In early December 2025, the robotics industry witnessed a moment that, on the surface, might seem like a simple social media post. Brett Adcock, the chief executive of Figure AI, shared a short video clip on the Threads platform. The footage was brief, but its content carried considerable weight for anyone tracking the state of humanoid robotics. The video showed the company’s latest machine, the Figure 03, running.
The clip, as described in reporting from Forbes, was not merely a demonstration of a robot walking on a treadmill or taking measured steps across a lab floor. Instead, it showed the Figure 03 breaking into a genuine run, with a visible quick start from a stationary position, smooth changes of direction while in motion, and a sharp braking action to come to a stop. The setting was described as the interior of a complex, suggesting a real-world environment rather than a carefully staged outdoor track.
What made this footage notable was not just the fact that the robot could run, but the manner in which it did so. The movements appeared fluid and controlled. The acceleration was evident, the turns were executed without hesitation or stumbling, and the deceleration was decisive. For observers who have followed the humanoid robotics sector over the past several years, this visual demonstration stood in stark contrast to the typical footage of bipedal machines cautiously shuffling forward or taking deliberate, slow steps.
The significance of the video was underscored by a follow-up detail shared by Adcock. He indicated that the Figure 03 was running at a speed of 6.5 miles per hour. To put that figure into context, that pace sits near the upper end of what is generally considered a human jogging speed. For a machine that must balance on two legs, manage dynamic loads, and coordinate dozens of actuators in real time, sustaining that speed represents a considerable engineering achievement.
The video also prompted speculation about future events. Adcock reportedly mentioned the possibility of entering the robot in the second annual Humanoid Robot Marathon, which is scheduled for the following year. Whether that participation materialises remains to be seen, but the mere suggestion indicates a level of confidence in the platform’s dynamic capabilities.
It is worth noting the timeline of this disclosure. The video was shared in early December 2025, and the reporting from Forbes was published on 4 December 2025. The exact time of day the video was posted is not disclosed in the source material, but the month and year are clear.
Why it matters for European robot service
For readers of Robot Service Map, particularly those involved in the deployment, maintenance, and servicing of robotic systems across Europe, this development carries implications that extend beyond a single company’s marketing efforts.
The European market for robotic services has grown steadily, with a strong focus on industrial automation, logistics, and, increasingly, service robots intended for public and domestic environments. Humanoid robots, while still a niche segment compared to traditional industrial arms or autonomous mobile robots, have attracted significant attention from both investors and potential end users. The promise of a machine that can navigate environments designed for humans—staircases, narrow corridors, doorways, and workspaces—has driven interest in sectors ranging from healthcare to hospitality and warehouse operations.
One of the primary criticisms levelled at humanoid robots in recent years has been their speed, or rather, the lack of it. Many of the machines demonstrated to date have moved at paces that would be impractical for most commercial applications. A robot that walks at a leisurely 2 or 3 miles per hour might be acceptable for a guided tour or a controlled demonstration, but it is not suitable for tasks that require timely completion, such as responding to a service call, fetching an item from across a warehouse floor, or accompanying a worker on a patrol route.
The Figure 03 video challenges that assumption directly. If a humanoid robot can run at 6.5 miles per hour, accelerate quickly, change direction smoothly, and brake sharply, then the envelope of possible use cases expands considerably. For European service providers who are evaluating whether to integrate humanoid platforms into their offerings, this is a relevant data point.
Consider the logistics sector, which is a major driver of robotic adoption in Europe. Warehouses and distribution centres are often vast, and the distances between workstations can be significant. A robot that can only walk slowly would spend a disproportionate amount of its time in transit. A robot that can jog between stations, however, becomes a more viable candidate for tasks that involve moving items, tools, or documents across a facility.
Similarly, in facilities management and security, the ability to move quickly is often essential. A security robot that must respond to an alert or investigate an anomaly needs to reach the location in a timely manner. A patrol robot that can cover ground at a jogging pace is more effective than one that ambles.
The domestic and care sectors also stand to benefit. While speed is not the primary requirement for a robot assisting an elderly person with daily tasks, the ability to move briskly when needed—such as fetching a glass of water or responding to a call button—could improve the perceived utility of the device. Moreover, the smoothness of the motion, as evidenced by the Figure 03’s directional changes, is relevant for user comfort. A robot that lurches or stumbles is unsettling to be around; one that moves fluidly inspires more confidence.
It is important, however, to place this development in a broader context. The Figure 03 is not the only humanoid robot in development. Competitors such as Agility Robotics with its Digit platform and Sanctuary AI with Phoenix have published maximum walking speeds in the range of 3 to 4 miles per hour, according to the source material. These are respectable speeds for walking, but they are below the running pace demonstrated by Figure 03.
Figure AI itself has reported a more conservative figure for the Figure 03. The company has stated a speed of 1.3 meters per second, which equates to approximately 2.7 miles per hour. This discrepancy between the company’s stated specification and the performance shown in the video is worth noting. The Forbes report indicates that the robot appears to be operating at the upper end of human jogging speed, which is generally considered to be 4 to 6 miles per hour. The 6.5 miles per hour figure shared by Adcock is slightly above that range, suggesting the robot may have been pushed to its limits for the demonstration.
