China’s edge in the AI robotics race lies in the Greater Bay Area

Thanks to the Greater Bay Area’s tech champions Huawei, BYD, Tencent and DJI, China commands nearly 40 per cent of the global supply chain for embodied AI

Shaoshan LiuandDing Ning

Published: 9:30am, 23 Mar 2024

Why you can trust SCMP

The recent unveiling of Figure AI’s humanoid, powered by OpenAI’s technologies, has taken the world by storm. Demonstrating unparalleled capabilities in understanding and reacting to its environment, this humanoid represents a monumental leap in the realm of intelligent, interactive robotics.

At the heart of this groundbreaking fusion lies embodied artificial intelligence. Embodied AI transcends traditional AI by integrating intelligence into physical entities like robots, enabling them to perceive, learn and dynamically interact with their environment.

This revolutionary approach not only allows robots to adapt and evolve in response to their surroundings but also paves the way for an accelerated shift towards the “autonomy economy”, transforming society.

The strategic significance of embodied AI places it at the forefront of the global technological race. In analysing the supply chain, our findings highlight the pivotal role China’s Greater Bay Area, with Shenzhen at its core, plays in the country’s embodied AI supply chain. This dominance underscores the area’s critical contribution to China’s competitive edge.

Embodied AI technologies encapsulate two major areas, the robotic “brain” empowered by AI foundation models and cloud computing platforms, and the robotic “body” which includes the integrated hardware and computing components.

The core components of the embodied AI supply chain include 3D vision sensors, force sensors to measure the force exerted on or by a robot, and micro-drive systems, which are miniature motors and actuators that enable precise movements and controls.

They also include the precision force control systems that regulate force output with high accuracy, the robot joint modules that allow for the articulation and movement of different parts, and the AI foundation models that provide a broad understanding of the world, language and tasks.

Also vital are specialised AI chips, robot control chips which are specialised microprocessors that manage the operations of robotic systems, and cloud computing platforms, which provide processing power for AI model training.

To gauge the embodied AI supply chain distribution, we identified the leading supplies from the United States, Europe, Japan and China in each of these categories. The results reveal that China commands 38 per cent of the supply chain, with the US, Europe and Japan commanding 26 per cent, 24 per cent and 12 per cent respectively.

In particular, China excels in sectors such as 3D vision sensors and robot joint modules, where cost effectiveness gives it a significant edge. Despite these achievements, China remains behind the US in critical technological domains like AI foundation models and AI chips.

02:48

China World Robot Expo 2023: Your next bartender could be a humanoid

Furthermore, an intriguing aspect of our analysis reveals that the Greater Bay Area accounts for over 55 per cent of China’s embodied AI supply chain. Remarkably, if considered on its own, the Greater Bay Area would command 24 per cent of the global embodied AI supply chain, on a par with Europe and surpassing Japan.

The Greater Bay Area, which includes pivotal cities such as Shenzhen, Hong Kong and Guangzhou, has evolved into a global hub of manufacturing and technological innovation since China joined the World Trade Organization in 2001, playing a vital role in propelling China’s tech-driven economic expansion.

Over recent decades, the area has seen the rise of tech giants such as Huawei Technologies, BYD, Tencent and DJI. Each has become a significant global force in AI and robotics, providing a robust foundation for China’s advancement in the embodied AI arena.

China has 12½ times more robots in its workforce than experts expected: report

Nonetheless, a gap is evident in the robotic AI chips sector, with a dominant player yet to emerge from the region. And despite Huawei and Tencent developing their own AI foundation models, their technological capabilities still do not match the leading standards set by OpenAI and Google, indicating areas for growth and improvement in China’s embodied AI development strategy.

In summary, the global race towards embodied AI underscores a pivotal transformation in how societies will navigate the autonomy economy. China’s Greater Bay Area is developing into the epicentre of the embodied AI supply chain, with not only a sophisticated and mature supply chain, but also tech giants to drive the development of advanced AI technologies.

To secure a competitive edge in the embodied AI race, China should focus on the development of embodied AI in the Greater Bay Area and amplify investments in AI foundation models and AI chips.

Dr Shaoshan Liu is a member of the ACM US Technology Policy Committee, and a member of the US National Academy of Public Administration’s Technology Leadership Panel Advisory Group

Dr Ding Ning is the managing dean of the Shenzhen Institute of Artificial Intelligence and Robotics for Society (AIRS)

Post