New Delhi: Taiwanese electronics manufacturer Foxconn and US-based chipmaker Nvidia have expanded their collaboration to develop AI solutions for smart cities, electric vehicles (EVs), and advanced manufacturing. The announcement was made during the ongoing Computex 2025 event in Taipei.
Focus on Domain-Specific AI Models
Foxconn, officially known as Hon Hai Technology Group, is leveraging its new AI foundational model, FoxBrain, to support customers in deploying generative AI across automotive and industrial sectors. The FoxBrain model is based on open-source architectures including Meta’s Llama-3 and Llama-4 and is optimized for domain-specific tasks, particularly in smart manufacturing and mobility.
According to Foxconn Chairman and CEO Young Liu, FoxBrain is designed for “agentic reasoning workflows” and is hosted at the company’s Kaohsiung data center. The model runs on Nvidia’s high-performance AI chips – H100, GB200, and GB300 demonstrating Foxconn’s reliance on Nvidia’s computing infrastructure.
Foxconn’s EV Push Gains Ground
Foxconn is working with Mitsubishi Motors to introduce its smart EV reference models. Liu said Mitsubishi will be the first original equipment manufacturer (OEM) to integrate Foxconn’s reference architecture into its future EVs. Foxconn’s EV arm, Foxtron, recently reached an in-principle agreement with Mitsubishi to jointly develop a new electric vehicle.
The reference models are intended to reduce development time for OEMs by approximately 80%, allowing them to focus more on design and customization.
Data Center to Support AI Workloads
To handle the increasing demand for AI processing, Foxconn will build a 100MW data center in Taiwan. The facility, equipped with Nvidia’s hardware and software, will be constructed in phases starting with 20MW. The data center will support what Liu referred to as an “AI factory”, a computing backbone aimed at enabling large-scale AI deployment across smart cities, EV platforms, and industrial applications.
Discussion about this post