At GTC Berlin 2026, NVIDIA announces major semiconductor supply-chain expansion plans, including a 2.5-fold increase in substrate capacity, more than 300% growth in advanced packaging capacity and a $1.1 billion advance to support new US testing facilities.
NVIDIA has announced a major expansion of its global semiconductor supply chain at GTC Berlin 2026, introducing new manufacturing agreements designed to increase production capacity for advanced AI processors, networking technologies and computing infrastructure.
The October 21 announcement includes partnerships with Amkor Technology, IBIDEN, King Yuan Electronics (KYEC), Samsung Electro-Mechanics, Siliconware Precision Industries (SPIL) and Unimicron Technology Corporation.
Together, the agreements target a 2.5-fold increase in substrate capacity over three years, more than 300% additional advanced packaging capacity at Amkor by the end of 2028 and a substantial expansion of semiconductor testing and optical-engine production.
NVIDIA is also providing a $1.1 billion capacity reservation advance to KYEC to support new testing facilities in the United States, with high-volume production targeted for April 2027.
The investments highlight how growing demand for AI computing is reshaping the semiconductor supply chain, directing capital towards manufacturing technologies that support the production, integration and testing of increasingly complex processors.
NVIDIA Targets a 2.5-Fold Increase in Semiconductor Substrate Capacity
One of the central components of NVIDIA’s strategy involves expanding the production of semiconductor package substrates, which provide the electrical connections and structural support required to integrate advanced chips into computing systems.
As AI processors become more complex, demand for sophisticated substrates is increasing alongside the need for advanced packaging technologies.
NVIDIA plans to increase aggregate substrate capacity for its products by 2.5 times over the next three years through agreements with manufacturers in Japan, South Korea and Taiwan.
In Japan, IBIDEN is expanding its Ono Plant with support from a multibillion-dollar capacity reservation advance announced earlier in 2026. The facility is targeting three times its existing capacity, with additional production expected to begin ramping during the first quarter of 2028.
Samsung Electro-Mechanics is joining NVIDIA’s substrate supply chain, providing immediate capacity from existing operations while supporting factory upgrades and additional facilities in Sejong and Busan, South Korea.
Meanwhile, Taiwan’s Unimicron is expanding its KF substrate manufacturing operations, with new capacity expected to become available beginning in the second quarter of 2027.
The agreements reflect NVIDIA’s effort to secure manufacturing resources ahead of anticipated demand rather than waiting for supply constraints to emerge.
Amkor Plans More Than 300% Growth in Advanced Packaging Capacity
Advanced packaging represents another significant area of investment.
Unlike conventional chip assembly, advanced packaging technologies allow multiple semiconductor components to be integrated within sophisticated computing systems, supporting higher performance and improved communication between processors and memory.
NVIDIA and Amkor Technology are expanding their collaboration in South Korea, targeting an increase of more than 300% in interposer wafer capacity by the end of 2028.
Interposers provide high-density connections between semiconductor components, making them important for complex AI accelerators and next-generation computing systems.
The expansion builds on an earlier agreement supporting Amkor’s advanced packaging operations in Arizona.
In July 2026, Amkor announced a $1.5 billion multiyear agreement with NVIDIA to support packaging and testing development in the United States.
On October 21, Amkor announced an additional $1.3 billion agreement focused on South Korea, bringing the companies’ multiyear agreements across both countries to approximately $3 billion, according to Amkor.
These commitments demonstrate the increasing importance of outsourced semiconductor assembly and testing companies within the AI infrastructure market.
NVIDIA Provides $1.1 Billion Advance for US Semiconductor Testing
Semiconductor testing is also becoming a major area of supply-chain investment.
Advanced processors require extensive validation before they can be integrated into high-performance computing infrastructure.
Testing helps manufacturers identify defects, verify performance and ensure that semiconductor components meet operational requirements.
NVIDIA expects its total semiconductor testing capacity to double between 2026 and 2027, supported by manufacturing expansion across the United States, Taiwan and Singapore.
