity would support production of its current and next-generation accelerated-computing platforms.
The Kinsus agreement is separate from NVIDIA’s broader supplier plan announced on October 21. That earlier plan named Amkor, IBIDEN, KYEC, Samsung Electro-Mechanics, SPIL and Unimicron. It targeted an aggregate 2.5-times increase in substrate capacity over three years and did not name Kinsus.
Kinsus’s official site also lists a November 2026 announcement describing more than $1 billion in investment from NVIDIA to expand advanced substrate capacity for AI. The accessible listing does not explain the investment structure or payment terms, while NVIDIA’s release does not state a value for the Kinsus agreement.
NVIDIA also said it amended its existing substrate supply agreement with Samsung Electro-Mechanics to improve operations at existing factories and support increased production. Samsung’s original October partnership announcement said dedicated lines in Sejong and Busan, South Korea, would supply NVIDIA beginning in 2026. The November update did not disclose a new capacity multiple or value for the amendment.
In advanced packaging, NVIDIA said a new multiyear agreement with SPIL is backed by multibillion-dollar NVIDIA funding and is intended to increase contracted capacity dedicated to NVIDIA by mid-2029. The company did not quantify the planned increase. NVIDIA also restated previously announced Amkor agreements in Arizona and South Korea totaling nearly $3 billion. NVIDIA classifies expectations about the supplier arrangements, capacity increases and production timing as forward-looking statements subject to risks, not guarantees of future performance.
More news

AWS adds managed NVIDIA Dynamo support to SageMaker HyperPod Inference

NVIDIA unveils Nemotron science stack and Nsight porting agent

NVIDIA Introduces Four-Node DGX Spark HPC Developer Kit
