NVIDIA backs six chip suppliers in $279B push to break AI capacity bottleneck
What's the deal? NVIDIA announced at GTC Berlin on October 21 that it will fund six supply chain partners to expand manufacturing capacity across semiconductor substrates, advanced packaging, testing, and optical engines. The partners are Samsung Electro-MechanicsDealroom has a profile for this one. Try Dealroom →, Japan's IbidenDealroom has a profile for this one. Try Dealroom →, Taiwan's Unimicron, US-based Amkor, and Taiwan's KYEC and SPILDealroom has a profile for this one. Try Dealroom →. Together, the substrate agreements aim to lift capacity 2.5-fold over three years.
What's the endgame? NVIDIA is moving from GPU supplier to backer of the components needed to make them. The company previously funded cloud buyers like CoreWeave; now it is securing scarce inputs at the production end too.
The details: NVIDIA signed a $1.3 billion deal with Amkor to raise advanced packaging capacity in South Korea more than 300% by the end of 2028. It paid KYEC a $1.1 billion advance to build testing facilities in the US. SPIL agreed to grow optical-engine capacity 15-fold by mid-2027.
For substrates, NVIDIA provided a multibillion-dollar advance to triple Ibiden's Ono plant in Japan, with new output from early 2028. Samsung Electro-Mechanics will expand lines in Sejong and Busan, while Unimicron adds Taiwanese capacity from mid-2027.
Why now? "Demand is extraordinary. Capacity is the constraint," said chief executive officer Jensen Huang. The push reflects NVIDIA's view that manufacturing — not chip design — now limits how fast AI infrastructure can ship.
What could go wrong? The upfront commitments improve supply access but raise NVIDIA's exposure if demand or project timelines slip. In August, the Wall Street Journal reported NVIDIA's supplier commitments had reached $279 billion in its latest quarter, up from $119 billion the prior quarter.
The signal: NVIDIA is spreading capital across both sides of the AI chip market, deepening its grip on the entire supply chain. By tackling bottlenecks beyond GPU fabrication, it is betting AI infrastructure demand stays strong — and that controlling the inputs is the surest way to keep systems moving.
Read more: chosun.com
Image credit: pestoverde
