Foxconn said that the $1.4 million supercomputing cluster it is building in Taiwan with Nvidia will be ready by the first half of 2026. Once completed, it will be Taiwan’s largest advanced GPU culture.
The 27-megawatt data center will be powered by Nvidia’s new Blackwell GB300 chips and is also set to be Asia’s first GB300 AI data center, said Neo Yao, CEO of a new unit Foxconn, has established for AI supercomputing and cloud operations called Visonbay.ai.
“As GPU technology accelerates, building individual facilities may no longer make economic sense,” said Alexis Bjorlin, a Nvidia vice president, at the contract electronics manufacturer’s tech day, which was attended by Foxconn’s partners and clients including Nvidia, OpenAI and Uber. “Renting compute resources may offer a far better return on investment, enabling flexibility and enabling companies to scale their compute according to both product and business cycles,” she added.
READ: Foxconn, SoftBank team up on Stargate data center project in Ohio (
Foxconn, which is Apple’s top iPhone assembler, has been expanding into electric vehicles and AI data centers. It is now Nvidia’s main maker of AI racks, which are server racks tailored for AI workloads that house chips, cables and other equipment. The company is also teaming up with SoftBank to build a Stargate data center in Ohio.
This has made the company a big beneficiary of the data center boom, as cloud computing firms spend billions of dollars to expand their AI infrastructure and research capacity. The company offered a bullish outlook on AI last week, noting that it would be a big driver of growth in 2026. Chairman Young Liu said in an earnings call that the importance of AI would massively increase next year, though the company would also keep a wary eye on geopolitical and currency issues.
Liu told Reuters that Foxconn would invest $2 billion to $3 billion a year in AI. Foxconn’s founder, Terry Gou, also made an appearance at the tech day as did Spencer Huang, a product line manager at Nvidia’s leading robotics product who is also the son of Nvidia founder Jensen Huang. Huang said that Nvidia was working with Foxconn to bring AI to factories and manufacturing lines.
READ: Foxconn remains bullish on AI, hints at OpenAI announcement (
Liu said the company now had the capability to manufacture 1,000 artificial intelligence racks per week, and it expected that rate to increase next year. He also added that Foxconn’s EV volumes were just at about the level where automakers could outsource more production to the company and Chief Strategy Officer Jun Seki showcased the company’s “Model A” electric vehicle on stage.
Liu also said that the Model A was designed by Japanese engineers and that Foxconn planned to eventually set up a company there to serve Japanese customers. The Model A will eventually be made in Japan too, he said.