NVIDIA has set the stage for a dramatic escalation of global capacity to manufacture key AI infrastructure components, announcing a multi-year plan that leverages extensive partnerships and billions in financial commitments to shore up its supply chain across multiple continents.
On October 21, 2026, NVIDIA disclosed it would work with manufacturing and packaging partners in Japan, South Korea, Taiwan, Singapore, and the United States to address surging demand for its AI hardware platforms, ranging from advanced GPUs to networking systems. The company intends to increase substrate capacity by 2.5 times within three years, and—with Amkor Technology and Siliconware Precision Industries (SPIL)—aims to expand advanced packaging output by over 300% by the close of 2028.
Semiconductor testing is also in sharp focus, with NVIDIA planning to double total testing capacity from 2026 to 2027, particularly by expanding U.S. and Asian facilities. In a move that signals its urgency, NVIDIA advanced $1.1 billion to King Yuan Electronics (KYEC) to secure capacity for new U.S. testing lines, seeking high-volume output by April 2027.
Collaboration has extended to Foxconn, which is set to further integrate NVIDIA’s next-generation GPU and networking platforms into its own vast global production ecosystem, bolstering the company’s ability to meet customer needs.
Strategic Partnerships and U.S. Focus
This latest expansion builds upon NVIDIA’s recent multiyear agreement with Corning, announced in May 2026, which seeks to increase American optical connectivity capacity tenfold. Corning is now undertaking three advanced facility projects in North Carolina and Texas, expected to generate over 3,000 new manufacturing jobs and reinforce the company’s U.S. supply base for critical components like high-performance optical fiber—essential for powering hyperscale AI data centers.
Repurposing and Scaling Amid Intense Industry Competition
NVIDIA’s infrastructure strategy also involves consideration of alternative site deployments. The firm is reportedly evaluating the conversion of more than 130 former telecom properties, originally owned by Brightspeed, into smaller-scale AI data centers across as many as 20 U.S. states. This potentially accelerates new AI compute capacity while repurposing underused real estate assets.
However, as NVIDIA pursues these bold expansions, competitors are intensifying their own investments. AMD recently committed over $10 billion to boost advanced manufacturing in Taiwan, while OpenAI is reportedly exploring chip sourcing diversification—planning purchases from Google in addition to reliance on NVIDIA GPUs—signaling growing fragmentation and hedging across the AI chip market.
Supply Chain and Sustainability Questions
NVIDIA has amassed a reported $530 billion in financial guarantees across its supplier base—a sum raising questions about sustainability even as it seeks to entrench itself at the heart of the global AI supply chain. Meanwhile, the environmental footprint of massive AI infrastructure buildouts remains a source of debate: in a parallel development, AI firm Anthropic has inked a $3.4 billion deal to secure gas generator capacity for its data centers, prompting concerns over the sector’s carbon impact.
While NVIDIA’s latest announcement outlines the broad geography and scale of its capacity plans, some specifics—including precise facility addresses—have yet to be made public. The company maintains that these new manufacturing capabilities will begin coming online between 2027 and 2028, with further ramp-up expected into the decade’s end.
Looking Forward
This sweeping wave of investment and partnership underscores not only the insatiable market hunger for AI compute power but also the complexity and stakes of global tech supply chains in the race for leadership. As efforts progress, the spotlight will remain on how NVIDIA and its partners balance rapid scale, sustainability, and technological innovation in the years to come.