The American semiconductor group Nvidia will spend up to $500 billion building supercomputers for artificial intelligence entirely in the US for the first time, as it becomes the latest US technology firm to bow to the Trump administration’s push for domestic manufacturing.
The $2.7 trillion chipmaker said it had commissioned more than a million square feet of manufacturing space to build and test supercomputers in Texas, partnering with Foxconn in Houston and Wistron in Dallas, two Taiwanese electronics manufacturers. Mass production at both plants is expected to ramp up in the next 12 to 15 months.
Nvidia’s chips are primarily made by Taiwan Semiconductor Manufacturing Company (TSMC) in Taiwan, although some of its Blackwell chips have also started production in TSMC’s factory in Arizona. Arizona-based Amkor and Siliconware Precision Industries, the Taiwanese semiconductor specialist, will provide packaging and testing operations.
• Tariff confusion leaves Big Tech braced for further market turmoil
The move followed similar announcements from a clutch of tech firms that have pledged to bring manufacturing back to the US amid threats of steep import tariffs.
Apple, which assembles most of its iPhones in China, in February promised half a trillion dollars in the US investments in the next four years, including a factory in Texas for artificial intelligence servers. Microsoft said it would spend $80 billion in 2025, with half of this investment focused on the US. Meta, the owner of Facebook , has said the company will spend up to $65 billion in 2025, including construction of a vast data centre in Louisiana.
“It is unlikely Nvidia would have moved any production to the US if it was not for pressure from the Trump administration,” said Gil Luria, an analyst at DA Davidson. “The half-a-trillion number is likely hyperbole, in the same way Apple made a half-a-trillion promise.”
Jensen Huang, Nvidia co-founder and chief executive, said: “The engines of the world’s AI infrastructure are being built in the United States for the first time. Adding that American manufacturing helps us better meet the incredible and growing demand for AI chips and supercomputers, strengthens our supply chain and boosts our resiliency.”
Supercomputers are “the engines” of a new type of data centre created for the sole purpose of powering AI, the company said, predicting that “tens of ‘gigawatt AI factories’ would be built in the coming years”.
Huang has forecast that Nvidia’s data centre infrastructure revenue to hit $1 trillion by 2028.
On Friday, hopes were raised that Nvidia and other technology companies might be handed a reprieve from Trump’s punitive tariffs on China after the White House published a list of exemptions on electronics imports. However, Trump later indicated that smartphones and other electrical items shipped from overseas would still be subject to tariffs, with rates due to be announced in the next week.
Nvidia was launched after a meeting between a trio including Huang at Denny’s restaurant in San Jose in 1993. Its chips are now used by more than 40,000 companies, from carmakers and drug-discovery businesses to weather forecasters and social media giants chasing superfast computing speeds to make the most of AI technology.
• Jensen Huang: the rise and rage of Nvidia’s visionary billionaire
It became the first chip company to reach a $1 trillion market valuation in 2023 and peaked at almost $3.4 trillion at the end of last year, as investors clamoured to gain exposure to AI infrastructure providers.
However, it has lost about $600 billion in value since then after advances made by DeepSeek, the Chinese AI start-up, rattled investors and Trump’s global tariffs have threatened demand.
The shares dropped $0.31, or 0.28 per cent, lower to $110.62 in afternoon trading in New York.
