Nvidia announced on the 14th (local time) that it plans to produce up to $500 billion (approximately 700 trillion won) in artificial intelligence (AI) infrastructure in the United States over the next four years with its partners. Nvidia aims to manufacture hardware necessary for developing and executing AI, including AI chips and AI supercomputers, in the United States.
Nvidia announced this plan through its blog that day and noted it has secured over 1 million square feet (93,000㎡) of manufacturing space for this purpose.
Nvidia explained that it has already started production of the latest AI chip, Blackwell, at the world's largest foundry company, Taiwan Semiconductor Manufacturing Company (TSMC), located in Phoenix, Arizona, and is conducting packaging and testing operations with Amkor Technology and Silicon Ware Precision Industries.
In Texas, it is building supercomputer manufacturing facilities in collaboration with Foxconn and Wistron and expects to begin mass production within 12 to 15 months.
Nvidia's announcement is analyzed as a measure to alleviate burdens such as rising expenses resulting from President Donald Trump imposing high tariffs on trade partners to encourage corporations to produce products in the United States.
The Trump administration announced on the 11th that it would exempt chips, smartphones, computers, and other technology products and components from reciprocal tariffs while stating that it would introduce new tariff measures on imported semiconductors.