Foxconn announced on Friday that its $1.4 billion supercomputing facility—developed in partnership with Nvidia
—is on schedule for completion by mid-2026. Once operational, the centre will become Taiwan’s largest GPU cluster and Asia’s first supercomputing site powered by Nvidia’s advanced Blackwell GB300 chips.
The 27-megawatt data centre, housed under Foxconn’s new AI-focused unit Visonbay.ai, will significantly strengthen the island’s role in global AI infrastructure. Neo Yao, CEO of Visonbay.ai, said the centre positions Taiwan as a regional hub for AI cloud operations and machine learning development in sectors like robotics, automation, and autonomous driving.
During Foxconn’s technology day, Nvidia vice president Alexis Bjorlin emphasized that as GPU computing advances so rapidly, localized, smaller-scale builds are becoming uneconomical. Instead, large-scale cloud compute clusters, such as this upcoming facility, enable on-demand flexibility and efficient scaling for both enterprise and research-level workloads.
Foxconn, widely known as Apple
’s primary iPhone assembler, has steadily diversified beyond contract manufacturing into emerging technologies such as electric vehicles and artificial intelligence hardware. The company now serves as Nvidia’s main integrator for AI server racks—high-performance frameworks that power next-generation cloud platforms and corporate data centres.
The ongoing AI data centre boom has proven a strategic advantage for Foxconn. With cloud computing leaders massively expanding their AI capacity, Foxconn projects that related demand will drive substantial revenue in 2026. Chairman Young Liu recently told Reuters that the company plans to invest between $2 billion and $3 billion annually toward next-generation AI research and infrastructure.
Foxconn founder Terry Gou also appeared at the event alongside Spencer Huang, Nvidia’s robotics product manager and son of founder Jensen Huang. Spencer highlighted Nvidia’s collaboration with Foxconn to bring AI-driven efficiency to manufacturing plants, where automated systems and edge AI could reduce assembly time and energy use significantly.
Liu further revealed that Foxconn can currently manufacture up to 1,000 AI racks weekly, a production rate expected to rise as global AI infrastructure needs expand. The company also showcased its “Model A” electric vehicle, engineered by Japanese designers, underscoring Foxconn’s goal to set up localized assembly operations in Japan and strengthen its footprint in the Asian EV market.
TAIPEI — Taiwanese tech giant Foxconn and OpenAI announced on Friday an agreement to design and build AI data center hardware, the latest in a string of infrastructure deals for the US creator of ChatGPT. https://t.co/kGUU64Zhue
— The Manila Times (@TheManilaTimes) November 21, 2025
