Nvidia plans to spend several hundred billion dollars over the next few years on expanding its chip manufacturing in the United States, adding to an industrywide push to develop artificial intelligence infrastructure in the country.
The Silicon Valley company, a maker of semiconductors that power artificial intelligence, outlined plans to ultimately shift manufacturing from overseas at a tech conference this week in San Jose. CEO Jensen Huang said in an interview with the Financial Times during the gathering that the newest chips designed by his company and Nvidia-powered servers for data centers could now be produced at U.S. factories operated by suppliers such as Taiwan Semiconductor Manufacturing Company and Foxconn.
Nvidia and other big tech firms have become highly dependent on TSMC’s chipmaking facilities in Taiwan, which are subject to risks including threats from China, earthquakes and taxes on imported semiconductors. A move to the U.S. would also help the company avoid any Trump administration tariffs on imports and protect its supply chain, Huang told the FT.
Nvidia is already using the TSMC’s Arizona factory to help produce some of its graphics processing units, which have become an essential component for the current wave of artificial intelligence investment. The supplier recently announced it was committing $100 billion to build chip manufacturing facilities in the state, which came on top of a $65 billion investment that was agreed to under the Biden administration.
“The most important thing is to be prepared,” Huang told the FT. “At this point, we know that we can manufacture in the U.S., we have a sufficiently diversified supply chain.”
Other big technology firms have recently moved to shore up their stateside manufacturing operations in light of Trump’s “America First” trade policies. Apple last month said it would invest more than $500 billion in the United States in the next four years, with plans to hire 20,000 workers and open a manufacturing facility in Texas to make servers to support its AI system, Apple Intelligence. The company said it plans to keep expanding data center capacity in North Carolina, Iowa, Oregon, Arizona and Nevada, while opening an Apple Manufacturing Academy in Detroit to help companies pivot to “advanced manufacturing.”
At the San Jose developers’ conference, which has been alternately dubbed “the AI Woodstock” and “the Super Bowl of AI,” Huang stressed in his keynote speech that his company, the world’s most valuable chipmaker, wants to be more than that — it plans to make the equipment that powers the AI universe where we all may someday live.
Robots and supercomputers
The conference was also a chance to convince the tech industry that Nvidia’s chips are still crucial to AI at a moment when investors have questioned whether the pace of investment in the infrastructure was warranted, given others had shown progress with far fewer resources.
AI models launched this year by Chinese startup DeepSeek claimed AI could be developed at a fraction of the cost. Those concerns have compounded jitters about worldwide economic instability in the wake of the Trump administration’s trade policies.
But Huang told the Financial Times that he believed the Trump administration could accelerate the development of America’s AI industry. “Having the support of an administration who cares about the success of this industry and not allowing energy to be an obstacle is a phenomenal result for AI in the US,” he said.
The new products unveiled by Nvidia at the conference included more powerful chips, including a model called Vera Rubin that’s named after the American astronomer who discovered dark matter, which Huang said could ultimately be developed to train much larger AI models and accomplish more sophisticated tasks for a greater number of users. It also introduced a platform called Isaac GR00T N1 that it said would “supercharge humanoid robot development.” Nvidia is working with Walt Disney Co. and Google’s DeepMind on the project.
The firm also touted new personal supercomputer systems from Dell, HP and other manufacturers that will enable developers and scientists to use the devices to work on AI models at their desks. Some of Nvidia’s biggest customers, a group that includes Microsoft Corp. and Amazon, made their own commitments during the week to keep spending.