Nvidia commits $1.5B to Amkor to expand US chip packaging and secure AI supply chain

Nvidia Secures Critical Chip Packaging Capacity with $1.5 Billion Deal

Nvidia has committed $1.5 billion to Amkor Technology in a prepayment agreement designed to expand advanced semiconductor packaging and testing operations in the United States1

1

. The announcement on July 23 sent Amkor shares surging roughly 12% in after-hours trading, later climbing 17% in extended trading as investors recognized the strategic importance of the partnership1

1

3

3

.

The multi-year agreement addresses a critical bottleneck in AI infrastructure development. While chip packaging may be the least glamorous stage of chipmaking, it has become one of the most contested1

1

. This is where finished silicon is stitched together with memory and interconnects into the dense modules that power AI servers, and demand for this work has outpaced available capacity.

Advanced Semiconductor Packaging Becomes Strategic Priority

Shortages of advanced packaging, rather than raw wafer fabrication, have at times been the factor limiting data-center chip shipments for Nvidia1

1

. This makes locking in capacity a strategic priority rather than a procurement footnote. The prepayment structure allows Amkor to build ahead of demand it might otherwise hesitate to fund alone, with the money specifically funding expansion of Amkor's operations in Arizona, where the company is already constructing a $2 billion packaging and test facility in Peoria1

1

.

Nvidia has increasingly used its balance sheet to turn scarce inputs into guaranteed ones. Over the past year, the company secured a warrant-backed deal with Corning that brought new US plants and thousands of jobs, and recently locked in SK Hynix for HBM4 memory1

1

. The Amkor prepayment applies the same logic to the back end of the production line.

Developing Next-Generation Packaging Technologies for AI and Accelerated Computing

The two companies will jointly develop packaging and testing technologies for what they describe as next-generation AI and accelerated computing platforms1

1

3

3

. This work includes high-density interconnects and heterogeneous integration, the practice of combining different types of silicon—logic, memory, and more—inside a single package1

1

2

2

.

As increasingly powerful computing systems push chipmakers to find new ways to combine processors, memory, and other components, advanced packaging has become critical to keeping power consumption and data-transfer bottlenecks under control2

2

. Instead of relying only on making individual chips smaller, manufacturers can combine different types of silicon and components in a single package, improving performance and energy efficiency while allowing specialized chips to work together more closely.

Building Supply Chain Resilience and Strengthening Semiconductor Supply Chains

"AI is driving a generational shift in technology, transforming every industry and creating a unique opportunity to reinvigorate American manufacturing and supply chains," said Debora Shoquist, Nvidia's executive vice-president of operations1

1

2

2

. Kevin Engel, Amkor's chief executive, noted that the partnership "underscores the central role advanced packaging plays" in enabling the future of AI1

1

2

2

.

Both sides framed the expansion as a hedge as much as a capacity play. Concentrating packaging in a handful of Asian sites has long been a single point of failure for the industry, and a domestic alternative is meant to give AI accelerator production more geographic resilience1

1

. The Arizona expansion fits a broader onshoring push, with Amkor's Peoria plant already awarded $407 million under the CHIPS Act1

1

.

Strategic Positioning in Crowded Market

The deal lands on top of an already crowded year for Amkor. The company packages chips for both Nvidia and AMD, and in June signed a 10-year agreement with TSMC for advanced packaging in Arizona1

1

3

3

. Amkor already supplies advanced packaging for Nvidia's product portfolio, including data center processors, networking chipsets, and accelerated computing systems2

2

3

3

, so the new commitment widens existing work rather than starting it.

The facility sits close to TSMC's Arizona fabs, whose wafers it is well placed to package without shipping them back to Asia first1

1

. Keeping fabrication and packaging within the same state closes one of the last gaps in onshore semiconductor supply chains that policymakers have spent years and billions trying to assemble. Most advanced packaging capacity still sits in Asia, and reproducing it in the US is slower and more expensive, but Nvidia's suppliers have been moving regardless. The collaboration will span the United States and Asia, linking technology development with manufacturing capacity as AI computing platforms continue to evolve2

2

.