The technology, ecology, and economics of humanoid robot batteries

Since the beginning of this year, humanoid robots have hardly taken a break, with videos of them dancing, running marathons, performing yangko dances, and boxing continuously flooding the internet. On the other hand, related companies have experienced a rare peak in financing, with valuations exceeding tens of billions of yuan, and many well-established internet giants have become investors behind the scenes.

Most opinions suggest that 2025 will be the inaugural year for the mass production of humanoid robots, with the global humanoid robot market expected to reach 6.339 billion yuan. Looking to the future, the market size for humanoid robots is anticipated to exhibit exponential growth. Goldman Sachs predicts that by 2035, the global humanoid robot market will reach 38 billion US dollars.

The current growth is mainly attributed to the urgent demand for automation and intelligent upgrades in industries such as manufacturing and logistics warehousing. For example, Tesla plans to deploy thousands of Optimus robots in its factories to perform tasks like welding and handling. Meanwhile, a larger market potential will come from applications in home services and healthcare and wellness scenarios, which together might account for 45%.

Not limited to Tesla, domestic companies like Zeekr, Midea, BAIC, BYD, and Hengtong are increasingly exploring how to introduce humanoid robots into production lines. However, transitioning from social media to real-world applications, where they can work continuously to replace humans in repetitive and high-cost labor tasks, humanoid robots still have a long way to go.

Among them, as the "source of power," batteries play a crucial role in the commercialization journey of humanoid robots, with the energy system—especially battery technology—being of paramount importance. Batteries are not only the "heart" of robots, providing them with power, but also represent the "Achilles’ heel" that determines their economic viability, practicality, and application boundaries.

Faced with the emerging market of humanoid robots, battery manufacturers are shifting from early observation to active strategic deployment. Leading domestic power battery companies such as CATL, Sunwoda, Gotion High-Tech, EVE Energy, and SVOLT have all announced the inclusion of humanoid robots in their battery technology plans.

The Dual Role of the Battery's "Economic Account"

From the perspective of hardware manufacturing, the proportion of power batteries in humanoid robots is not high. According to the cost estimation of Tesla's Optimus robot by Minsheng Securities, its core hardware costs account for about 69% of the total cost. In the BOM cost of Tesla Optimus, the battery cost is only 2,180 yuan, accounting for 0.5% of the total cost.

However, this proportion is not the only one; the industry generally believes that the cost of the battery in humanoid robots is around 1%, which is lower than that of core motion components such as reducers, servo motors, and controllers.

Despite this, the demand potential for power batteries remains immense. It is estimated that by 2030, the global shipment of humanoid robots will reach 5 million units. Based on a battery capacity of 2.5 kWh per unit, there will be a demand for at least 12.5 GWh of batteries, equivalent to 250,000 pure electric vehicles with a range of 500 km.

This is an unmissable market opportunity for battery manufacturers. Sunwoda is a leading global lithium-ion battery enterprise, accurately capturing the trends in consumer batteries, power batteries, and energy storage batteries. In terms of shipment growth rate from 2023 to 2024, Sunwoda is the fastest-growing company among the world's top ten power battery manufacturers and the world's top ten energy storage battery manufacturers.

At the same time, the importance of power batteries is also reflected in the fact that their performance directly determines the endurance, charging frequency, and overall work efficiency of robots. Their technical selection and performance parameters have a profound and decisive impact on subsequent operational costs.

Therefore, even though the BOM cost proportion is low, companies still need to consider long-term economic factors when choosing a battery solution, rather than just focusing on the initial purchase price. Operating costs not only include energy consumption but also encompass downtime due to charging or battery swapping, maintenance labor costs, and the depreciation and replacement costs of the battery itself.

Inefficient charge and discharge cycles directly translate into high operational costs and low investment returns. Therefore, when evaluating the economics of humanoid robots, it is more reasonable to use the total lifecycle cost as the measure.

From this perspective, the performance indicators of batteries, such as energy density, charging speed, cycle life, and safety, directly determine the effective working duration, maintenance frequency, and energy efficiency of robots.

The mainstream power batteries currently can only operate for 3 to 6 hours, and their heavy weight further restricts their flexibility. "It's like a tourist whose phone runs out of battery and leaves disappointed; the potential of robots is similarly constrained by their energy core," said Zhou Shuangjun, a small power battery technology expert at Sunwoda.

Limited battery life not only restricts the continuous operational capability of robots but also significantly increases their operating costs by necessitating more frequent recharges and downtime, thus becoming a key obstacle to their large-scale commercial deployment.

In this year's humanoid robot half marathon, Tiangong Ultra won the first place. However, behind this victory, Tiangong Ultra changed its battery three times in order to complete the whole course.

It is obvious that the endurance capability of humanoid robots is still far from keeping pace with the rhythm of commercialization, which directly weakens the cost advantage of humanoid robots replacing human labor and extends their investment payback period. At this stage, the trade-off is to innovate through operational models (such as self-recharging functions) to compensate for the endurance shortfall.

