We have transitioned into the Acceleration Era of technological convergence, characterized by the interdependent evolution of artificial intelligence, quantum computing, advanced networking, robotics, biotechnology, advanced materials, and various other technologies. They are steadily strengthening one another.

Few technological advancements demonstrate that convergence more explicitly than the humanoid robot. For decades, humanoid robots were predominantly confined to the domains of speculative fiction, scholarly laboratories, and rigorously supervised displays. That is in the process of alteration. The technology is advancing into production sites, distribution centers, experimental settings, and, more frequently, additional tangible contexts. The relevance is not solely that robots possess the capability to bipedally traverse.

The emergence of artificial intelligence is equipping machines with the capacity to perceive, learn, reason, and adapt, as improvements in hardware enhance their physical capabilities. The consequence is the rise of what I refer to as embodied intelligence: AI that transcends mere information generation to engage with the tangible world.

Moving from Experimental Showcases to Occupational Settings

The shift is already evident. Boston Dynamics has discontinued its hydraulic Atlas and unveiled a fully electric model engineered for industrial purposes. The organization claims that the new Atlas is under development for actual material-handling contexts and has commenced customer field evaluations with Hyundai.

Amazon is conducting trials with Agility Robotics’ Digit, a bipedal mobile manipulator adept at movement, grasping, and managing items in warehouse environments. Amazon has identified tote handling as a foundational use case, highlighting a crucial feature of humanoids: their proficiency in functioning within environments already configured for human use without the need for extensive modifications to the infrastructure.

BMW has similarly assessed Figure’s Figure 02 humanoid at its Spartanburg manufacturing site. Throughout an experiment, the robot proficiently carried out a skillful operation involving the placement of sheet-metal components into fixtures employed in chassis assembly. These cases are important because they showcase a movement from robots tailored for limited industrial functions to robots adapted for the environments in which humans already engage.

The economic landscape is also commencing to transform. In 2024, Goldman Sachs Research anticipated that the global market for humanoid robots could approximate $38 billion by 2035, with annual shipments possibly reaching 1.4 million. The investigation further revealed that anticipated manufacturing expenditures had diminished by nearly 40%, now estimated between $30,000 and $150,000 for diverse levels of humanoid proficiency.

These statistics should be regarded as predictions rather than assurances, but they indicate the path forward. The focus is evolving from the question of whether humanoid robots can be created to their capacity for economic sustainability.

China is Advancing from Showcasing to Widespread Application.

The international humanoid lineage is not limited to the United States. In reality, a crucial evolution in the past two years has been the accelerated commercialization of humanoid robotics in China. Estimates within the industry differ markedly depending on whether they evaluate shipments, installations, production, or distinct types of humanoid robots. This holds importance due to the absence of a globally acknowledged approach for evaluating this developing market.

However, the path is clear. It is anticipated that the worldwide distribution of humanoid robots will amount to 13,318 units in 2025, marking a nearly fivefold rise from the preceding year. Chinese manufacturers topped the charts, as AgiBot dispatched approximately 5,168 units, Unitree about 4,200, and UBTech nearly 1,000.

Various research entities have yielded superior estimations. According to IDC, global shipments are anticipated to reach about 18,000 units by 2025, while Counterpoint Research forecasts around 16,000 installations, predominantly in China, which constitutes more than 80% of the total.

The variations are not as crucial as the fundamental trend: China has created a powerful manufacturing infrastructure focused on humanoid robotics. Enterprises such as Unitree, AgiBot, and UBTech are leveraging China’s substantial industrial foundation, extensive supply chains for electronics and batteries, established electric vehicle manufacturing expertise, and state support for robotics and embodied AI.

The rate of progress is also striking. The A2 Ultra model of AgiBot has been recognized by Guinness World Records for walking the longest distance by a humanoid robot—over 106 kilometers. The Tiangong humanoid from China has exhibited proficiency in climbing outdoor stairs and utilizing vision for navigation in rugged environments.

