When evaluating inspection robots for industrial environments, the conversation has moved well beyond simple telepresence or camera-on-wheels solutions. The current generation of legged robots is being built for specific, high-stakes tasks — and gas leak detection has emerged as one of the most compelling use cases. The recent update from ANYbotics, a Swiss manufacturer, illustrates exactly where this technology is heading and what end users should be paying attention to.
The first thing to understand is the physical platform itself. ANYmal is a quadruped — a four-legged robot — designed from the ground up for heavy-duty industrial inspection work. It is not a consumer toy or a research curiosity. The machine weighs roughly 30 to 50 kilograms depending on the specific model configuration. That weight range places it firmly in the category of serious, mission-capable equipment rather than lightweight drones or handheld devices. The chassis is protected by an IP67-rated casing, which means it is sealed against dust ingress and can withstand temporary submersion in water. For industrial sites where dust, moisture, and debris are constant companions, this level of protection is not a luxury — it is a baseline requirement.
The sensor payload is where ANYmal distinguishes itself from simpler inspection tools. The robot carries a 20× zoom optical camera, which allows operators to get a close look at equipment, gauges, and infrastructure from a safe distance. It also includes a thermal imager capable of reading temperatures in the range of -40 to 550 degrees Celsius. That thermal range is significant because it covers both cryogenic applications and high-heat industrial processes. The robot is equipped with 360-degree LiDAR for mapping and navigation, ultrasonic microphones for detecting acoustic anomalies, and — critically for this latest update — gas detectors. The combination of these sensors means that a single pass through a facility can yield visual, thermal, acoustic, and chemical data simultaneously.
Battery life is another practical consideration. ANYmal can run for approximately 90 minutes on a single charge. When the battery runs low, the robot can auto-dock to recharge. This is not a trivial feature — in large facilities, manually retrieving and recharging a robot can consume significant operator time. Auto-docking enables longer autonomous missions with minimal human intervention.
Speed is worth noting as a trade-off. ANYmal walks at a normal pace of roughly 0.75 meters per second. That is not fast — a human walking briskly can easily outpace it. But the value proposition is not speed; it is persistence, sensor coverage, and the ability to operate in environments where humans should not be. The robot can climb stairs, navigate uneven terrain, and access areas that wheeled robots cannot reach. For facility managers, the question is not whether the robot can keep up with a walking pace, but whether it can perform inspections that would otherwise require shutting down equipment, donning protective gear, or risking human safety.
Modularity is another key attribute to look for. ANYmal has a payload capacity of around 10 kilograms for additional sensors or tools. This means the base platform can be adapted to different inspection tasks by swapping out modules. The gas detection capability, for example, comes in the form of a specialized "sniffer" module. This module uses acoustic imaging and concentration sensors to pinpoint gas leaks by both sound and chemical detection. The acoustic imaging component is particularly interesting — it allows the robot to localize a leak by the ultrasonic noise it generates, while the concentration sensors confirm the presence and magnitude of the gas. This dual-mode detection reduces false positives and helps operators prioritize repair work.
Another available add-on is the Leica BLK ARC sensor, which enables high-precision 3D scanning. This module can create digital twins of facilities with millimeter accuracy. For industries that need accurate as-built documentation, this capability transforms the inspection robot from a simple data collector into a documentation tool that can support engineering analysis, regulatory compliance, and long-term asset management.
The gas leak detection capability is not merely an incremental feature — it positions ANYmal as a preferred choice for mission-critical inspections in harsh environments. The manufacturer explicitly positions the robot as a complement to other platforms like Boston Dynamics' Spot. While Spot has gained attention for its agility and even its ability to dance or perform backflips, ANYmal focuses on a different set of priorities: ruggedness, sensor density, and the ability to operate in explosive or otherwise dangerous atmospheres. In particular, the ANYmal X variant includes an explosion-proof feature, which gives it a capability that Spot does not currently offer. For industries such as oil and gas, chemical processing, and utilities, explosion-proof certification is not optional — it is a legal and safety requirement for entering certain zones.
The practical implication is straightforward: if you have a gas leak in a danger zone at 3 AM, you want a robot that can go in, identify the source, and report back without risking human life. That is the use case ANYmal is designed to serve. The robot does not need to be entertaining; it needs to be reliable, precise, and survivable.
