CES 2026, the world’s largest technology trade show, has come to a close—and this year, one theme stood out above all others: robots. From humanoids and quadrupeds to sports and healthcare applications, robotics was everywhere on the show floor, signaling that the era of Physical AI—where AI and robotics come together in the real world—is quickly taking shape.

With Jae-bum Park, Senior Researcher at POSCO Research Institute, we look at what CES 2026 revealed about the future of robotics and where POSCO Group’s strengths could create new opportunities in this fast-growing industry.

Powering the Energy Transition: Robots Were Everywhere at CES 2026

At this year’s CES, robots were impossible to miss. Even in booths hosted by companies not traditionally associated with robotics, robot-related technologies appeared throughout the exhibition. The message was clear: AI-powered robots are moving beyond demonstrations and entering everyday life and industrial environments.

One of the most talked-about exhibits was Atlas, the humanoid robot presented by Boston Dynamics, Hyundai Motor Group’s robotics affiliate. Equipped with 56 degrees of freedom (DoF), Atlas demonstrated highly flexible movement and the ability to handle heavy-duty tasks involving loads of more than 50 kilograms—offering a glimpse of its potential in real industrial settings.

The exhibition also introduced a collaboration case involving POSCO and Boston Dynamics. On screen, Spot, Boston Dynamics’ quadruped robot, was shown moving through high-temperature facilities at a POSCO steelworks. In a noisy and heat-intensive environment, Spot was able to detect gas leaks and carry out precise inspections of equipment and infrastructure. The scene offered a compelling look at how robots are evolving from crowd-pleasing exhibits into reliable co-workers on industrial sites.

Another exhibit that drew attention came from Korean company Bodyfriend, which showcased a massage chair designed to feel almost like wearing a robotic suit. Its wearable AI healthcare robot, with independently moving arms and legs, attracted strong interest for its ability to analyze and stretch the user’s joints.

Robots were also making an impact in sports. Table tennis robots and combat robots once again proved popular with visitors, while sports robots introduced by Chinese companies such as Unitree and others showed just how far the technology has advanced. Some were able to analyze an opponent’s movements and track the trajectory of a ball in real time, providing highly detailed coaching.

Degrees of Freedom: What Makes Robot Movement More Sophisticated

As CES 2026 made clear, robot movement is becoming increasingly refined as technology advances. One term that many robotics companies use when discussing precision and flexibility is DoF, or Degree of Freedom. In simple terms, DoF refers to the number of joint axes a robot can move or control.

Generally speaking, the higher the DoF, the more independently controllable joints a robot has. This is particularly important for humanoid robots, which are expected to move in ways that resemble the human body. Industrial robot arms used in automated manufacturing processes typically have around 6 DoF, while humanoid robots usually range from 30 to 60 DoF. With 56 DoF, Boston Dynamics’ Atlas is close to the level required to mimic full-body human joint movement.

That said, increasing DoF also means adding more components such as motors, reducers, and sensors. In other words, the more sophisticated the movement, the higher the production cost. That is why robotics companies carefully design and allocate DoF based on each robot’s purpose and application.

Why Actuators Matter

An actuator is the drive unit that physically moves a robot’s joints. It can be understood as a complete joint system that combines a motor, reducer, sensor, and control circuit.

A motor on its own produces rotational motion. But for a robot arm to extend in a straight line or bend with precision, rotational movement alone is not enough. By integrating gears and electronic controls, an actuator enables more complex forms of motion, including both rotational and linear movement. If a motor is like a muscle cell, an actuator is the complete muscular joint that allows a limb to move in a controlled way.

For a humanoid robot with 56 DoF, roughly 56 precision actuators are needed.

Why Hyper NO Matters in the Age of Robotics

Electrical steel is also a critical material in motor manufacturing, accounting for roughly 20% to 30% of total motor production cost. This is especially important in robot motors, which must generate high output despite their compact size. In many cases, the performance of the electrical steel used inside the motor directly affects the motor’s overall efficiency and output.

To achieve better performance, the steel sheet must be made as thin as possible. If it is too thick, eddy currents* can form inside the material, causing energy loss in the form of heat.

*Eddy currents are swirling electric currents generated by electromagnetic induction when the magnetic field around a conductor changes rapidly.

POSCO’s high-performance electrical steel, Hyper NO, is designed to maximize magnetic performance even at extremely thin, paper-like thicknesses. The thinner electrical steel becomes, the more difficult it is to manufacture. In fact, only about five to six steelmakers worldwide, including POSCO, are capable of stably mass-producing electrical steel at the Hyper NO level. That makes it a strategic material with high technological barriers to entry—and an increasingly important one in the era of robotics and electrification.