The company is providing a $1.1 billion capacity reservation advance to KYEC, supporting new testing facilities and equipment in the United States.
High-volume production at the planned US facilities is targeted for April 2027, while expanded operations in Singapore are expected to begin high-volume production in February 2027.
These agreements illustrate how NVIDIA is using advance capacity commitments to help suppliers finance manufacturing expansion and secure future production availability.
Such arrangements can provide suppliers with greater visibility into expected demand, although their financial benefits will ultimately depend on production volumes, operating costs and customer requirements.
Optical Interconnects Become a New Manufacturing Priority
Beyond chips and packaging, NVIDIA is expanding capacity for optical engines used in co-packaged optics.
As AI computing systems grow, processors must exchange increasingly large amounts of data across servers and computing clusters.
This creates demand for high-bandwidth communication technologies that can support large-scale infrastructure while managing energy consumption.
NVIDIA and SPIL are planning to expand optical-engine production in Taiwan, targeting a 15-fold increase in capacity by mid-2027.
The companies are also aiming to double that expanded capacity again by the end of 2028.
Co-packaged optics integrates optical communication technologies more closely with computing and networking components, potentially improving performance and efficiency in demanding data-centre environments.
Although the technology remains subject to manufacturing and commercial deployment challenges, the scale of NVIDIA’s planned expansion highlights its strategic importance for future AI infrastructure.
What the Agreements Mean for Semiconductor Investors
The announcements demonstrate how AI infrastructure spending is creating opportunities beyond the companies designing GPUs and AI models.
Semiconductor substrates, advanced packaging, testing services and optical components are essential parts of the manufacturing process.
As demand for AI computing increases, businesses specialising in these technologies may benefit from additional manufacturing contracts and longer-term capacity commitments.
For NVIDIA, advance payments and manufacturing agreements can help secure access to production resources and reduce exposure to certain supply-chain bottlenecks.
For suppliers, these arrangements may support capital expenditure by providing a clearer indication of customer demand.
However, larger capacity commitments also introduce financial risks.
Manufacturing expansion requires substantial investment, and returns depend on whether future demand justifies the additional production capability.
Delays in equipment installation, changes in semiconductor technology, competition and fluctuations in AI infrastructure spending could affect expected outcomes.
Investors should therefore distinguish between announced capacity targets and realised revenue, profitability or production output.
AI Infrastructure Growth Reshapes the Global Semiconductor Industry
NVIDIA’s manufacturing expansion reflects a broader transformation in the semiconductor industry.
Artificial intelligence infrastructure requires a complex network of suppliers, from advanced chip designers and manufacturing foundries to packaging specialists, testing providers and networking equipment manufacturers.
The increasing scale of AI computing systems means that constraints in any one part of this network can affect the availability of finished products.
NVIDIA’s latest agreements aim to strengthen capacity across several of these critical stages simultaneously.
The geographic distribution of the partnerships also demonstrates the international nature of semiconductor manufacturing, with major capacity additions planned across Japan, South Korea, Taiwan, Singapore and the United States.
For businesses and investors, the developments reinforce the importance of evaluating AI infrastructure as an interconnected industrial supply chain rather than focusing exclusively on individual chipmakers.
As demand for advanced computing continues to expand, the financial performance of suppliers will depend on their ability to convert these manufacturing commitments into reliable output, competitive production costs and sustainable returns.
Sources
- NVIDIA Newsroom — NVIDIA and Supply Chain Partners Expand Capacity to Accelerate AI Infrastructure Buildout (21 October 2026)
- Amkor Technology — Amkor Technology and NVIDIA Advance Strategic Partnership for AI Infrastructure (21 October 2026)
- Amkor Technology — Strategic Partnership With NVIDIA to Expand Advanced Packaging and Test (23 July 2026)
- NVIDIA — GTC Berlin 2026: Live Updates From Jensen Huang’s Keynote (21 October 2026)
- Reuters — NVIDIA and Amkor Strike $1.5 Billion Chip Packaging Deal (23 July 2026)
