In the past year, NVIDIA has been one of the most impressive stocks in the U.S. market, with its share price skyrocketing 174% throughout 2024. Its strategic position in the development of artificial intelligence is undeniable. Now, NVIDIA has taken a further step by relocating the manufacturing chain of AI supercomputers back to the United States, reflecting its dual considerations of localization and tariff policy hedging.
NVIDIA’s AI supercomputers are regarded as the engines of a new type of data center, featuring ultra-high GPU computing capabilities, dedicated networks, and storage architectures that can support the training processes of models such as GPT, Gemini, and Stable Diffusion. The so-called “AI factories” is a metaphor referring to data centers specifically designed to “produce AI models.” NVIDIA has stated that it plans to construct dozens of “gigawatt-level” AI factories in the coming years.
According to NVIDIA’s announcement, the company has already initiated over 1 million square feet of production space, with its Blackwell chips currently in production at TSMC’s foundry in Phoenix, Arizona. NVIDIA is also constructing supercomputer manufacturing plants in Texas, with Foxconn responsible for the Houston plant and Wistron for the Dallas plant. Both factories are expected to enter mass production within the next 12 to 15 months.
NVIDIA indicated that through partnerships with global leading manufacturers such as TSMC, Foxconn, Wistron, Amkor, and SPIL, it aims to produce AI infrastructure worth up to $500 billion in the United States. This move reflects NVIDIA’s strategic direction of shifting part of its supply chain and production capacity back to the U.S. Such localized production also helps to avoid high import tariffs, strengthen supply chain resilience, and manage geopolitical risk.
NVIDIA founder and CEO Jensen Huang stated:
“The core engine of global AI infrastructure is now being built domestically for the first time in the United States. Relocating the manufacturing chain back to the U.S. will help us better respond to the growing demand for AI chips and supercomputers while enhancing our supply chain and resilience.”
This adjustment in NVIDIA’s production capacity layout is closely related to the “Day of Liberation” tariff policy recently promoted by U.S. President Trump. This policy has had a significant impact on the global technology supply chain, and NVIDIA’s actions are seen as a proactive response to mitigate trade policy risks.
