NVIDIA expands advanced packaging capacity
New manufacturing agreements target higher capacity for advanced packaging, testing and optical engines to support growing AI demand.
NVIDIA is expanding manufacturing capacity across its semiconductor supply chain through agreements with packaging, testing, substrate and optical-engine suppliers, underscoring the pressure to scale production for next-generation AI computing systems.
Announced on 21 October, the plans involve Amkor Technology, Siliconware Precision Industries (SPIL), King Yuan Electronics (KYEC), IBIDEN, Samsung Electro-Mechanics and Unimicron Technology.
The agreements target capacity constraints across several stages of semiconductor manufacturing, including advanced packaging, testing and co-packaged optics.
Amkor plans to increase its advanced packaging interposer wafer capacity in South Korea by more than 300% by the end of 2028.
The expansion builds on its existing collaboration with NVIDIA and is intended to support higher production volumes for current and next-generation AI platforms.
NVIDIA is also supporting KYEC’s planned US testing facilities through a US$1.1 billion capacity reservation advance.
The company expects its total semiconductor testing capacity to double between 2026 and 2027, supported by expansion across the United States, Taiwan and Singapore.
Elsewhere, NVIDIA expects agreements with substrate suppliers to contribute to a 2.5-fold increase in substrate capacity over the next three years.
SPIL is targeting a 15-fold increase in optical-engine manufacturing capacity for co-packaged optics by mid-2027, with plans to double that capacity again by the end of 2028.
The developments highlight the increasingly interconnected roles of outsourced semiconductor assembly and test (OSAT) providers, substrate manufacturers and optical-component suppliers in supporting AI hardware production.
For the advanced packaging industry, the expansion reflects growing demand for high-density integration and the manufacturing capacity needed to assemble increasingly complex computing systems.
The investment in optical engines also points to the growing importance of photonic connectivity as AI infrastructure scales.
As chip architectures become more complex, the availability of advanced packaging and testing capacity is becoming a critical factor in bringing next-generation AI platforms to market.
