South-California-based NVIDIA, in conjunction with Taiwanese foundry TSMC, announced the production of the first wafer—the basic‐substrate for its next‐gen “Blackwell” chips architecture—that was manufactured at TSMC’s semiconductor manufacturing facility situated in Phoenix (Arizona USA).
The event was marked by the appearance of NVIDIA’s CEO Jensen Huang at TSMC’s Phoenix/fab site. He describes the milestone as “the world’s most important chip” being made in the U.S. for the first time in recent memory.
TSMC’s Arizona facility is announced to support advanced process nodes (2 nm, 3 nm, 4 nm) as well as “A16” class chips, aimed at AI, telecommunications and high-performance computing applications.
While manufacture of the wafer is a major first step, the full chip assembly, packaging and test operations still need to scale. Observers note that this is not yet full mass production of fully packaged chips made entirely in the U.S.
It's worth noting that NVIDIA first introduced its Blackwell platform last year. The company presented it as a designed to transform the AI landscape. Major tech players such as OpenAI, Google, and Amazon, were among the first to commit to using the new architecture.
Related: OpenAI to Buy “Tens of Billions” Worth of AMD Chips.
The company claimed the platform delivers significantly higher performance while using less energy and cutting costs by up to 25 times compared to its previous generation. Now that the Blackwell wafers are being produced at TSMC’s Arizona facility, NVIDIA has gained a stronger foothold in domestic manufacturing.
This milestone helps NVIDIA reduce its exposure to global trade uncertainties and geopolitical risks. It also supports the U.S. goal of rebuilding its chip production ecosystem.
As Blackwell enters large-scale production, the company is expanding its operations across the country. NVIDIA plans to invest around half a trillion dollars in AI infrastructure through partnerships with TSMC and Foxconn to strengthen America’s position in advanced computing.
