TAIPEI (Taiwan News) — Taiwanese manufacturers are using AI in their factories to accelerate AI server production, Nikkei Asia reported Wednesday.
At Nvidia GTC in Taipei last month, CEO Jensen Huang (黃仁勳) announced that the company's next-generation Vera Rubin AI server platform had entered full-scale production.
Huang said assembling one Grace Blackwell server rack previously took about two hours but now takes just five minutes. A standard Vera Rubin rack combines 18 processor trays with cooling, power, networking, and data management systems to function as a single high-performance AI computer, per Nikkei Asia.
Foxconn Chair Young Liu (劉揚偉) said the company can manufacture one Vera Rubin server tray every minute without human intervention. Beyond factory automation, Foxconn uses AI to analyze sensor and production data in real time, helping managers identify problems and optimize manufacturing operations.
Through its Genesis smart manufacturing initiative, Foxconn has increased production line capacity and efficiency by 50%, while improving fault diagnosis accuracy to 90%, according to Yang Chiu-chin (楊秋瑾), head of the company's AI software business.
Other Taiwanese manufacturers are adopting similar technologies. Server maker Wistron uses AI-generated data to create defective samples in about a second, training inspection models on different defect patterns. The approach has boosted detection accuracy by 5% to 20%.
The rapid adoption of AI coincides with Taiwan’s AI server industry's continued expansion. According to Nikkei Asia, information and communication equipment exports, including servers, rose 83.8% year-on-year to NT$5.5 trillion (US$170.71 billion) in the first half of this year, accounting for more than 40% of the country’s total exports.
