logo

Author: Jenny Wu

At the 2025 Viva Technology conference in Paris, humanoid robots were one of the most attention-grabbing topics, indicating the field is moving beyond proof-of-concept into real-world application. A total of 36 startups and companies related to robotics showcased the potential of humanoid robots in manufacturing, logistics, and service industries.

The event also featured a keynote speech by Nvidia CEO Jensen Huang, the global leader in AI chips. In his address, he emphasized the transformative impact humanoid robots will have on the future of industry, stating, “Soon, anything that moves will be a robot.”

What Are Humanoid Robots and Why Have They Advanced So Quickly?

“Humanoid robots” refer to machines designed to resemble and function like humans, typically with arms, legs, a head, and a torso. They can walk, grasp objects, converse, and perceive their environment. Their human-like form allows them to function seamlessly in human-centered environments with minimal need for adaptation. Several key technological advancements have propelled the development of humanoid robots:

1. Rapid Progress in Artificial Intelligence (AI): Improvements in generative AI, deep learning, and natural language processing have enabled robots to better understand and respond to human language, mimic expressions, and even make decisions—enhancing both the practicality and naturalness of human-robot interactions.
2. Breakthroughs in Sensing and Control Technologies: The combination of visual sensors, motion capture, real-time feedback, and machine learning algorithms has made humanoid robots more agile and adaptable, enabling them to walk on uneven surfaces or perform delicate tasks.
3. Improved Chip and Computing Performance: With high-performance computing platforms designed specifically for AI and robotics—such as those from Nvidia—robots now have more powerful “brains” to handle real-time processing, edge AI inference, and complex tasks.
4. Rising Demand for Industrial Applications: Labor shortages, aging populations, and the need for automation in hazardous environments have all increased the demand for humanoid robots across industries.

How Does AI Teach Robots to Walk and Work? — The Role of Virtual Simulation and Digital Twins

Traditionally, teaching robots to walk or perform tasks required extensive real-world testing, which was time-consuming, expensive, and risky. But with the development of AI simulation platforms such as NVIDIA Isaac Sim, developers can now create a “digital twin”—a virtual mirror of the physical robot and environment—for training in virtual space.

This “Sim-to-Real” approach dramatically shortens development cycles, reduces physical testing costs, and mitigates risk. AI can train thousands of virtual robots simultaneously using reinforcement learning to discover optimal solutions, which are then transferred to real-world machines.

This data-driven, simulation-based training is a key reason why humanoid robots today can acquire complex human-like behaviors in a fraction of the time it would take a person to learn them—bringing practical deployment within reach.

Market Response to Technological Breakthroughs: A Rapidly Expanding Industry

With simultaneous progress in simulation, sensing, control, and AI, humanoid robots are quickly moving from labs into real-world manufacturing and service settings, driving a new industrial revolution. The market is responding enthusiastically.

According to a May 2025 report by Research and Markets, the global humanoid robot market is projected to grow from $2.98 billion in 2025 to $24.34 billion by 2035, with a compound annual growth rate (CAGR) of 49.2%.

Morgan Stanley’s 2025 analysis suggests that when including software, maintenance, sensors, components, and related services, the overall market could surpass $5 trillion by 2050, with as many as one billion robots deployed globally. These projections demonstrate that humanoid robots are not just high-tech novelties—they are becoming central to the restructuring of labor and industry worldwide.

Foxconn and Nvidia: From Concept to Real-World Deployment

Humanoid robots are no longer limited to tech showcases; they’re now entering industrial production lines. According to a Reuters report from June 20, 2025, Nvidia and Foxconn Technology Group (also known as Hon Hai) plan to deploy humanoid robots in a new AI server factory in Houston, Texas, aiming to promote automation and smart manufacturing.

This partnership marks a significant milestone—the first large-scale deployment of humanoid robots on an actual production line. Foxconn will handle the factory's construction and operations, while Nvidia will provide core AI computing platforms and simulation training tools.

The robots are expected to perform tasks such as assembly, material handling, and quality inspection, trained in virtual environments via Nvidia’s Isaac Sim before being transitioned to the physical plant. This collaboration is more than a technological trial; it could have far-reaching implications for global manufacturing and supply chains.

Will Taiwan Miss Out or Lead in This AI Industrial Revolution?

Taiwan plays an indispensable role in the global AI server supply chain. As generative AI and intelligent automation rapidly advance, global demand for AI infrastructure is surging—benefiting Taiwan’s industrial ecosystem at every level.

At the chip level, TSMC (Taiwan Semiconductor Manufacturing Company) is the major manufacturer of Nvidia’s high-end AI chips like the H100 and GH200. TSMC’s technical prowess and capacity management directly affect the global supply of AI computing resources.

Further downstream, Taiwanese companies dominate key components such as PCBs, server motherboards, cooling systems, enclosures, and power supplies. Companies like Wistron, Quanta, GIGABYTE, Inventec, and Foxconn are critical suppliers to Nvidia, Microsoft, and Google.

These firms are not just hardware manufacturers—they are expanding into systems integration and edge AI applications, further strengthening Taiwan’s strategic position in the AI ecosystem. Even if Taiwan does not own leading humanoid robot brands or consumer-facing applications, its deep involvement in AI servers and computing infrastructure keeps it at the heart of this technological transformation.

Looking ahead, as demand for computing power and robotic applications continues to grow, the need for edge devices and data processing modules will also expand. If Taiwan can maintain its high-value supply capabilities and deepen its collaboration with global AI and automation brands, it could solidify its role as a global hub in the AI industrial revolution—and potentially move up the value chain into platforms and applications.

Disclaimer

This report is based on publicly available information and third-party media sources and is intended solely to provide readers with an overview of trends in humanoid robotics and AI technology development. The mention of specific companies, technological applications, or market forecasts does not constitute any form of commercial endorsement or investment advice. Readers are advised to conduct their own research and carefully evaluate associated risks before taking any action.

Reference

VivaTech 2025: Attack Of The Humanoid Robots

Exclusive: Nvidia, Foxconn in talks to deploy humanoid robots at Houston AI server making plant

人形機器人： 邁向人類2.0時代| Packaging Media

AI 的下一個關鍵 為何是人形機器人？

Humanoids: A $5 Trillion Market | Morgan Stanley

Humanoid Robots Market Industry Trends and Global Forecasts to 2035: Global Humanoid Robots Market to Surge from $2.98B in 2025 to $243.4B by 2035, Driven by AI and Automation - ResearchAndMarkets.com

AI需求急升  台灣居供應鏈核心：從科技巨擘到新創企業  搶進算力賽道