The Rise of Humanoid Robots in Manufacturing by 2025

In a significant development for the manufacturing sector, Boston Dynamics is set to deploy its all-electric humanoid robot, Atlas, within a Hyundai factory later this year. This marks a pivotal moment as it will be the first instance of Atlas being utilized in commercial manufacturing, following its evolution from the hydraulic model that gained fame through viral demonstrations since 2013.

Boston Dynamics, acquired by Hyundai in 2021 for $1.1 billion, has not disclosed the specific tasks that Atlas will undertake. However, company representatives indicate that this robot is engineered to outperform human workers in strength and reliability. The capabilities of Atlas include lifting heavy items and managing awkward loads, tasks that can be challenging for human employees.

The year 2025 is poised to see a broader introduction of multipurpose humanoid robots into the workforce, moving beyond the confines of research environments. Companies like Agility Robotics and Figure are already seeing their humanoid robots, including Digit and the Figure biped, begin to engage in paid work, primarily in warehousing.

Tech giants, including Apple and Meta, are also reportedly exploring the development of consumer-oriented humanoid robots. According to a report from Goldman Sachs, the market for humanoid robots is projected to reach $38 billion by 2035, showcasing a sixfold increase from previous estimates.

The fundamental appeal of humanoid robots lies in their ability to perform various tasks, similar to human workers. This is in contrast to traditional assembly line automation, which typically requires a manufacturing environment tailored to specific tasks. Experts like Jonathan Hurst, co-founder of Agility Robotics, suggest that humanoid robots will complement existing automation rather than replace it entirely. For operations that do not require constant activity, flexible robots could enhance productivity.

Boston Dynamics emphasizes that its design approach for Atlas considers the need for robots to function effectively in human-centric environments. This design philosophy aims to ensure that robots can navigate spaces originally intended for human activity.

Despite the promising advancements, challenges remain in the commercialization of humanoid robots. Tesla's Optimus robot, highly anticipated since its announcement in 2021, faced scrutiny after a demonstration revealed that the showcased robots were controlled by humans, raising doubts about their autonomous capabilities. Recent statements from Tesla indicate plans to produce thousands of Optimus units in 2025, although production may be affected by restrictions on rare-earth metal exports from China due to trade tariffs.

A significant hurdle in deploying multipurpose humanoid robots has been the lengthy process of training them for various tasks. However, experts believe that advancements in large language models will facilitate this learning process. Companies like Google have recently introduced AI models designed to enhance robots' adaptability, thereby improving their ability to learn on the job.

With the capacity to learn tasks dynamically, humanoid robots could perform a wide range of functions within retail environments, from unpacking shipments to managing inventory. This versatility is expected to unlock substantial value in various work settings. Furthermore, improved natural language processing capabilities may allow robots to receive verbal instructions, simplifying the programming of new tasks.

Nevertheless, safety concerns persist regarding the operation of heavy robots in proximity to human workers. Experts highlight the importance of reliability in robotic performance. For instance, a robot tasked with restocking shelves may perform flawlessly for an extended period before encountering a critical failure. Such incidents could lead to costly consequences, underscoring the need for robust reliability measures.

Despite these challenges, the rapid advancements in AI technology provide hope for the future of humanoid robots in the workforce. Experts express optimism that the integration of large language models could lead to significant improvements in robotic capabilities, paving the way for a new era of automation in manufacturing and beyond.