NVIDIA Mobilizes Global Semiconductor Partners as AI Infrastructure Demand Accelerates

Semiconductor industry is entering a new phase of capacity expansion as the rapid development of artificial intelligence infrastructure increases pressure on the manufacturing ecosystem behind high-performance computing. NVIDIA’s latest supply chain initiative reflects this shift, bringing together key semiconductor manufacturing partners to strengthen production capabilities across several critical stages of chip development and integration.

NVIDIA has announced plans to expand global manufacturing capacity alongside Amkor Technology, IBIDEN, King Yuan Electronics (KYEC), Samsung Electro-Mechanics, Siliconware Precision Industries (SPIL), and Unimicron Technology Corporation. The initiative focuses on substrates, advanced packaging, semiconductor testing, and optical engines – technologies that are becoming increasingly important as AI data centers scale to accommodate demanding computing workloads.

Rather than focusing exclusively on chip production, the expansion addresses a broader industrial requirement: ensuring that the components surrounding high-performance processors can be manufactured, packaged, tested, and connected at the pace required by AI infrastructure deployment.

• The growth of AI computing is increasing demand across multiple layers of the semiconductor supply chain. High-performance processors require sophisticated packaging technologies, specialized substrates, extensive testing, and increasingly capable interconnect solutions to operate within large-scale computing systems.
• NVIDIA CEO Jensen Huang described the situation as an infrastructure buildout of unprecedented scale, highlighting manufacturing capacity as a major constraint on the industry’s ability to respond to demand.
• The company’s latest initiative seeks to strengthen production capacity across multiple geographic locations while improving supply chain resilience and redundancy.

Key developments shaping the expansion include:

• Substrate manufacturing: NVIDIA targets a 2.5-fold increase in substrate capacity over the next three years.
• Advanced packaging: Planned collaboration with Amkor aims for an increase of more than 300% in interposer wafer capacity by the end of 2028.
• Semiconductor testing: Total testing capacity is expected to double between calendar years 2026 and 2027.
• Optical connectivity: SPIL and NVIDIA are targeting a 15-fold expansion in optical-engine manufacturing capacity by mid-2027, with plans to double that capacity again by the end of 2028.

These targets illustrate how the AI infrastructure investment cycle is extending beyond processor design into the manufacturing processes required to deliver complete computing platforms.

Advanced substrates play a critical role in connecting semiconductor dies, supporting electrical pathways, and enabling the integration of increasingly complex computing packages. As AI accelerators become more sophisticated, substrate availability and manufacturing capability are becoming important considerations for scaling production.

• NVIDIA plans to increase substrate capacity for its products by 2.5 times over three years, supported by expansions involving IBIDEN, Samsung Electro-Mechanics, and Unimicron.
• In Japan, NVIDIA is supporting IBIDEN’s expansion at its Ono Plant through a multibillion-dollar capacity-reservation advance. The project targets production capacity at the facility reaching three times current levels, with additional capacity expected to begin ramping in the first quarter of 2028.
• South Korea is also gaining a larger role in NVIDIA’s substrate supply chain. Samsung Electro-Mechanics is joining as a supplier, initially bringing capacity from its existing operations while supporting facility upgrades and new factories in Sejong and Busan. By 2028, production allocated to NVIDIA is projected to represent a significant share of the company’s annual package-substrate production capacity.

Meanwhile, Unimicron is expanding KF substrate manufacturing capacity in Taiwan under a capacity-reservation agreement with NVIDIA. Additional capacity is expected to begin coming online in the second quarter of 2027, with the expansion targeting full capacity by mid-2028.

Together, these developments demonstrate how semiconductor supply chain partnerships can help secure manufacturing capacity well ahead of the anticipated production requirements of future computing platforms.

Advanced Packaging Moves Further Into the AI Spotlight

As semiconductor systems become more complex, advanced packaging is emerging as a critical enabler of performance, integration, and manufacturing scalability.

NVIDIA and Amkor plan to expand advanced packaging capacity in South Korea, targeting a greater-than-300% increase in interposer wafer capacity by the end of 2028. The initiative will complement Amkor’s development of new manufacturing facilities in the country.

The expansion is intended to support increased production volumes for NVIDIA’s current and next-generation platforms.

For the broader semiconductor industry, the development reinforces the growing importance of packaging technologies that integrate multiple components into high-performance systems. As AI computing requirements rise, access to advanced packaging capacity can become a key consideration alongside wafer fabrication and processor availability.

The planned expansion also underscores the importance of geographically diversified manufacturing partnerships in supporting the industry’s longer-term production needs.

Testing Capacity Expands Across Three Major Manufacturing Hubs

• Producing advanced chips is only part of the process. Semiconductor devices must undergo extensive testing to assess performance, reliability, and functional requirements before they can be integrated into computing systems.
• NVIDIA expects its total semiconductor testing capacity to double from calendar 2026 to 2027, supported by expansion projects in the United States, Taiwan, and Singapore.
• A major element of the initiative is NVIDIA’s $1.1 billion capacity-reservation advance to KYEC. The arrangement will support new US testing facilities and equipment, with high-volume production targeted for April 2027.
• In Singapore, expanded testing operations are expected to enter high-volume production in February 2027.
• These investments highlight the growing strategic importance of semiconductor testing infrastructure as manufacturers work to accommodate increasingly complex chips and rising system-level performance requirements.
• Expanding testing capacity across multiple locations could also improve supply chain flexibility by providing additional manufacturing support in different regions.

What This Means for the Global Semiconductor Industry?

NVIDIA’s latest expansion plans reflect a broader shift in semiconductor manufacturing priorities. The AI infrastructure buildout is creating demand not only for powerful processors, but also for the specialized technologies and manufacturing services required to bring complete systems to market.

Several industry implications stand out:

• Supply chain diversification: Capacity expansion across the United States, Japan, South Korea, Taiwan, and Singapore strengthens the geographic breadth of NVIDIA’s manufacturing ecosystem.
• Greater emphasis on packaging: Advanced packaging and substrates are receiving increased attention as chip integration becomes more complex.
• Testing infrastructure investment: Expanded testing capacity highlights the need to scale validation and quality-control processes alongside production.
• Optical connectivity development: Investment in optical engines points to the growing role of advanced interconnect technologies in large AI computing environments.
• Long-term capacity planning: Capacity-reservation agreements and multiyear expansion targets illustrate how suppliers and technology companies are coordinating investments to address anticipated demand.

However, announced targets and planned investments should not be interpreted as immediately available production capacity. The actual contribution of these projects will depend on construction, equipment installation, qualification, manufacturing ramp-up, and customer demand.

NVIDIA’s collaboration with its manufacturing partners illustrates how the next stage of AI development depends on a wider network of semiconductor technologies. Substrates, advanced packaging, testing, and optical engines are becoming central to the industry’s ability to manufacture and deploy increasingly capable computing systems.

As expansion projects progress toward their targeted timelines between 2027 and 2028, execution will be an important factor in determining how quickly additional capacity reaches production.

The defining takeaway: The global AI race is no longer only about designing faster chips. It is increasingly about building the manufacturing depth, packaging capabilities, testing infrastructure, and optical connectivity needed to deliver AI computing at scale.