Pentagon Warrior Approaching the Limit

Ultimately, improving the performance of the battery itself is the fundamental solution to the problem of robot endurance.

According to relevant data, humanoid robots currently primarily use ternary cylindrical lithium batteries, which account for approximately 70% of the market. The energy density of the cells is between 250-300Wh/Kg, and the voltage platform is mostly concentrated between 48-58V. Some manufacturers also choose lithium iron phosphate batteries and are piloting semi-solid/solid-state batteries.

Leading lithium battery manufacturer Sunwoda, after conducting research with multiple humanoid robot manufacturers, has learned that the endurance target for humanoid robots is 8 hours, which still has some way to go at the current stage.

In response to this demand, the technical team at Sunwoda is working on enhancing the energy density of battery cells through innovations in cell materials (such as high-nickel high-silicon chemical systems) and process innovations, with the initial target to increase it to 350Wh/kg, aiming to solve the endurance challenge.

The next step is to hope for the maturation of semi-solid-state and solid-state battery technologies. The theoretical energy density of solid-state batteries can reach over 500Wh/kg, which is 2 to 3 times that of traditional liquid lithium batteries, significantly extending the working time of robots.

Sunwoda began its solid-state battery development in 2015 and now has 10 years of R&D experience. Currently, it has achieved small-batch production at 320Wh/kg and 360Wh/kg, and has completed installation and flight tests on unmanned aerial vehicles at the hundred-kilogram level.

"Humanoid robot batteries need to find a balance between energy density, cost, and safety," said Ouyang Minggao, an academician of the Chinese Academy of Sciences. "2025 will be a technological watershed. If semi-solid-state batteries can reduce costs to below $150/kWh, they are expected to be the first to open up the market."

It is worth mentioning that energy density is not the only requirement for battery manufacturers of humanoid robots; power output is also a crucial aspect.

Based on the performance of the humanoid robot half marathon competition in Beijing Yizhuang, the 48V voltage platform is insufficient to meet the demands for kinetic energy output, heat dissipation efficiency, and energy consumption control under long-term high load conditions. The battery needs to be upgraded to a higher voltage and stronger power to accommodate the power demands in complex scenarios.

In addition, during daily operations, robots may perform complex actions such as carrying, rapid movement, and dancing, which demand higher power. Manufacturers have found in practice that the transient power of batteries is difficult to meet, and there are significant fluctuations in kinetic energy throughout the full discharge cycle.

When fully charged, the voltage is high and the robot's power is strong. As the battery level decreases, the kinetic energy also decreases, affecting the precision and continuity of movements. It is necessary to achieve a stable 3C level of continuous discharge capability.

At the same time, the pursuit of high C-rate poses higher demands on the cycle life of battery cells. Currently, high C-rate battery cells on the market generally last only about 200 cycles, becoming a bottleneck for the lifespan of humanoid robots. Most manufacturers hope to increase this to over 600 cycles while ensuring system safety.

It is understood that Sunwoda's adopted strategy is to use an all-tab design to reduce the internal resistance of the cell, enhancing transient power output capability and kinetic stability. Additionally, Sunwoda has developed a BMS (Battery Management System) that integrates EIS (Electrochemical Impedance Spectroscopy monitoring), end-cloud collaborative control, SOX algorithms, and AI safety warning, making the battery smarter and safer.

On the other hand, at the current stage, the batteries used in humanoid robots generally have the characteristics of poor cell consistency and weak adaptability to irregular structures.

Some humanoid robot manufacturers have reported that the battery performance significantly declines after multiple uses, with severe capacity degradation. Additionally, due to the limitations of the battery cell size, the battery system is difficult to adapt to the irregular structure of the robot's body and cannot meet the spatial needs of different parts.

In this regard, Sunwoda innovatively adopts a hybrid architecture of "main trunk battery + joint micro-battery", combined with lightweight structural materials, to meet the adaptability of irregularly shaped bodies while reducing the overall weight of the battery.

The requirements for batteries in humanoid robots far exceed those in traditional consumer electronics or electric vehicles. They need to achieve an extreme balance in energy density, power density, safety, environmental adaptability, and cost, akin to a "pentagon warrior." These nearly stringent comprehensive performance requirements are driving battery technology towards new physical and chemical limits.

Xinwangda is fully aware of the demand differences across various industries. According to a representative from Xinwangda, "We provide a fully customized battery solution that precisely matches the unique structures of humanoid robots; leveraging our accumulated expertise in materials, structure, and intelligent management, we primarily focus on industrial and high-frequency service scenarios. These scenarios demand higher battery life and power stability, which better highlight our technical advantages. Additionally, we collaborate with industry partners to participate in the formulation of industry standards and related projects, accelerating the commercial scale-up of humanoid robots."