NVIDIA is moving production to the US. Not video cards, but AI clusters for now
When Trump imposed tariffs on the rest of the world last week, many companies, including nVidia, were seriously threatened. Although the law was immediately "frozen" for 90 days, the uncertainty of the future policy poses serious risks.
NVIDIA, which is no longer associated only with video cards, has announced large-scale plans to build new supercomputers in the United States. The company has put into operation more than 300 km2 of new production lines in the US and plans to invest more than USD 500 billion in localisation of production over the next 4 years.
NVIDIA Blackwell chips, the next-generation chips behind the company's artificial intelligence computing power, have already begun production at Taiwan Semiconductor Manufacturing Company (TSMC) in Phoenix, Arizona.
In addition to chip manufacturing, the company is building supercomputer fabs in Texas, teaming up with Foxconn in Houston and Wistron in Dallas. NVIDIA expects both sites to reach production targets within 12-15 months.
"The engines of the world’s AI infrastructure are being built in the United States for the first time - Jensen Huang".
The company is not only relying on traditional production methods. NVIDIA plans to use its own artificial intelligence, robotics, and digital twin technologies to design and operate its fabs. Using platforms such as NVIDIA Omniverse and NVIDIA Isaac GR00T, the company will model and automate complex manufacturing processes at its new facilities in the US. AI is literally taking over AI manufacturing.
Currently, no one doubts the further development of AI and its prospects. The United States is systematically moving strategic production away from China so that the Chinese government does not have access to it in the event of Taiwan's takeover. However, given the role of American investors in the development of DeepSeek and other Chinese high-tech giants, we cannot rule out the possibility that the Americans themselves will help China build the infrastructure for AI.
