At the seventh annual Hon Hai Tech Day 2026 (HHTD26) in Taipei, Hon Hai Technology Group (Foxconn) demonstrated its industrial leadership in token factory solutions. The event highlighted the components and system technologies that enable next-generation artificial intelligence infrastructure, emphasizing the company's deep-running industry partnerships.

As the world's largest electronics manufacturer and AI server provider, Foxconn is leveraging its vertical integration and precision manufacturing capabilities to support advanced platforms, including the NVIDIA Vera Rubin platform. The exhibition showcased everything from rack-level full AI supercomputing system integration to sub-micron-level component manufacturing technologies.

The Evolution of AI Infrastructure: The NVIDIA Vera Rubin Platform

A central highlight of the event was the introduction of the NVIDIA DSX POD. This standardized, factory-built reference design allows partners to preintegrate NVIDIA Vera Rubin rack-scale AI infrastructure inside modular containers. By enabling critical validation before shipment, the system allows units to be connected directly to prepared site infrastructure immediately upon delivery.

This modular approach represents a shift in how large-scale AI infrastructure is deployed. Foxconn Chairman Young Liu explained the company's strategic direction:

"Foxconn is transforming into a technology manufacturing platform service provider. From testing to the tangible, our vertical integration and precision manufacturing makes the impossible possible. We are building this kind of strength across ecosystems of smart consumer electronics to AI and EV, as well as in space and robotics."

To honor these collaborative milestones, the event featured a dedicated customer zone displaying key moment photos and milestone technology timelines that trace the history of these behind-the-scenes partnerships.

Inside the Token Factory: Precision Engineering at Scale

Of the record 321 exhibits presented at the Taipei Nangang Exhibition Center, a quarter were located within the specialized Token Factory zone. This zone illustrated the extreme co-design required to build modern AI hardware.

A major visual centerpiece was a 20-meter long wall displaying:

  • PCBs mounted with tens of thousands of micron-level precision-engineered components.
  • An AI server compute tray design that is cable-free, hose-free, and fanless in a single generation.
  • Co-packaged optics (CPO), which represents the next major AI megatrend.
  • Re-imagined power, cooling, and high-speed data solutions designed to maximize the efficiency of AI infrastructure.

Co-Designing the AI Factory: Simulated Data Halls and Visionbay.ai

The event featured an immersive simulated AI data hall to demonstrate how AI compute, networking, power, and cooling are co-designed for an AI factory environment. This setup included NVIDIA 800 VDC power distribution systems.

Visionbay.ai, Foxconn's dedicated business unit for AI supercomputing and cloud operations, and an NVIDIA Cloud Partner, detailed its facility design plans. The business unit is planning its facility design to follow NVIDIA DSX MaxLPS principles. This approach aims to maximize AI factory performance per watt within a fixed power envelope, optimizing computing, power, and cooling configurations prior to actual construction.

Ecosystem Partnerships and Embodied Intelligence

The conference brought together global industry leaders to discuss the future of infrastructure and mobility. Onstage panel discussions featured representatives from:

  • NVIDIA, Google LLC, and Schneider Electric, who discussed the requirements of AI infrastructure.
  • Honda R&D Co Ltd and Mitsubishi Electric Mobility, who focused on next-generation electromobility.

The event also assembled a highly concentrated display of embodied intelligence in the Robotics and Humanoid zone. This area showcased the potential of a growing ecosystem, demonstrating real-world application examples of industrial products used in robotics.

In the Precision Manufacturing zone, Foxconn highlighted its heritage in ultra-precision manufacturing processes. These processes are backed by the company's proprietary machining, materials processing, and mold making equipment. Notable exhibits in this zone included:

  • Poketomo AI companion robots.
  • One-punch enclosures for laptops.

Prioritizing AI Safety and Community Engagement

As AI systems become more autonomous, safety remains a critical focus. Foxconn teams demonstrated how the NVIDIA Open Agent Safety Platform can support agent safety across multiple domains, including:

  • Smart healthcare.
  • Smart manufacturing and robotics.
  • Smart city applications.

The two-day annual conference, held at the Taipei Nangang Exhibition Center, Hall 1, 1F and 4F, also introduced a family-friendly pavilion. This space allowed children to engage with hands-on exploratory themes focused on sustainability, mobility, smart factory, and smart city concepts, extending the reach of technology education to the next generation.