Wistron opens $700 million Texas AI facility to build NVIDIA GB300 chips

Wistron's first U.S. plant in Fort Worth has begun producing NVIDIA GB300 Grace Blackwell Ultra boards as the company expands domestic AI hardware assembly and backs broader U.S. onshoring efforts.

33d ago

Fact Check

All four sources—NVIDIA's official blog, Wistron's own PR Newswire release, Reuters, and CNBC—independently confirm the core facts: Wistron opened a $700 million, ~324,000 sq ft facility in Fort Worth, Texas, that produces the NVIDIA GB300 Grace Blackwell Ultra Superchip (the first U.S. mass production) and will produce the Vera Rubin Superchip. The event occurred on July 21-22, 2026. The only slight imprecision in the claim's phrasing is 'build NVIDIA GB300 chips'—the facility assembles/mass-produces GB300 Superchip systems rather than fabricating raw silicon—but this does not materially contradict the sources, which describe the plant as producing GB300 Superchips.

Summary

Wistron has inaugurated its $700 million D1 AI Smart Facility in Fort Worth, Texas, its first manufacturing plant in the United States and a key site in its global AI infrastructure network. The roughly 324,000-square-foot facility is producing NVIDIA GB300 Grace Blackwell Ultra systems, with the latest report saying the plant is designed to turn out tens of thousands of boards per month. The new facility opened less than a year after Wistron announced a Texas investment that was previously reported at $761 million, with Fort Worth selected as the site in August 2025 and the plant operating by July 2026. Nvidia CEO Jensen Huang cast the project as part of a wider push to rebuild advanced manufacturing capacity in the United States, and Wistron said the site was planned and optimized using Nvidia Omniverse digital twin technology. The Fort Worth plant is expected to support about 1,000 jobs by the end of 2026, adding to earlier reporting that the project was expected to create more than 600 full-time roles. Wistron's shares rose about 9.7% after the opening announcement. The factory adds a domestically based assembly and testing link for AI server hardware as Nvidia and its manufacturing partners increase U.S. production capacity.

Terms & Concepts

digital twin technology: Software-based virtual replicas of physical facilities or processes used to simulate and optimize real-world operations.
AI infrastructure: The hardware, systems and production capacity needed to build, run and scale artificial intelligence workloads.
Grace Blackwell Ultra: NVIDIA's high-end AI computing platform used in advanced server systems for training and running large AI models.