NVIDIA announced on October 21, 2026 that it is working with major supply-chain partners to expand capacity for accelerated computing and AI factories. The announcement identifies substrate, advanced packaging, semiconductor testing, and optical-engine partners—exactly the kinds of nodes that determine whether demand can become deliverable product.
The named nodes
NVIDIA said substrate agreements with IBIDEN, Samsung Electro-Mechanics, and Unimicron are intended to drive a 2.5-times increase in capacity over three years. Amkor plans to increase advanced-packaging capacity by more than 300% by the end of 2028. KYEC is expected to build new U.S. facilities supported by a $1.1 billion NVIDIA capacity-reservation advance, while SPIL is targeting a major expansion in optical-engine capacity.
Why the graph matters
- Substrates: constrain how quickly advanced packages can be assembled at scale.
- Packaging: connects chip supply to usable system output and can become a bottleneck even when wafers are available.
- Testing: determines whether completed devices can move into volume production.
- Optics: affects the interconnect layer required by large AI factories.
Investor interpretation
This announcement is not a guarantee of revenue and it is not an equity offering. It is a strategic effort to increase resilience and throughput by investing in the supplier network. The risks to monitor are execution timing, partner concentration, capacity utilization, and whether end-customer demand arrives quickly enough to absorb the new infrastructure.
Bottom line
NVIDIA’s competitive position depends on more than GPU design. The October announcement makes the surrounding supply chain visible: substrates, packaging, testing, and optics are the network’s critical paths. Mapping those relationships against future SEC disclosures can show where growth is supported—and where a bottleneck could still form.
Sources: NVIDIA supply-chain announcement; NVIDIA SEC filing.