Wistron Launches First US AI Server Factory, Mass-Producing Nvidia GB300 

Wistron has officially opened its first US AI server manufacturing facility in Fort Worth, Texas, marking a major milestone in Nvidia's effort to localize advanced AI hardware production in the United States. The US$700 million (NT$22.7 billion) D1 plant is the first facility to mass-produce Nvidia's GB300 Grace Blackwell Ultra Superchip computing boards on US soil. It will also manufacture the next-generation Vera Rubin Superchip. The opening ceremony was attended by Nvidia CEO Jensen Huang, Wistron Chairman Simon Lin, Taiwan's Representative to the US Alexander Yui, and local officials.

The 324,000-square-foot facility is expected to become a key pillar of Nvidia's US AI infrastructure strategy. Wistron is ramping production to tens of thousands of AI computing boards per month while creating more than 500 jobs, with headcount expected to reach 1,000 by the end of the year. The company has also begun planning a much larger second Texas facility, D2, as demand for AI servers continues to accelerate.

The Texas plant showcases Wistron's advanced smart manufacturing capabilities. Before construction began, the entire factory was designed and simulated using Nvidia's digital twin technologies—including Omniverse, Metropolis, Nemotron, and Cosmos—allowing engineers to optimize production lines and train workers virtually. The facility currently operates at around 60% automation, focusing on AI server motherboard assembly and testing while enabling faster delivery through localized final assembly.

The investment reflects a broader expansion of Taiwanese technology manufacturers into the United States as global AI infrastructure spending accelerates. Nvidia has pledged to manufacture up to US$500 billion of AI platforms in the US, working with partners including Wistron, TSMC, and Foxconn. Jensen Huang described the initiative as part of the "largest infrastructure buildout in history." At the same time, Wistron emphasized that future AI competitiveness will increasingly depend on resilient, localized manufacturing and supply chains rather than individual production sites.

Reference Sources

1. Yahoo!Finance/Quartz - Wistron opens first U.S. factory for Nvidia AI servers in Texas
2. Taiwan News - Taiwan’s Wistron opens US plant for AI systems

Rubio to Meet Chinese Counterpart in Manila, US Official Says

US Secretary of State Marco Rubio will meet with Chinese Foreign Minister Wang Yi on the sidelines of the Association of Southeast Asian Nations gathering in Manila.

The two will meet in Manila, where foreign ministers from the 11-Member ASEAN and major partners have gathered this week to discuss a range of global issues.

Analysts have said a Rubio meeting with his Chinese counterpart was likely to focus on preparations for a second meeting between US President Donald Trump and Chinese President Xi Jinping after the two met last May.

Trump has said Xi will visit the US at the end of September, and Rubio on Sunday said that the summit was still anticipated.

Reference Sources

1. aawsat - Rubio to Meet Chinese Counterpart in Manila, US Official Says
2. usnews - Rubio Says He Discussed China President's US Visit in Meeting With Wang Yi
3. thejakartapost - TheJakartaPost
4. abscbn - Rubio says at ASEAN meeting he expects 'positive' US visit by China's Xi
5. nikkei - Rubio discusses Chinese president's US visit in meeting with Wang Yi

OpenAI Plans to Spend Over $30 Billion on Georgia Data Center

OpenAI plans to spend tens of billions of dollars on a massive new data center near Savannah, Georgia, as part of the ChatGPT maker’s push to keep up with surging demand for artificial intelligence software.

The startup has contracted with Georgia Power Co. and secured 3.2 Gigawatts of energy for the data center complex in Effingham County, potentially setting the facility up to rank among OpenAI’s largest to date. Several hundred Megawatts are expected to come online starting in 2028, with the rest of the development extending through 2032. A single Gigawatt is enough to power about 750,000 us homes at any given time.

OpenAI committed to spending $20 billion in the county to qualify for an incentive package that will be voted on this week. At the full 3.2-Gigawatt size of the project, the cost would probably be more than $30 billion, said Sachin Katti, vice president of compute strategy at OpenAI.

The startup is currently looking for partners to help finance and develop the data center, Katti said, and may opt to do some of the operation itself.

Openai Ceo Sam Altman has previously said he expects the company to spend trillions of dollars on physical infrastructure to support its long-term AI ambitions.

The company has partnered with developers on large data centers in Europe, the Middle East and several us States as part of its Stargate initiative.

Reference Sources

1. bloomberg - OpenAI Plans to Spend Over $30 Billion on Georgia Data Center

US threatens sanctions against Chinese AI models over IP theft

The US Treasury Secretary, Scott Bessent, has warned that the US could sanction Chinese artificial intelligence (AI) firms if they are found to be stealing American intellectual property.

Bessent stated that the US would examine open-source AI models created overseas, including those from China, and take action on any evidence of intellectual property theft from US companies.

He added that the US supports open-source models, but what it does not support is IP theft.

The administration is finding Watermarks of American large language models on many of the Chinese models, and “That’s unacceptable,” Bessent said.

He also suggested that pressure could be placed on companies using Chinese AI.

The rapidly increasing global competitiveness of China’s AI models is ringing alarm bells in Washington, posing a new test for the standard protectionist playbook.

If the Trump administration decides to try to curtail China’s open-weight models, however, it would run into challenges unseen with previous protectionist moves. AI giants like Anthropic, OpenAI, and Google have each raised alarms about China-based labs using unauthorized distillation – a technique in which a more advanced model is used to train a new one without permission.

Reference Sources

1. bloomberg - OpenAI President Says Kimi K3 ‘Pretty Good,’ Unsure If Distilled
2. bloomberg - China’s ‘AI for All’ Offensive Defies US Containment Playbook
3. techcrunch - US threatens sanctions against Chinese AI models over IP theft
4. bloomberg - Bessent Says US Will Scrutinize Chinese AI Models for IP Theft
5. newyorkpost - Chinese AI firms that rip off US models could face sanctions, Treasury Secretary Scott Bessent warns

The CommonWealth English daily news digest is a service curated by CommonWealth English team with the help of AI tools.  

Have you read?

• Taiwan to Choke Mobile Internet Speed to Simulate War, Disasters
• Jeff Bezos and UK Government Back CuspAI to Reinvent Chip Materials with AI
• Boeing Sees Demand for 44,000 New Fuel-Efficient Jets Over Next 20 Years

Uploaded by Ian Huang