This possesses geopolitical consequences. Humanoid robotics is transforming from a simple automation field into a crucial manufacturing asset. The state that innovates the most effective amalgamation of AI models, actuators, batteries, sensors, semiconductor technology, robotics software, and manufacturing prowess could secure a significant advantage that surpasses factory automation.

Recent progress underscores this notion explicitly. In September 2026, Reuters indicated that China comprised about 95% of global humanoid shipments in 2025 and is investigating military applications spanning logistics, reconnaissance, and dangerous operations. Thus, the humanoid race is part of the larger conversation surrounding economic competitiveness, supply-chain durability, and national security.

Japan is Revitalizing its Edge in Humanoid Technology.

Japan presents a unique but comparably vital account in the realm of robotics. For decades, the country has been connected to humanoid robotics, exemplified by Honda’s ASIMO and SoftBank’s Pepper. At present, Japan is endeavoring to integrate its historical heritage with a new era of AI-enhanced robotics.

The Kyoto Humanoid Association, abbreviated as KyoHA, collaborates with institutions like Waseda University, Murata Manufacturing, tmsuk, and SRE Holdings to foster the development of indigenous humanoid robots. The goal is to reestablish a Japanese hardware ecosystem proficient in supporting the next wave of physical AI.

Japan is commencing to extend its reach beyond research facilities. In 2026, Japan Airlines, in collaboration with GMO AI & Robotics, commenced a demonstration initiative at Haneda Airport in Tokyo to assess humanoid robots for ground-handling duties, including tasks related to baggage and cargo.

This holds particular relevance due to Japan’s demographic and workforce strains that make automation economically appealing. At this moment, China could be at an advantage in terms of scale and manufacturing output, while the United States excels in artificial intelligence, software, and investment capital. Japan showcases unparalleled skill in precision manufacturing, robotic components, sensors, automotive engineering, and user-oriented robotics.

The developing competition is thus not merely concerned with which country can create the most extraordinary robot. It concerns the ecosystems that can incorporate the entire stack.

The Robot Intellect: AI Has Transformed the Formula

The foremost reason for the current viability of humanoids is not attributed to their legs. It constitutes the mind’s center. In my recent articles for Forbes titled “Tracking The AI Revolution” and “The Rapid Trajectory of Artificial Intelligence,” I analyzed the evolution of AI from standard machine learning to generative, agentic, neuromorphic, and eventually more intricately integrated forms of intelligence. That development is directly applicable to robotics. Please see:

Standard industrial robots were usually designed to carry out specific tasks in controlled environments. Humanoids demand a distinct approach. They are required to discern their visual stimuli, grasp instructions, foresee repercussions, manage novel objects, traverse dynamic settings, and recover from mishaps.

This is the juncture where vision-language-action models, commonly referred to as VLAs, become crucial. Google DeepMind’s RT-2 showcased that a vision-language model could be tailored to directly create robotic actions, permitting insights obtained from large-scale web training to enhance physical robotic behavior. Google indicated a more than threefold enhancement in particular generalization evaluations compared to earlier models.

NVIDIA has implemented a parallel methodology with Isaac GR00T, a multifunctional foundational model platform for humanoid robots that amalgamates robotic data, synthetic data, simulations, and AI models for generalized reasoning and skills in robotics.

This signifies a crucial change in understanding. A robot does not now need a program to generate a unique command for each imaginable circumstance. On the contrary, AI models can more and more deliver a framework of perception, reasoning, and action. This does not indicate that humanoids have reached the level of general intelligence found in humans. They have refrained from doing that.

The complexities of physical-world reasoning are still profoundly daunting, as reliability, energy usage, dexterity, safety, and expense present considerable hindrances. Yet, the course is clear. AI has revolutionized the robot from a tool that predominantly adheres to directives into a system that can progressively comprehend and adjust.