Energy and Materials Will Help Power the Humanoid Robot Era

No matter how advanced a robot’s movement may be, it cannot function properly without enough energy. Today, humanoid robots typically operate continuously for only about two to four hours, making battery technology one of the biggest limiting factors in the industry.

At CES 2026, swappable battery systems emerged as a notable trend aimed at addressing that limitation. Boston Dynamics’ Atlas, for example, is known to operate continuously for around four hours, and it is even equipped with a function that allows it to replace its own battery when power runs low.

Tesla CEO Elon Musk recently predicted that there could be 10 billion humanoid robots within 25 years. If production reaches that scale, demand for actuators would rise into the hundreds of billions, making growth in demand for high-performance electrical steel almost inevitable. Once replacement demand is taken into account, the number of batteries required could also reach into the tens of billions.

This would also drive a sharp increase in demand for lithium, one of the key raw materials used in batteries. From this perspective, POSCO Group’s high-performance electrical steel technology and lithium assets are likely to draw growing attention as the robotics market continues to expand.

As robotics moves closer to real-world adoption, the future of the industry will depend not only on software and AI, but also on the materials and energy technologies that make advanced movement possible. In that future, POSCO Group is well positioned to play an important role.

Establishing a Collaboration Model Between Workers and Humanoid Robots to Expand Physical AI to Heavy and Heavy Industries

POSCO Group is accelerating the adoption of physical AI in manufacturing sites by pursuing a project to apply humanoid robots to steel product logistics management.

On the 3rd, POSCO Group signed an MOU at the POSCO DX Pangyo headquarters with representatives from four companies: POSCO, POSCO DX, POSCO Investment, and Persona AI, to implement industrial humanoid robots in the field.

Under the agreement, POSCO will identify potential work sites at its steelworks and assess their feasibility. POSCO DX will design and build a robot automation system and jointly develop a steelworks-specific model. POSCO Investment will support the implementation of a Proof of Concept (PoC), a small-scale preliminary test of ideas and technologies to determine their practical feasibility. Persona AI will be responsible for the development and implementation of a humanoid robot platform tailored to the steelworks industrial environment.

POSCO Group and Persona AI have agreed to begin testing Persona AI’s humanoid robot business in the logistics management of steel coils produced at the steelworks in February. Unloading rolled coils weighing tens of tons requires crane operation. The humanoid robot will collaborate with field workers to attach the crane belt to the coils. This is expected to be an example of human-robot collaboration creating a safer work environment.

Logistics work involving handling coils weighing 20 to 40 tons carries a high risk of accidents and the potential for musculoskeletal disorders due to repetitive tasks. The deployment of humanoid robots is expected to address these issues. Since last year, POSCO Group has been considering the introduction of humanoid robots in terminal logistics processes, such as transportation and material preparation, considering the characteristics of heavy construction sites. If this demonstration process confirms the mechanical safety and collaboration potential of humanoid robots with workers, the group plans to expand their deployment and apply them to various logistics sites.

Persona AI is a US-based humanoid robot startup founded in 2024 by renowned robotics engineers and industry experts. Prior to this agreement, POSCO Group invested a total of $3 million in Persona AI last year. Combining NASA’s robotic hand technology with its own precision control technology, Persona AI is developing industrial humanoid robots capable of everything from assembling fine components to handling heavy loads.

To respond promptly to the industrial paradigm shift, including AX, POSCO Group plans to expand its Intelligent Factory initiative across its manufacturing sites and create a safe and comfortable, technology-based workplace.

▲ (From the second from the left in the front row, moving right) Kyuho Chung, Head of POSCO DX Strategy Office; Minseok Shim, President & CEO of POSCO DX; Nicolaus Radford, CEO of Persona AI; Geunhwan Kim, President of POSCO Venture Capital, along with business representatives, pose for a commemorative photo after signing a Memorandum of Understanding (MOU) at POSCO DX’s Pangyo office for the on-site application of industrial humanoid robots.

Recently, demand for humanoid robots in China has surged, leading to a wave of large-scale supply contracts. In response, Morgan Stanley released its Humanoid Robots 100 report, projecting that the global humanoid robots market could reach as much as USD 60 trillion within the next decade. This article examines the potential of humanoid robots as an industrial mega trend, global technology development strategies, and Korea’s approach to this emerging sector.

Senior Researcher Jeoung-Heon Woo POSCO Research Institute

Humanoid Robots in History

Humanoid robots—machines designed to resemble humans—have appeared throughout history, sometimes as loyal assistants, other times as perceived threats. Examples include the bronze giant Talos from ancient Greek and Roman mythology, the water-clock-powered automaton of China’s Han Dynasty, and Leonardo da Vinci’s “robot knight” from the Renaissance. Across ancient civilizations, human-like machines have emerged in various forms.

