In July 2025, UBTech Robotics, the Shenzhen-based manufacturer known for its humanoid and service robots, released a video demonstrating a capability that has long been anticipated in the robotics industry: a bipedal humanoid robot that can change its own battery without any human assistance. The robot in question, the Walker S2, is shown ambling over to a charging station, using its own robotic arms to remove a depleted battery pack from its upper back, inserting that pack into a charger, and then retrieving a fully charged replacement unit to slot back into place. Once the swap is complete, the robot heads off, presumably to resume whatever task it had been performing.
The significance of this demonstration, as reported by Forbes contributor Leslie Katz, lies in what it represents for the broader trajectory of autonomous machines. UBTech has framed the Walker S2’s ability to perform continuous operations across dynamic industrial scenarios as one of its key advantages. The company’s claim is not merely about battery life, but about the elimination of a fundamental bottleneck: the need for human intervention when a robot’s power runs low.
According to the source material, the Walker S2 can detect when its battery is low, navigate autonomously to a charging station, and complete the entire battery replacement procedure in less than three minutes. The robot’s battery provides roughly four hours of operational runtime per charge. When that charge is depleted, the robot does not wait for a technician to arrive with a fresh pack. Instead, it manages the process itself, from detection to navigation to physical manipulation of the battery pack.
This is not a theoretical capability described in a press release. The video evidence shows the robot physically performing the task, using its arms to handle the battery pack with enough precision to remove it from its own body, place it in a charger, and install a replacement. The fact that the robot can manipulate a component of its own anatomy—reaching behind its own back, so to speak—adds a layer of complexity that goes beyond simple battery management systems found in conventional industrial robots.
It is worth noting that UBTech is not alone in pursuing this kind of capability. The source material also references Boston Dynamics’ Atlas humanoid robot, which operates on dual battery packs providing approximately four hours of runtime. When power runs low, Atlas autonomously navigates to a charging station, replaces its packs in three minutes, and returns to work. Boston Dynamics has stated that recharging takes 90 minutes, meaning Atlas can operate around the clock with minimal downtime. The company has said its Atlas production is “fully committed” for 2026, with plans to eventually build 30,000 units per year.
Agility Robotics is also moving in a similar direction with its Digit v5 humanoid, which is set to launch in December 2025. Digit v5 emphasizes safe, cooperative work alongside humans without traditional safety fencing, and offers 20-hour operational days. Agility has indicated that Digit will eventually be able to swap its own end-effectors—the robot’s hands or grippers—though the source material does not specify a timeline for that capability.
The Walker S2’s self-battery-swapping feature, however, appears to be a concrete demonstration of a capability that other manufacturers are still developing or have only announced as a future plan. Whether UBTech’s implementation is more advanced than its competitors’ is not something the source material directly addresses, but the video evidence places UBTech among the first to show a humanoid robot performing this task autonomously.
Why it matters for European robot service
For European businesses and service providers, the implications of self-swapping batteries in humanoid robots extend far beyond the novelty of watching a machine tend to its own power needs. The ability to operate continuously without human intervention addresses one of the most persistent operational challenges in industrial automation: downtime.
In traditional manufacturing and logistics environments, robots—whether fixed industrial arms or mobile platforms—require scheduled maintenance, recharging, or battery replacement. Each of these interruptions requires human staff to be available, which in turn imposes shift patterns, scheduling constraints, and labor costs. A robot that can manage its own energy needs effectively removes the human from that loop, enabling truly unattended operations.
The source material highlights that the Walker S2 can operate around the clock in dynamic industrial scenarios. For European operators, this could mean rethinking how production lines and warehouse operations are designed. If a robot can work through the night without a human present to swap batteries, then the economics of shift work change. Facilities could potentially run longer hours without corresponding increases in labor costs, or they could deploy robots in environments where human presence is undesirable or unsafe.
There is also a service dimension to consider. Robot service providers in Europe—those who install, maintain, and support robotic systems—will need to adapt their offerings. If robots can self-serve their energy needs, then the frequency of human intervention drops. However, this does not mean service requirements disappear entirely. Batteries still degrade over time, charging infrastructure still needs maintenance, and software updates still need to be applied. What changes is the nature of the service call: instead of routine battery swaps, service providers will focus on predictive maintenance, system optimization, and handling exceptions that the robot cannot manage on its own.
The source material notes that the Walker S2 has a four-hour battery life. This is a critical detail for European operators to consider. Four hours is not an entire shift, nor is it a full day. The robot’s ability to swap its own battery in under three minutes means that, in theory, it could operate indefinitely as long as there is a charging station with a fresh battery available. But this raises logistical questions: How many spare batteries must a facility keep on hand? How many charging stations are needed to support a fleet of robots? What happens if the robot’s navigation system fails and it cannot find the charging station?
These are the kinds of practical considerations that European buyers and operators will need to address. The technology is promising, but its deployment requires careful planning around infrastructure, battery inventory, and fail-safes.
