Foxconn to Invest NT$42B in AI Supercomputing Hub

On October 27, Foxconn (Hon Hai Technology Group) announced that its board has approved an investment plan of up to NT$42 billion (about RMB 9.74 billion) from December 2024 to December 2026. The funds will be used to purchase information equipment for AI computing and cloud infrastructure, forming a large-scale AI cluster and supercomputing center.

According to Foxconn, all funding will come from its own resources. The move aims to expand its cloud-based AI computing platform and accelerate the development of its three core intelligent platforms — smart manufacturing, smart electric vehicles, and smart cities.

Industry analysts point out that this could become Taiwan's largest NVIDIA Cloud Partner (NCP) construction project in recent years. In addition to Foxconn, NVIDIA, TSMC, and Taiwan's National Science and Technology Council (NSTC) may also utilize the computing capacity from this initiative.

Foxconn Chairman Young Liu recently revealed that the company is crafting its next five-year plan, emphasizing that "AI applications are just beginning." He noted the tremendous market potential of AI, with estimates suggesting that building a 1GW-class data center could cost at least USD 50 billion. Liu believes that AI will soon permeate all industries, offering vast growth opportunities.

Earlier in mid-October, Foxconn also announced a partnership with NVIDIA to introduce an 800V DC power architecture in Kaohsiung, Taiwan, co-developing the infrastructure for future AI Factories. This project marks a key milestone in Foxconn's AI transformation, serving as a model for integrating AI servers, data centers, and renewable energy.

In its capital expenditure plan, Foxconn disclosed in August that its first-half CapEx reached NT$79.8 billion, up more than 25% year-on-year, and it remains committed to maintaining over 20% growth for the full year.

Foxconn's research institute added that the company is actively developing AI foundational model training technologies, leveraging its vast internal data to advance smart platforms. It emphasized that AI model training depends heavily on data volume and sufficient budget, underscoring Foxconn's strategic commitment to AI infrastructure and self-sufficiency.