Foxconn & Nvidia Launch 100MW AI Data Center in Taiwan

At Computex 2025, two tech titans—Foxconn and Nvidia—announced a groundbreaking partnership to build a 100-megawatt (MW) AI data center in Taiwan. This ambitious project positions both companies at the forefront of the AI infrastructure boom, aiming to meet the soaring global demand for generative AI, robotics, and high-speed data processing.

As reported by Reuters, the new facility will be rolled out in phases, beginning with an initial 20 MW capacity and scaling to 100 MW—making it one of the most powerful AI-focused data centers in Asia.

Why This Collaboration Matters

With the rise of AI applications like LLMs, robotics simulations, and real-time data processing, the need for advanced infrastructure has never been greater. Nvidia plans to power the data center using its cutting-edge Blackwell B200 GPUs and NIM (Nvidia Inference Microservices) platform, forming the computational backbone of this facility.

Meanwhile, Foxconn—best known for assembling iPhones—is evolving into a major cloud and AI infrastructure player through its Foxconn Industrial Internet (Fii) arm. Their strategic pivot is in sync with their broader AI vision, supported by their role as a manufacturer and now, as a service provider.

This aligns with Nvidia’s broader push for scalable, modular AI deployments, as highlighted during their unveiling of the new NVLink Fusion interconnect technology, which facilitates high-bandwidth chip-to-chip communication.

Project Details

• Location: Kaohsiung, Taiwan

• Initial Phase: 20 MW

• Target Completion: End of 2026

• Applications: AI inference, training models, robotics, digital twins

Foxconn's vision is to make this data center a central hub for Asia-Pacific AI development. The company also intends to power the facility with sustainable energy sources to meet green computing demands.

Part of a Larger AI Boom

This move is not isolated. Multiple tech firms are doubling down on AI hardware and infrastructure. For instance, Intel recently introduced its new Arc Pro B60 and B50 GPUs for workstations and edge computing. As covered by the Times of India, these GPUs are designed to support AI workloads for enterprise users.

In the realm of hardware and humanoid robotics, Ray Kurzweil’s startup is also making waves. According to Reuters, the company is currently in talks for a $100 million investment to build AI-powered humanoids, further reflecting the industry's collective momentum.

Why This Is a Big Deal

For Nvidia, this partnership gives it a stronger footprint in the Asia-Pacific region, while advancing its software stack—CUDA, Omniverse, and NIM. Foxconn, on the other hand, is stepping into a new era, where AI data centers and platform services are just as vital as assembling consumer electronics.

And they’re not alone. Other OEMs, like TECNO, are also using Computex to showcase AI-first hardware and ecosystem strategies, showing that the AI era is a global race.

The Future Ahead

As AI-as-a-Service (AIaaS) continues to scale, data centers like the one planned by Foxconn and Nvidia will form the computational backbone of tomorrow's AI economy. From autonomous driving to real-time industrial simulations, these mega infrastructures are the brains of the intelligent world.