By Wen-Yee Lee
TAIPEI (Reuters) -Taiwanese electronics manufacturer Wistron said on Friday that its new U.S. manufacturing facilities for its customer Nvidia would be ready next year and the firm was in talks with potential other customers.
Part of the facilities will be used by Nvidia to support its plan to build artificial intelligence servers worth up to $500 billion in the U.S. over the next four years.
The U.S. firm said in April it planned to build supercomputer manufacturing plants in Texas, partnering with Foxconn in Houston and Wistron in Dallas. Both sites are expected to increase production within 12 to 15 months.
“I believe it’s exactly in line with what the customer announced. All our progress will follow the customer’s lead,” Wistron CEO Jeff Lin said in his first public comments since Nvidia’s announcement.
He said the company was in discussion with other customers to use the U.S. facilities, which will produce high-performance computing and AI-related products, declining to disclose their names.
Wistron’s board approved a total investment of $500 million in its new U.S. subsidiary to support business development and strategic growth in the U.S.
Commenting on U.S. restrictions on exports of advanced chips to China, Lin said demand outside of China remained very strong.
“We expect to grow alongside our customers … As for developments in the Middle East, most of them are essentially our indirect customers.”
The United Arab Emirates and the U.S. signed an agreement this week for the Gulf country to build the largest AI campus outside the U.S., in a deal that according to sources could involve purchase of 500,000 of Nvidia’s most advanced AI chips per year starting in 2025.
As the U.S. threatens sweeping tariffs on many of its trading partners, Wistron said it would also consider producing notebooks in Mexico, noting that such products would not face tariffs under the United States-Mexico-Canada trade agreement.
(Reporting by Wen-Yee Lee; Editing by Miyoung Kim and Tomasz Janowski)
Comments