Materials Science Constitutes the Other Integral Component of the Revolution

Solely relying on artificial intelligence does not yield a practical humanoid. The tangible body is of comparable importance. Over the decades, robotics engineers have grappled with a core issue: motors, actuators, batteries, and structural materials have predominantly been heavier, less versatile, and less energy-efficient compared to biological alternatives.

This transformation is occurring as a result of advancements in materials science and actuator technology. Integrated sensing electric actuators, lightweight composite substances, advanced metallic alloys, 3D printing, flexible electronic components, and novel battery technologies are augmenting the proficiency of robotic platforms.

Scholars are additionally engineering artificial muscles founded on materials and systems crafted to replicate specific features of biological muscle. Studies on twisted and coiled artificial muscles have shown remarkable energy density, significant load-to-weight ratios, and extensive deformation abilities, yet considerable engineering challenges must be addressed before these systems can effectively replace conventional robotic actuators on a large scale.

Electronic skin embodies a significant advancement in exploration. Scientists are designing pliable, stretchable sensor technologies that can identify mechanical and environmental signals and utilize machine learning to analyze this information. The amalgamation of electronic skin technology and AI could potentially grant robots a significantly enhanced tactile perception, surpassing the current reliance on visual systems.

This is the juncture where the convergence becomes notably captivating. AI delivers intellectual prowess. Materials science offers physical proficiency. Instruments enable cognition. Cutting-edge computing enables immediate decision-making. Batteries supply stamina. Robotics amalgamates each of these components. The humanoid is thus not an individual technology. It represents a system of technology.

The Cybersecurity Challenge: A Robot’s Potential Exceeds Merely Data Exfiltration.

This integration engenders a completely unprecedented cybersecurity obstacle. It is judicious to investigate the security consequences of the growing interconnectivity of systems, such as IoT, smart cities, AI, and vital infrastructure. My worry about humanoids is uncomplicated: they are not just linked devices. They are linked devices proficient in performing physical actions. A hacked laptop can disclose data. A robot that has been compromised could potentially navigate, manipulate tools, breach restricted areas, disrupt manufacturing processes, or endanger human safety.

That modifies the risk assessment. Consequently, humanoid robots must be considered as significant cyber-physical systems. The security design ought to encompass identity management, zero-trust strategies, secure software development practices, software bills of materials, hardware-rooted security measures, encrypted communications, model integrity protections, behavioral monitoring, and secure update systems.

The AI itself additionally serves as a target for attacks. A rival might endeavor to taint training data, alter perception frameworks, take advantage of flaws in a foundational model, or utilize prompt injection on an AI-based robotic system.

The line separating a cyberattack from a physical attack might consequently grow more ambiguous. This bolsters an assertion I presented in my recent Forbes article “AI, Quantum And The New Cybersecurity Framework Imperative”: cybersecurity must shift from a perimeter-oriented model to one centered on continuous evaluation, resilience, adaptive risk management, and trust. For humanoids, security must not be appended after implementation. It should be incorporated into the structure from the initial stages. Please see:

Transitioning from Humanoid Robots to Biological Machines

The most intriguing advancements may eventually transpire beyond the current industrial humanoid. I perceive the progression unfolding along a continuum. The initial stage involves the industrial humanoid. These systems will concentrate on logistics, production, storage, construction, evaluation, upkeep, and perilous settings.

The subsequent stage will involve social and service-oriented humanoids, where progress in conversational AI, perception, dexterity, and human-machine interaction enable robots to engage more seamlessly in healthcare, elder care, hospitality, education, and domestic settings.

Beyond that exists a more profound realm: the bio-machine frontier. Investigators are currently examining biological substances, living organisms, synthetic muscles, biological detectors, neuromorphic architectures, and brain-computer interfaces. The aim is not explicitly to create a machine that genuinely transforms into a human. Instead, the objective is to amalgamate the advantages of biological and technological systems.

My latest article in Forbes, titled “The Emerging Computing Ecosystem: AI, Quantum, Biological, And Chemical,” explored the extensive convergence and the rise of computing architectures that transcend traditional silicon.

