Embracing the Dawn of AI and Humanoid Robotics: How to Capitalize on the Industry's Breakout Year

The development of AI is expected to bring revolutionary changes to the humanoid robot industry. End-to-end AI is completely different from prior rule-based control, meaning the software system itself can execute the task from original commands and scenarios to final outputs under AI self-generated rules instead of software engineers' pre-programmed rules.

Goldman Sachs recently raised the 2035 total addressable market forecast to US$38bn, a 4-fold increase in shipments estimate to 1.4mn units with a much faster path to profitability on a 40% reduction in bill of materials. Tech analyst Jacqueline Du has revised her projections, indicating that factory applications for technology could become economically feasible earlier than previously anticipated, within the timeframe of 2024 to 2027 (compared to the initial forecast of 2025 to 2028). Furthermore, she predicts that consumer applications are likely to follow suit, becoming economically viable a bit later, from 2028 to 2031.

■ How could AI expand the application fields of humanoid robots

The advancements in AI have taken the world by storm. The progress made in end-to-end AI, which operates on principles distinct from rule-based control systems, has the potential to expedite the development cycles of humanoid robots significantly. This is evidenced by the noticeable enhancements in the manipulation and interaction features of several products that emerged in 2023.

AI has expanded the application of robots to more fields: manufacturing sectors that have not yet been replaced by industrial robots; fatal manual work (such as hazardous chemical industry, disaster rescue, and nuclear industry maintenance), elderly care, and household fields (expected to be commercialized by 2028).

Research shows the current state of manufacturing in China shows that 70% of production tasks are managed by existing machinery and automation systems, with 20% still reliant on human labor and the remaining 10% overseen by managerial personnel. Given this distribution, the integration of industrial robots, which excel in performing structured and precise operations, alongside humanoid robots—known for their versatility and ability to navigate complex environments as opposed to just flat surfaces—presents an opportunity to further expand the automation potential within the industry.

Humanoid robots are also attractive for performing tasks that fall into the category of "dangerous, dirty, and dull" (3D), owing to the high risk of fatalities associated with such jobs and people's general aversion towards undertaking them. This aversion often translates into a willingness by customers to pay a premium for automation solutions that can safely handle these types of work. Data from the US Bureau of Labor Statistics indicates that there are approximately 5,000 fatal work injuries in the United States each year. If we include the automotive manufacturing industry, assuming labor substitution rates of 5%, 10%, and 15% across the above potential applications, the demand for humanoid robots could potentially reach between 1.1 million and 3.5 million units on a global scale.

■ Who are the main manufacturers?

Several companies worldwide are at the forefront of manufacturing humanoid robots, each with its own unique focus and contribution to the field.

Among the major players, Boston Dynamics has gained widespread attention for its robots, including the humanoid robot Atlas, which showcases advanced mobility, balance, and agility.

Major Chinese manufacturers include Fourier and Kepler. Chinese robots have traditionally been used in the logistics and package delivery industry, and are gradually expanding into the medical and educational fields.

■ Manufacturing costs drop faster than expected

The production expenses for humanoid robots have seen a substantial decrease, with costs plummeting from an estimated range of $50,000 to $250,000 per unit last year to the current $30,000 to $150,000, which spans from basic to advanced models. This roughly 40% reduction is primarily due to the more affordable components now available, a wider range of supply chain choices beyond the confines of research laboratories, and improvements in design and manufacturing processes. This cost efficiency suggests that the commercial deployment of humanoid robots could accelerate, outpacing previous expectations based on an analysis that compares the trajectory of the humanoid robot's cost against its output and performance.

■ What production components are involved?

The production of humanoid robots includes actuators, sensors, control systems, power supply, chips, software, and communication systems for those requiring human-like appearance or tactile sensing.

6) Software: This includes the operating system, drivers, and applications that perform data processing, decision-making, machine learning, and movement coordination. Before the advent of the AI era, software was the biggest bottleneck for robots.

Risk Disclosure: Robot revenues may still account for a small proportion of the revenue of the companies mentioned in the article.