1 hour ago
17
Jensen Huang has a habit of showing up where the action is. On July 21, 2026, the Nvidia CEO presided over the grand opening of Wistron’s D1 AI manufacturing facility in Fort Worth, Texas, a 324,000-square-foot building representing a $700 million bet that the future of AI hardware does not have to be assembled exclusively overseas.
This is Huang’s second Texas appearance in roughly two months. He was also on hand for the Coherent facility groundbreaking in June 2026.
What the D1 facility actually does
The Fort Worth plant is not a research lab or a data center. It is a production facility, and its first job is building the GB300 Grace Blackwell Ultra Superchip, Nvidia’s current flagship AI processor architecture. After that ramps up, the facility is slated to move into production of the next-generation Vera Rubin Superchip.
The facility is built using Nvidia’s own accelerated computing stack, including AI models like Nemotron and Cosmos, along with digital twin technology.
Wistron, for context, is a Taiwanese contract manufacturer and longtime Nvidia partner. This Fort Worth facility is its first footprint on US soil, which makes the $700 million investment a meaningful geographic pivot for the company.
The broader reshoring play
Huang described the current moment as “the largest infrastructure buildout in history” during his remarks at the opening.
The plant is expected to create 1,000 jobs by the end of 2026.
Huang also used the platform to push back on narratives that frame Taiwan as a vulnerability in the global semiconductor supply chain. His line was direct: “Taiwan saved the American semiconductor industry.”
What this means for investors watching AI infrastructure
When Nvidia announces production capacity for next-generation chips like the Vera Rubin Superchip, it is signaling where compute availability is heading over the next 12 to 24 months.
A US-based production node for GB300 chips reduces, at least partially, the concentration risk that comes from having critical AI hardware manufactured entirely in Asia.
Wistron’s willingness to commit $700 million to a first US facility is a signal about contract manufacturer confidence in sustained AI hardware demand.
Advanced manufacturing facilities take time to reach full yield rates. Early production of cutting-edge chips like the GB300 is rarely smooth, and any yield issues at the D1 facility would create bottlenecks that could affect Nvidia’s ability to fulfill orders.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·