In the field of robotics, this convergence may ultimately yield machines equipped with biologically inspired sensors, flexible artificial muscles, self-repairing materials, and progressively advanced neural interfaces. Simultaneously, individuals will progressively be enhanced via exoskeletons, sophisticated prosthetics, wearable technology, and brain-computer interfaces. They would become cyborgs and the distinction between human and machine may consequently become less dichotomous.

Economic and Social Impacts

The emergence of humanoids ought not to be perceived merely as a narrative concerning robots supplanting employees. The more intricate inquiry pertains to the allocation of tasks between humans and machines. Significant prospects exist for employing humanoids in roles that are perilous, labor-intensive, monotonous, or challenging to fill. Construction, disaster management, mining, logistics, manufacturing, infrastructure assessment, and other perilous settings are clear contenders.

An increasingly elderly global demographic presents an additional compelling motivation. Goldman Sachs anticipates that humanoid robots may ultimately alleviate certain manufacturing labor deficits and the need for elderly care, highlighting the potential significance of demographics alongside technological prowess in influencing adoption.

However, economic upheaval is unavoidable. Entities must reevaluate employee training, job structuring, insurance policies, liability considerations, safety accreditation, and human-machine interaction. Authorities must contemplate regulations for autonomous physical systems akin to the frameworks established for aviation, medical apparatus, self-driving vehicles, and cybersecurity. The policy dilemma is to promote innovation while ensuring that technology does not surpass safety and regulatory measures.

The Humanoid Epoch Is a Narrative of Infrastructure

The significance of humanoid robotics lies not in the mere novelty of a bipedal machine. It embodies what the humanoid signifies. The humanoid is evolving into a nexus for artificial intelligence, sophisticated semiconductors, energy storage, sensors, materials science, computer vision, cloud and edge computing, telecommunications, simulation, and cybersecurity.

This renders the technology strategically important well beyond the realm of robotics. The nations and enterprises that thrive will not inherently be those possessing the most remarkable singular robot. They will be the individuals who can construct the most proficient ecosystem surrounding the robot.

China is showcasing the strength of its manufacturing capacity. The United States exhibits remarkable capabilities in artificial intelligence, software development, and advanced computing. Japan possesses profound proficiency in precision robotics and human-centric engineering. Europe, South Korea, Taiwan, and various other technological ecosystems possess essential components of the solution.

The contest has only just begun. The era of humanoids has commenced, yet it remains in its nascent stages. The machinery operating in factories today is rudimentary in comparison to what could develop in the forthcoming decade. As artificial intelligence advances, materials are enhanced to be lighter and more robust, battery technology progresses, sensors gain complexity, and robotic models assimilate knowledge from expansive datasets of the physical realm, humanoid robots may emerge as a pivotal technology of the Acceleration Era.

The primary inquiry has shifted from whether machines will integrate into human surroundings. They have already done so. The question pertains to what we instruct them in, how we safeguard them, how we regulate them, and how we guaranty that their growing physical intelligence ultimately enhances rather than constricts human potential. This will constitute a pivotal technological and policy challenge of the forthcoming decade.

  • Goldman Sachs Research — $38 billion humanoid-robot market and 1.4 million-unit 2035 forecast
  • Goldman Sachs Research — earlier labor-shortage and elderly-care projections
  • Omdia shipment data — 13,318 humanoids shipped globally in 2025
  • IDC — approximately 18,000 global humanoid shipments in 2025
  • Counterpoint Research — approximately 16,000 humanoid installations in 2025
  • Boston Dynamics — Electric Atlas
  • Amazon — Digit and warehouse robotics
  • BMW — Figure 02 manufacturing trial
  • Google DeepMind — RT-2 vision-language-action model
  • NVIDIA — Isaac GR00T humanoid foundation model
  • Kyoto Humanoid Association / Murata — Japanese humanoid initiative
  • Japan Airlines — Haneda Airport humanoid-robot trial