Interest in humanoid robots has endured for centuries, accompanied by caution over potential risks. Notably, science fiction writer Isaac Asimov introduced the Three Laws of Robotics in his 1942 short story Runaround, raising philosophical questions about the relationship between humans and machines.

The rapid advancement of AI in recent years suggests that robotics may evolve toward a humanoid robots-centered future. While a “machine” is generally defined as a tool designed to perform production activities using power, a “robot” is an intelligent machine capable of making autonomous decisions under certain conditions.

Modern humanoid robots go further, combining advanced AI with a human-like form factor—a body structure modeled on human anatomy—allowing them to learn work methods, optimize performance, and actively assist in a wide range of human activities.

Physical AI: Extending Intelligence into the Real World

The concept of Physical AI is also gaining attention. Physical AI refers to AI embedded in physical devices—such as robots or autonomous vehicles—that interact directly with the real world. Traditional AI communicated with humans through digital interfaces like text or images, but Physical AI operates in real-world environments, collaborating with humans, perceiving surroundings, and responding accordingly. At CES 2025, NVIDIA CEO Jensen Huang identified Physical AI as a major future growth driver, emphasizing NVIDIA’s role at the center of this technological shift.

ⓒ gettyimagesbank

Why the World is Paying Attention to Humanoid Robots

Humanoid robots are gaining social acceptance and technological attention for two main reasons: form factor suitability and socio-economic potential.

1. Form Factor Perspective

Robots designed for specific repetitive tasks benefit from specialized form factors. However, for Physical AI performing diverse, non-specialized actions in real-world environments, a human-like form factor is advantageous because our physical infrastructure is built for human proportions.

Door handles, stair dimensions, and control panel placements are all designed for human use. Humanoid robots can operate in these environments without costly infrastructure changes, offering high versatility. In contrast, having different standards for each form factor would be inefficient.

2. Socio-Economic Perspective

Humanoid robots’ human-like appearance enables a wide range of human-robot collaboration scenarios, extending beyond manufacturing into customer service, caregiving, education, and guidance. Recent advances in language processing, facial expression recognition, and gesture control have improved emotional engagement, signaling the evolution of robots into social entities.

However, psychological barriers remain. Masahiro Mori, Professor Emeritus at Tokyo Institute of Technology, proposed the Uncanny Valley theory, which suggests that robots that look too human can cause discomfort. This highlights the need to consider psychological acceptance and emotional distance alongside technical perfection.

Economically, humanoid robots are highly promising. Their development requires not only AI but also sensors, actuators, motion control systems, and energy supply technologies. These demands drive innovation across multiple industries, making humanoid robots development a potential growth engine for the future.

Humanoid Robots Industry Structure and Potential Players

Morgan Stanley’s February 2025 report divides the humanoid robots industry value chain into three core areas: Brain, Body, and Integrator, and identifies potential players in each.

• Brain: Combines software and hardware. Software includes AI models, data science, simulation technology, and vision software. Hardware includes memory and vision computing.
• Body: Includes actuators, components, motors, sensors, batteries, power semiconductors, analog semiconductors, aluminum casting, connectors, heat treatment, and automation systems.
• Integrator: Companies that assemble and integrate the brain and body into finished products. Potential players include Hyundai Motor, Boston Dynamics, Apple, Samsung Electronics, LG Electronics, Alibaba, Amazon, Naver, ABB, and KUKA.

The industry can also be categorized into core components and modules, finished product assembly, and service areas, with the service sector expected to see diverse business models emerge.

Global Technology Leaders: Tesla and NVIDIA and Hyundai Motor

▲ Tesla’s Optimus Gen 2 (Source:Tesla’s official YouTube channel)

The emerging humanoid robots industry is being led by Tesla and NVIDIA, each pursuing distinct strategies.

Tesla is leveraging its expertise in EV production and autonomous driving to develop Optimus, a humanoid robot intended to automate production lines. First unveiled at Tesla AI Day in 2021, Optimus has evolved to perform a variety of tasks. The 2024 Optimus 2 features 40 actuators—12 in the hands alone—allowing it to perform delicate actions such as cracking an egg. Tesla plans to enter the humanoid robots sales market in 2026.

NVIDIA, on the other hand, aims to dominate the humanoid robots “Brain” platform rather than build its own robot. Its Jetson Thor computer, based on the latest Blackwell GPU architecture, enables large-scale AI inference and vision-based decision-making directly on local devices—capabilities previously limited to server environments.

Tesla’s approach resembles Apple’s integrated hardware-software model, while NVIDIA’s strategy is akin to Android’s platform dominance.

