Foxconn will build a data center for NVIDIA Corporation (NVDA.US) in Taiwan, which is expected to be operational in the first half of next year.

Foxconn announced plans to create the largest GPU cluster in Taiwan and deploy its first supercomputing center based on NVIDIA Corporation's GB300 NVL72 system. The supplier of Apple Inc. and NVIDIA Corporation stated that the GB300 NVL72 system is expected to be operational in the first half of next year, marking the acceleration of Taiwan's local AI infrastructure development. Foxconn stated that the new AI factory infrastructure supports scalable enterprise computing capability. The 27-megawatt data center is reportedly worth about $1.4 billion. However, Foxconn and NVIDIA Corporation declined to comment on the investment amount and facility power capacity. It is reported that Foxconn's parent company, Hon Hai Precision, has established Visionbay.ai, a business unit dedicated to AI supercomputing and cloud operations. According to the company, Visionbay's business model is centered around an AI factory ecosystem, providing GPU as a service, compute leasing, NVIDIA Corporation's native software solutions, and a cloud-based AI app store. The company added that by instantly accessing high-performance training, fine-tuning, development, and inference capabilities, enterprises can adopt AI with the lowest barriers, supported by NVIDIA Corporation's next-generation accelerated computing architecture. In addition, Foxconn announced a partnership with OpenAI to produce key components for its data centers in the United States. The company had previously hinted at an announcement related to OpenAI in its third-quarter earnings report and provided an optimistic outlook for next year in the report, stating that AI-related demand will be a key driver of growth in 2026.