NVIDIA has announced that it is working with its manufacturing partners to design and build factories that will produce AI supercomputers entirely in the United States for the first time. More than a million square feet of space has already been commissioned in Arizona for the production and testing of Blackwell chips, and a build and test AI servers have been established in Texas.

Production of Blackwell chips has begun at TSMC's Fab 21 in Phoenix, Arizona, and NVIDIA is building AI server manufacturing facilities in Texas, with Foxconn's facility in Houston and Wistron's in Dallas, both of which are expected to ramp up to volume production over the next 12 to 15 months. Because the AI chip and supercomputer supply chain is so complex, it requires state-of-the-art manufacturing, packaging, assembly, and testing technologies, for which NVIDIA is also partnering with Amkor and SPIL for packaging and testing operations in Arizona.

NVIDIA plans to produce up to $500 billion worth of AI infrastructure in the U.S. over the next four years through a partnership between TSMC, Foxconn, Wistron, Amkor and SPIL. NVIDIA wants to keep deepening its partnerships to grow its business and increase supply chain resilience while expanding its global footprint. In addition, hundreds of thousands of jobs will be created and trillions of dollars of economic security will result.

Sources say NVIDIA needs to build its AI server chain by investing in the U.S. in exchange for a moratorium on expanding restrictions on exporting AI chips to China.