▲ Hyundai Motor Company unveils humanoid robot “Atlas” at CES 2026(Source:Hyundai Motor Group official YouTube channel)

In addition to these global leaders, Hyundai Motor introduced its humanoid robot Atlas at CES 2026. Purpose-built for industrial and logistics operations, Atlas offers advanced mobility, precise manipulation capabilities, and seamless integration with Hyundai’s autonomous vehicle and smart factory ecosystems. The debut underscores Hyundai’s ambition to position itself as a key integrator in the humanoid robots value chain, capitalizing on its manufacturing expertise, robotics R&D, and global production footprint.

Securing Leadership in the Humanoid Robots Value Chain

Following smartphones and EVs, the world has lacked a clear driver of technological innovation—until humanoid robots emerged as the next catalyst. As a convergence of cutting-edge technologies, humanoid robots are recognized as a key area for future growth, though challenges remain in cost competitiveness and safety in human-machine collaboration.

Importantly, the humanoid robots industry’s impact will extend beyond AI and software into materials, components, and services. Development and standardization will require core components that meet both functionality and reliability, along with mass production capabilities. In the service sector, opportunities will arise in humanoid robots deployment, human-robot collaboration models, and humanoid robots training and operation.

In June 2025, the Korean government launched the K-Humanoid Robots Alliance, a national robotics and AI consortium involving over 40 domestic industry, academic, and research institutions. AI companies and experts are collaborating with universities to develop AI models for robot manufacturers, with field trials supported by demand-side companies such as POSCO Group.For example, Aei Robot has signed MOUs with POSCO E&C and HD Hyundai Mipo Shipyard to develop humanoid robots for construction sites and shipyards.

Humanoid robots have moved beyond simple robotics to become a central axis of next-generation industrial innovation. Understanding and responding strategically to the technological, industrial, and service trends surrounding their evolution is more important than ever.

Leading the spread of AX in the industrial field with applications such as coil unloading systems for tens of tons of steel products

POSCO DX (President Shim Min-suk) is accelerating its manufacturing AX (AI Transformation) by advancing its ‘Physical AI’ technology based on virtual environment simulation and applying it to the field one after another.

POSCO DX announced that it has developed an AI model that replicates a real-world factory in a virtual environment and established a ‘physical AI’ development system that efficiently supports AI learning and verification prior to field application. Facilities and sensors at industrial sites are greatly affected by the physical environment such as inertia, acceleration, and noise, and by implementing conditions that can be encountered in the real world in a virtual space and conducting various simulations, AI can learn the optimal movement of facilities.

At actual industrial sites, there were limitations in acquiring field data and testing AI models due to disruption of operations or safety and security reasons. POSCO DX utilized virtual environment simulation to solve these issues and is expected to significantly reduce the time and cost of AI model development and field application.

POSCO DX has implemented virtual environment simulation by adopting Isaac Sim of NVIDIA Omniverse platform. Isaac Sim is a SW that supports verification by simulating the physical environment precisely with actual sensors and systems in a virtual environment.

POSCO DX recently built a specialized optical laboratory in its Pangyo office building to further advance its ‘physical AI’. The optical laboratory is a space where the response of sensors and precision according to specifications, which are difficult to identify in AI simulations, are tested by artificially creating an environment similar to the actual site in terms of lighting, temperature, and movement. By reflecting the sensor data verified here in the AI ​​simulation, it is evaluated that the gap between the virtual and actual sites has been minimized, enabling more efficient simulations.

On the other hand, ‘Physical AI’, which was the topic of this year’s CES 2025, refers to AI that recognizes the surrounding environment and interacts with the real world by controlling physical systems. ‘Physical AI’By applying , AI can recognize and judge various operation situations and control and operate facilities directly with the on-site control system (PLC, Programmable Logic Controller), enabling unmanned operation. POSCO DX is leading the spread of ‘Physical AI’ in industrial sites by converging AI technology with existing IT (information technology) and OT(automation technology).

POSCO DX ‘physical AI’ to cranes that transport atypical products with different sizes and packaging forms.is prioritizing the application of In the first half of this year, POSCO DX developed an AI model to automate the unloading of coil products transported on trailers into cranes, and plans to spread it horizontally by conducting virtual test drives and applying it to the field.

Yoon Il-yong, head of POSCO DX AI Technology Development Center, emphasized, “POSCO DX is pursuing the implementation of ‘physical AI’ technology with the goal of ‘autonomous physical systemization of manufacturing facilities’ based on the convergence of AI, IT, and OT technologies,” and added, “We will strive to establish ourselves as a leading company in intelligent factories by advancing ‘physical AI’ utilizing physical development methodologies such as virtual environment simulation.”

▲POSCO DX AI researchers are testing the physical environment, such as light and distance, of the coil loading and unloading site in the optical laboratory of the Pangyo building.