Nvidia, the dominant player in chips for artificial intelligence models, plans to produce as much as half a trillion dollars’ worth of AI infrastructure in the US over the next four years through manufacturing partnerships.
Production of Nvidia’s latest generation AI chip, known as Blackwell, has begun at Taiwan Semiconductor Manufacturing Co’s new plant in Phoenix, the company said in a statement Monday. Nvidia is also building supercomputer manufacturing plants in Texas with Foxconn and Wistron, and partnering with Amkor Technology and Siliconware Precision Industries for chip packaging and testing operations in Arizona.
“Adding American manufacturing helps us better meet the incredible and growing demand for AI chips and supercomputers, strengthens our supply chain and boosts our resiliency,” Jensen Huang, chief executive officer of the Santa Clara, California-based company, said in the statement.
The US$500 billion figure refers to the combined value of all the goods Nvidia anticipates selling in to the supply chain for AI. In large part, the number reflects a commitment from the biggest cloud computing companies to build out and upgrade data centres with the latest gear.
That group, which includes Microsoft, Amazon.com and Meta Platforms, is expected to spend US$371 billion this year on AI facilities and computing resources, a jump of 44 per cent from last year, according to a report published last month by Bloomberg Intelligence.
Nvidia also said the effort would mark the first time that AI supercomputers are produced in the US, a development that President Donald Trump touted on Monday.
During an appearance at the White House, Trump said that Nvidia made the decision because of tariffs. “It’s one of the biggest announcements you’ll ever hear – because Nvidia, as you know, controls that almost the entire sector,” he said.