Another point worth considering is the competitive landscape. The source material references Boston Dynamics’ Atlas and Agility Robotics’ Digit v5, both of which are pursuing similar capabilities. For European companies, this means there is not a single vendor to watch but a field of competitors, each with different strengths and timelines. UBTech’s Walker S2 appears to be demonstrating self-battery-swapping now, while Boston Dynamics has announced production plans for 2026, and Agility is launching Digit v5 in December 2025 with self-swapping end-effectors as a future capability.
For European buyers, the choice of which humanoid robot to deploy will depend on factors beyond battery swapping. The source material mentions that Atlas is rated at IP67, meaning it is protected against dust ingress and temporary immersion in water, and can operate in temperatures ranging from -20 to 40 degrees Celsius. These environmental specifications are relevant for European facilities that may operate in cold warehouses or outdoor environments. The Walker S2’s environmental ratings are not disclosed in the source material, so it is unclear how it would perform under similar conditions.
The source material also notes that Agility Robotics has chosen bulk material handling as its initial focus, citing the acute labor gap across the US and Europe. This suggests that the first wave of humanoid deployments in Europe may target dull, dirty, and dangerous tasks—moving materials, loading and unloading, and other repetitive activities where labor shortages are most acute.
What buyers and operators should know
For those considering the adoption of humanoid robots with self-battery-swapping capabilities, the source material provides several concrete data points that should inform decision-making.
First, the Walker S2’s battery life is approximately four hours. This is a relatively short window compared to a standard eight-hour shift, which means the robot will need to perform battery swaps multiple times per day if it is to operate continuously. The swap itself takes less than three minutes, which is remarkably fast, but the frequency of swaps means that a facility must have a reliable supply of charged batteries and a charging station that can recharge depleted packs while the robot continues working.
Second, the robot must be able to detect its own low battery state and navigate to the charging station. This implies that the facility must be mapped and that the robot’s navigation system must be reliable. If the robot cannot find the charging station, or if the station is occupied by another robot, the entire workflow could be disrupted. The source material does not disclose how the Walker S2 handles multiple robots competing for a single charging station, nor does it specify how many charging stations are recommended per robot.
Third, the physical act of swapping the battery requires the robot to use its arms to remove a pack from its own upper back. This is a complex manipulation task that requires precision and force control. The source material does not specify the weight of the battery pack, nor does it indicate what happens if the robot drops a battery or fails to insert it correctly. These are edge cases that buyers should probe with the manufacturer before committing to a deployment.
Fourth, the source material notes that Boston Dynamics’ Atlas operates on dual battery packs and takes 90 minutes to recharge. This is a different approach from UBTech’s, which appears to use a single swappable pack. The trade-off is worth understanding: a dual-pack system may provide redundancy (if one pack fails, the other may still provide power), but it also means more complex battery management. A single-pack system like the Walker S2’s may be simpler but offers no redundancy if the pack fails mid-operation.
Fifth, the source material does not disclose pricing, delivery timelines, or service agreements for the Walker S2. Buyers should not assume that these details are public or that they can be obtained from the source material. What is known is that UBTech is a Shenzhen-based manufacturer with a track record of producing humanoid and service robots. The company’s ability to support European deployments—including spare parts availability, technical support, and software updates—is not addressed in the source material and should be verified directly with the manufacturer.
Sixth, environmental specifications matter. The source material provides detailed environmental ratings for Boston Dynamics’ Atlas (IP67, -20 to 40 degrees Celsius) but does not provide equivalent specifications for the Walker S2. European operators in cold climates or dusty environments should ask UBTech for these details before making a purchase decision.
Seventh, the source material references Agility Robotics’ Digit v5, which is set to launch in December 2025. Digit v5 emphasizes safe, cooperative work alongside humans without safety fencing, and offers 20-hour operational days. This is a significantly longer operational window than the Walker S2’s four-hour battery life, even accounting for battery swaps. However, Digit v5’s self-swapping capability applies to end-effectors, not batteries, at least for now. The source material does not indicate whether Digit v5 can swap its own batteries.
Eighth, the source material mentions that Boston Dynamics plans to build 30,000 Atlas units per year, with production “fully committed” for 2026. This suggests that Atlas may be available in larger volumes than the Walker S2, but it also implies that Boston Dynamics is scaling up its manufacturing capacity. For European buyers, this could mean shorter lead times for Atlas compared to other humanoid robots, but the source material does not provide specific lead time information.
Finally, it is important to recognize what the source material does not say. There is no information about the Walker S2’s price, its payload capacity, its software ecosystem, or its safety certifications. There is no mention of how UBTech handles software updates, cybersecurity, or data privacy. There is no discussion of the robot’s failure modes or what happens when a battery swap goes wrong. Buyers should treat the self-battery-swapping capability as one feature among many, and they should conduct thorough due diligence before committing to a deployment.
The broader trend is clear: humanoid robots are moving toward greater autonomy, and self-battery-swapping is a significant milestone on that path. For European operators, the question is not whether this technology will arrive, but how quickly it can be integrated into existing workflows and what infrastructure investments are required to support it. The source material provides a glimpse of what is possible, but the details that matter for procurement—pricing, support, specifications—remain to be disclosed.
Published by Vigla Media OÜ (Estonia).