메인으로

<2026> Technology Development Status and Market Outlook for Humanoid Robots and Robot Batteries (~2040)

At CES 2026, humanoid robots emerged as a central pillar of the technology exhibition, with various new products combining AI and Physical AI showcasing the current state of industrial, service, and home robots. In particular, global companies such as Nvidia, LG, and Boston Dynamics demonstrated robots with real task execution capabilities and interactive functions, highlighting how AI-based control and perception technologies are accelerating robot commercialization.

Boston Dynamics unveiled its next-generation electric Atlas robot at CES 2026, targeting real-world deployment in manufacturing environments. Companies from Korea, China, and the United States competitively showcased models applicable to industrial, service, and entertainment sectors, indicating a shift from traditional demonstration-focused exhibitions toward real-environment testing and deployment.

In China, companies such as UBTECH, Unitree, and various startups presented advanced humanoid robots throughout the exhibition. Some demonstrated fully autonomous industrial tasks such as component sorting and handling, while others emphasized applications in factories, logistics, and services through interactive demos. Meanwhile, Korea and Japan also showcased not only robot platforms but also key components such as actuators, reducers, sensors, and batteries, highlighting their strategies to expand the ecosystem by strengthening core component competitiveness.

Meanwhile, Elon Musk stated that humanoid robots (Optimus) could become Tesla’s largest business in the long term, emphasizing that they could create far greater economic value than vehicles. He also announced plans to convert part of the Model S/X EV production lines into humanoid robot production lines, targeting annual production of 1 million units. Furthermore, he noted that Optimus would begin by replacing factory labor and gradually expand to substitute everyday human labor, with a goal of reducing the price to below $20,000 through mass production.

Recent humanoid robot development can be summarized into (1) the transition of hardware to mass production, (2) rapid advancement of intelligence (Embodied AI), and (3) the establishment of safety, reliability, and operational (SoS) systems for real-world deployment. In the past, the key question was whether robots could walk or not, but now the competition centers on how long they can operate, how cost-efficient they are, how safe they are, and how many units can be deployed. In particular, repetitive tasks in manufacturing and logistics have clear ROI calculations, making them the most promising initial commercialization markets.

Global research firms’ market outlooks for humanoid robots vary significantly depending on how the market is defined. Therefore, it is reasonable to divide them into (1) market size (revenue) forecasts and (2) shipment/deployment (unit volume) forecasts.

Goldman Sachs estimated the humanoid robot TAM (Total Addressable Market) at $38 billion by 2035. MarketsandMarkets projected the “Humanoid Robot Market” to grow from $2.92 billion in 2025 to $15.26 billion by 2030, while IDTechEx also suggested growth potential to around $30 billion by 2035. Meanwhile, UBS presented a range of $30–50 billion by 2035, and further outlined a broader ecosystem—including components, data, and services—expanding to $1.4–1.7 trillion by 2050.

According to SNE Research, alongside declining cell prices by battery chemistry (LFP, NCM, semi-solid, all-solid-state), the average cell ASP is expected to fall to $94/kWh by 2035 and around $70/kWh by 2040. As a result, the humanoid robot battery cell market is projected to exceed $10.5 billion by 2040.

From a shipment/deployment perspective, Omdia forecasts over 10,000 units by 2027 and approximately 38,000 units by 2030. Bank of America presents an aggressive long-term scenario with annual shipments reaching around 1 million units between 2030 and 2035. Morgan Stanley, instead of focusing on market revenue, approaches the topic from the perspective of labor and wages, estimating 8 million working humanoids in the U.S. by 2040 and an impact of $357 billion on wages.

Finally, a critical factor in robot deployment is the battery, which limits operating time. Currently, operating time is approximately 2–4 hours, but it can be significantly reduced depending on load and operating conditions, creating constraints on robot utilization.

Unlike EV batteries, humanoid robot batteries must simultaneously satisfy high power (peak output), frequent charge/discharge cycles, lightweight and compact design, and safety for close human interaction. Accordingly, global battery companies are recognizing robots as a new demand source and are moving toward strengthening high-discharge performance, safety design, swappable/modular packaging, and BMS algorithms.

LG Energy Solution and Samsung SDI can be seen extending their capabilities in high-power, high-reliability cells/modules for robots, along with safety evaluation know-how and BMS technologies, similar to their approach in UAM and drones. Chinese major players (CATL, EVE, etc.) are also clearly moving toward improving both energy density and safety through high-power product lineups and semi-solid/all-solid-state roadmaps for robots and drones. From a robot-focused perspective, companies like Grepow, which are strong in high-discharge packs for drones and robots, are expanding the market by offering integrated solutions including pack design, thermal management, connectors, and protection circuits.

From a battery form factor perspective, practical options can be summarized as (1) cylindrical-based modules (e.g., 2170/4680), (2) pouch-based customized packs, and (3) swappable cartridge (hot swap) packs. Cylindrical cells benefit from a large supply chain and strong cost/maturity advantages, making them suitable for mass production. Pouch cells offer greater design flexibility due to efficient space utilization in irregular structures such as the robot torso, back, and pelvis. Hot swap reduces downtime for 24-hour operations (factories/logistics), but increases system complexity due to connector durability, surge protection, and dual power system design requirements.

Application scenarios vary: in industrial settings, hot swap combined with fleet management (swap stations) is dominant; in service/retail, docking-based charging (semi-autonomous to autonomous) is more practical; and for home use, autonomous docking with fast top-up charging becomes important due to usability. Ultimately, humanoid battery competitiveness is likely to evolve into system-level competition encompassing not only cell chemistry but also pack structure, thermal management, charging/swapping UX, and BMS software.

This report is not limited to a single topic such as market outlook or application scenarios for humanoid robots. Instead, it comprehensively covers key technologies, core components such as actuators and batteries, operational software, and market outlooks, enabling a holistic understanding from technology to market in a single report.

Therefore, the strengths of this report are as follows:

① This report provides a deep-dive analysis of core technologies and key components, enabling insights into technical issues and future development outlooks of humanoid robots.

② It offers a more detailed description of the development history, current status, and information of global humanoid robot manufacturers, which have previously been relatively simplified, thereby greatly helping to understand the current landscape.

③ The market outlook is segmented into three areas—industrial/logistics, service/retail, and household—and utilizes highly reliable data based on UN organizations such as UNIDO and ILO, as well as major public sources, thereby enhancing the credibility of the analysis by closely reflecting real-world conditions.

④ For batteries, the outlook is categorized by chemistry types such as LFP, NCM, semi-solid, and all-solid-state, and also by form factors including cylindrical, pouch, prismatic, and structural/custom packs, allowing for more realistic and practical forecasting.

Contents