Nvidia has announced plans to move part of its AI chip production to the United States, unveiling a commitment to more than one million square feet of manufacturing space across locations in Arizona and Texas.
The company revealed it has begun manufacturing its Blackwell AI processors at TSMC’s facilities in Phoenix, Arizona. Additionally, Nvidia is establishing supercomputer production sites in Texas through collaborations with Foxconn in Houston and Wistron in Dallas. To support the chip manufacturing process further, Nvidia will partner with Amkor and SPIL for essential packaging and testing operations in Arizona.
Nvidia expects mass production at these new facilities to scale up within the next 12 to 15 months. Over the following four years, the company aims to produce AI infrastructure worth as much as half a trillion dollars domestically.
Jensen Huang, CEO of Nvidia, stated, “For the first time, the engines powering the world’s AI infrastructure are being built in the United States. Expanding our American manufacturing capabilities enables us to meet the rapidly escalating demand for AI chips and supercomputers, strengthens our supply chain, and enhances our overall resiliency.”
The announcement closely follows Nvidia’s recent negotiations surrounding export controls. Reports indicate the company narrowly avoided restrictions related to the export of its advanced H20 chip. The H20, the most sophisticated AI chip Nvidia currently exports to China, was reportedly exempted after Huang assured U.S. government officials about Nvidia’s investments in new U.S.-based AI data centers.