Practical steps
If you are considering deploying an inspection robot with gas leak detection capabilities, the decision process should be methodical. The following steps are grounded in what is known about the ANYmal platform and the broader category of industrial inspection robots.
First, define the inspection scenarios you need to cover. Walk through your facility and identify the areas that are hazardous, difficult to access, or require frequent monitoring. Gas leak detection is most valuable in environments where leaks can occur in confined spaces, at height, or in areas that require personal protective equipment to enter. Make a list of these zones and prioritize them by risk level and inspection frequency.
Second, assess your existing sensor infrastructure. Many facilities already have fixed gas detectors installed. The value of a mobile robot is not necessarily to replace those fixed sensors, but to complement them. A robot can investigate an alarm, search for leaks in areas without fixed coverage, and provide a spatial map of gas concentrations rather than a single point reading. Think of the robot as a mobile supplement to your fixed sensor network, not a replacement for it.
Third, evaluate the physical environment against the robot's specifications. ANYmal's IP67 rating means it can handle dust and water exposure, but you should still verify that your facility's conditions are within the robot's operational envelope. Consider floor surfaces, stairways, door widths, and elevation changes. The robot's walking speed of approximately 0.75 meters per second means you should estimate mission durations accordingly. If a facility tour takes a human inspector two hours, the robot will likely take longer — but it can operate autonomously and does not require breaks.
Fourth, plan for the sensor payload. The gas sniffer module uses acoustic imaging and concentration sensors to detect leaks. This means the robot can detect both the sound signature of a leak and the chemical presence of the gas. When planning missions, consider whether you need the gas detection module, the 3D scanning module, or both. The 10-kilogram payload capacity allows for some flexibility, but you cannot carry everything at once. Prioritize based on your immediate inspection needs.
Fifth, think about data integration. The robot collects thermal images, optical images, LiDAR scans, acoustic data, and gas concentration readings. This data is only valuable if it can be acted upon. Determine how the data will be exported, stored, and analyzed. Will it feed into your existing computerized maintenance management system (CMMS)? Can it be used to generate work orders automatically? The Leica BLK ARC module, for example, creates digital twins with millimeter accuracy — this data can be used for engineering analysis, but only if your team has the tools and skills to process it.
Sixth, consider the operational model. ANYmal can run for about 90 minutes per charge and auto-dock to recharge. This means the robot can be deployed for multiple missions per day with minimal human intervention. However, you should still designate a responsible operator who can monitor missions, review data, and handle exceptions. The robot is autonomous, but it is not unsupervised — someone needs to be accountable for its actions and its findings.
Seventh, look at the ecosystem. ANYbotics positions ANYmal as a complement to other robots like Spot. If you already have a Spot deployment, consider how the two platforms can work together. Spot may be better suited for certain tasks, while ANYmal's explosion-proof variant may be the only option for hazardous zones. A multi-robot strategy can cover more ground and provide redundancy.
Eighth, plan for the human side of adoption. Equinor, an early adopter of autonomous robotics, deploys "dogs" and drones for tasks that are dull, dirty, distant, and dangerous. This framing is useful for internal communication. When introducing an inspection robot to your workforce, emphasize that the robot is taking on the dull, dirty, distant, and dangerous tasks — not replacing skilled workers. The robot enhances worker safety by reducing exposure to hazards, and it improves maintenance outcomes by providing more frequent and more detailed inspection data.
Ninth, consider the integration with mobile worker platforms. The source material mentions Comos Mobile Worker as an integral part of the ANYmal ecosystem for plant inspections. This suggests that the robot's data is not meant to be siloed — it is meant to flow into a broader maintenance and operations workflow. When evaluating the robot, ask about software integration options and how the data will connect to your existing systems.
Tenth, budget for the total cost of ownership. The source material does not disclose pricing, so you should request a quote from the manufacturer. However, you should factor in not just the purchase price, but also the cost of additional modules, maintenance, training, and software subscriptions. The 90-minute battery life means you may want spare batteries for continuous operations. The modular payload system means you may need to purchase multiple modules to cover different inspection scenarios.
