Chipmaker launches Arizona–Texas production hub amid tariff uncertainty and surging global demand
Nvidia has announced plans to build and mass produce its AI supercomputers entirely within the United States for the first time, marking a major shift in the company’s supply chain strategy as geopolitical trade tensions and demand for AI infrastructure continue to accelerate.
The move involves commissioning over a million square feet of new manufacturing space in Arizona and Texas. Production of Nvidia’s advanced Blackwell chips has already begun at Taiwan Semiconductor Manufacturing Company’s (TSMC) facility in Phoenix, while supercomputers built using those chips will be assembled at upcoming plants in Houston and Dallas, operated by Foxconn and Wistron respectively. Mass production at both Texas sites is expected to ramp up over the next 12 to 15 months.
The company said it expects to produce up to US$500bn in AI infrastructure over the next four years through these US-based partnerships.
“The engines of the world’s AI infrastructure are being built in the United States for the first time,” said Nvidia CEO Jensen Huang. “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.”
The announcement comes as the Trump administration escalates its trade war rhetoric, threatening new tariffs on imported semiconductors while temporarily exempting certain tech categories—including AI chips—from its sweeping reciprocal tariff regime. President Trump described the Nvidia investment as a “Trump effect in action,” while Commerce Secretary Howard Lutnick has confirmed that semiconductors remain under active review as part of upcoming national security tariff investigations.
Strategic realignment amid policy volatility
Nvidia’s U.S.-based production will not only include chip fabrication via TSMC, but also testing and packaging partnerships with Amkor and Siliconware Precision Industries (SPIL) in Arizona. These partners will handle the intricate processes of integrating memory and logic modules into finished Blackwell processors.
Supercomputer assembly in Texas will rely on Nvidia’s DGX systems—high-performance AI computing units that can be scaled into vast AI factory clusters. Each DGX system contains eight Blackwell GPUs, up to 2TB of flash storage, and advanced networking components. Foxconn, which already manufactures most of Nvidia’s AI servers globally, and Wistron, which is investing US$50m into its Dallas site, will manage the new assembly lines.
Nvidia will use its own AI technologies—including the Isaac GR00T platform for humanoid robot development and its Omniverse digital twin software—to automate and manage factory operations.
Economic and political implications
The investment is being framed by both Nvidia and the White House as a milestone for domestic economic security. The company says the initiative will create hundreds of thousands of jobs and drive trillions of dollars in downstream economic value.
Huang’s statement and the scale of investment have drawn praise from the Trump administration, which is using the news to bolster its messaging around reshoring manufacturing and confronting China’s trade practices. Trump has already imposed a 32% tariff on Taiwanese imports and 145% on Chinese goods, though he has since temporarily exempted several tech categories—including AI chips—from immediate enforcement.
However, the exemption may be short-lived. Trump has warned that “nobody is getting off the hook,” and reiterated his intention to target the “entire electronics supply chain” in future tariff rounds.
Market reaction
Despite the scale of the announcement, Nvidia shares slipped slightly during Monday trading, falling 0.2% to US$110.71. The stock is down approximately 18% year-to-date, largely due to market concerns over escalating trade policy volatility and supply chain disruptions.
Investors remain cautious as questions persist over whether U.S. production can match the efficiency and cost structure of Nvidia’s current overseas supply base. Analysts also noted that while the company will build its supercomputers domestically, some components—including rare earth materials and memory modules—will likely still be sourced from overseas.