Finally, pilot before you commit. If possible, arrange a demonstration or a short-term rental to test the robot in your facility. This will allow you to verify that the robot can navigate your environment, that the gas detection module can identify leaks at the sensitivity levels you need, and that your team can operate the system effectively. A pilot also gives you the data you need to build a business case for a full deployment.
Common mistakes to avoid
The first mistake is treating the robot as a toy or a novelty. ANYmal is a serious piece of industrial equipment, and it should be evaluated with the same rigor as any other capital investment. Do not be swayed by videos of robots performing tricks — focus on the sensor payload, the environmental ratings, and the specific inspection capabilities.
The second mistake is ignoring the explosion-proof requirement. If your facility has hazardous areas, you need a robot that is certified for those zones. The ANYmal X's explosion-proof feature is a significant differentiator, and it is one reason the competition between ANYmal and Spot has become more nuanced. Do not assume that a standard industrial robot can enter an explosive atmosphere. Verify the certifications before deployment.
The third mistake is underestimating the importance of modularity. The gas sniffer module, the Leica BLK ARC scanner, and other payloads are what make the robot versatile. If you buy a robot without a clear plan for which modules you need, you will end up with a platform that cannot perform the tasks you actually need. Think about your inspection scenarios first, then choose the modules accordingly.
The fourth mistake is neglecting data management. A robot that collects thermal images, LiDAR scans, acoustic data, and gas readings is generating a large volume of data. If you do not have a plan for storing, analyzing, and acting on that data, the robot's value is significantly diminished. The digital twin capability, for example, is only useful if you have the software and expertise to work with 3D models.
The fifth mistake is expecting the robot to match human walking speed. At approximately 0.75 meters per second, ANYmal is slower than a typical human inspector. This is not a defect — it is a design choice that prioritizes stability, sensor accuracy, and battery life. Plan your missions accordingly and do not expect the robot to complete a facility tour in the same time as a human.
The sixth mistake is ignoring the complementary nature of different robot platforms. The source material explicitly notes that ANYmal complements Spot's territory by going into even harsher domains. If you are already invested in one platform, do not view a second platform as a threat — view it as an addition to your toolkit. Different robots have different strengths, and a multi-platform strategy can provide better coverage.
The seventh mistake is failing to involve the maintenance team in the decision process. The robot is ultimately a maintenance tool, and the people who will use it and benefit from it should have a voice in the selection and deployment process. Equinor's approach — using robots for dull, dirty, distant, and dangerous tasks — is a framing that resonates with maintenance workers because it directly addresses their safety and quality of work.
The eighth mistake is assuming that autonomy means zero supervision. ANYmal can auto-dock and recharge, and it can conduct inspections autonomously, but it still requires human oversight. Someone needs to review the data, respond to alarms, and intervene when the robot encounters an unexpected situation. Plan for a human operator or supervisor as part of your operational model.
The ninth mistake is neglecting to verify the specifics of the gas detection capability. The source material indicates that the sniffer module uses acoustic imaging and concentration sensors. This is a dual-mode approach that is more robust than a single-sensor solution. However, you should still ask about detection limits, calibration requirements, and response times for different gas types. The source material does not disclose these details, so you should request them from the manufacturer before making a purchase decision.
The tenth mistake is making decisions based on unverified claims. The source material provides specific specifications — weight, IP rating, battery life, payload capacity, sensor types, and walking speed — but it does not disclose pricing, delivery times, or service-level agreements. Do not assume that these details are the same as those of competing products. Request written specifications and verify them against your requirements.
Finally, do not forget the human element. The goal of deploying an inspection robot is to enhance worker safety and improve maintenance outcomes. If the robot is perceived as a threat to jobs, adoption will fail. Communicate clearly that the robot is taking on the dull, dirty, distant, and dangerous tasks — the tasks that no one wants to do and that put workers at risk. The robot is a tool that makes the workforce safer and more effective, not a replacement for skilled workers.
In summary, the ANYmal gas leak detection capability represents a significant step forward for autonomous industrial inspection. The platform's rugged design, modular payload system, and explosion-proof variant make it a compelling option for mission-critical applications. By following a methodical evaluation process and avoiding common pitfalls, you can determine whether this technology is right for your facility — and if so, deploy it in a way that maximizes safety and operational value.
Sources
ANYbotics adds gas leak detection capability to its inspection robot
Published by Vigla Media OÜ (Estonia).