For European buyers and operators, this raises a practical question: what is the actual, reliable, repeatable performance of the Figure 03 in day-to-day operation? The video shows a single run, presumably under controlled conditions. It does not indicate how the robot performs over longer distances, how battery life is affected by running, or how the system handles uneven terrain, wet floors, or other real-world variables. These details are not disclosed in the source material, and it would be inappropriate to speculate.
What can be said is that the demonstration signals a shift in the competitive landscape. For years, the narrative around humanoids has been one of caution and limited capability. The Figure 03 video suggests that at least one company is pushing hard to change that narrative. For the European robot service ecosystem, which includes integrators, maintenance providers, and end users, this means that the technology is evolving faster than some may have anticipated.
What buyers and operators should know
For those in Europe who are considering the adoption of humanoid robots, or who are already operating such systems, the Figure 03 demonstration offers several points of consideration.
First, speed is a differentiator, but it is not the only metric that matters. The video shows impressive acceleration, turning, and braking. These are dynamic capabilities that are difficult to achieve in bipedal robots. However, the source material does not provide information on other critical factors such as payload capacity, battery runtime, reliability, or the cost of the system. Buyers should not assume that a fast robot is necessarily a productive robot. The ability to run at 6.5 miles per hour is noteworthy, but it must be weighed against the robot’s ability to perform the actual tasks required in a given environment.
Second, there is a discrepancy between the company’s published speed specification and the performance shown in the video. Figure AI has reported 1.3 meters per second, or 2.7 miles per hour, for the Figure 03. The video appears to show the robot running at a much higher speed, with Adcock citing 6.5 miles per hour. This gap between specification and demonstration is not unusual in the robotics industry, where marketing demonstrations often showcase peak performance. However, buyers should be aware that the real-world, sustained performance of the robot may differ from what is shown in a short clip.
Third, the competitive landscape is evolving. Agility Robotics and Sanctuary AI have published walking speeds in the 3 to 4 miles per hour range. Figure AI’s demonstration, if it reflects the true capability of the Figure 03, places the company ahead of these competitors in terms of speed. This could influence purchasing decisions, but it also means that the market is becoming more dynamic. Buyers should monitor developments from multiple vendors before committing to a platform.
Fourth, the mention of a potential entry in the Humanoid Robot Marathon is an interesting signal, but it is not a commitment. Adcock said that perhaps Figure will enter its robot in the event next year. This is speculative and should not be interpreted as a confirmed plan. The marathon, if it takes place, would be a valuable test of endurance and reliability, but until a formal announcement is made, it remains a possibility rather than a certainty.
Fifth, the source material does not provide any information on serviceability, spare parts, or maintenance requirements for the Figure 03. Robot Service Map does not have data on these aspects from this source, and it would be misleading to suggest otherwise. Buyers who are serious about deploying humanoid robots should seek detailed information from the manufacturer on these topics, including expected service intervals, availability of replacement components, and the level of technical support offered in Europe.
Sixth, the demonstration was conducted in an indoor complex. The source material does not describe the floor surface, the ambient conditions, or the duration of the run. Outdoor operation, in particular, presents additional challenges such as uneven ground, varying light conditions, and weather. Buyers should not assume that the performance shown in the video can be replicated outdoors without further testing.
Seventh, the speed of 6.5 miles per hour is impressive, but it is worth noting that this is near the upper end of human jogging speed. The robot is not sprinting; it is jogging. For many applications, a brisk walk might be sufficient. The key takeaway is not that the robot can run, but that it can move with agility and control at a pace that is practical for real-world tasks. The smooth directional changes and sharp braking are arguably more important than the top speed, as these indicate a high level of dynamic stability and control.
Finally, it is important to recognise that this is a teaser video. The source material describes it as a short clip. It is not a comprehensive demonstration of the robot’s capabilities. The full range of the Figure 03’s abilities, limitations, and operational parameters has not been publicly disclosed. Buyers should approach the video with appropriate caution and seek additional information from Figure AI before making any decisions.
In summary, the Figure 03 running video is a significant data point in the evolution of humanoid robotics. It challenges the perception that these machines are inherently slow and suggests that dynamic performance is improving rapidly. For European robot service providers and end users, this is a development worth monitoring. However, it is also a reminder that marketing demonstrations are not the same as operational reality. The true test of the Figure 03 will come when it is deployed in real-world environments, performing real tasks, over extended periods.
The source material for this article is limited to the Forbes report and the details shared by Adcock. No additional information on pricing, availability, or technical specifications is available from this source. Readers are encouraged to consult the original article for further context and to monitor future announcements from Figure AI for more detailed information.
Sources
https://www.forbes.com/sites/johnkoetsier/2025/12/04/figure-ceo-teases-running-humanoid-robot-it-looks-fast-agile-smooth/
Published by Vigla Media OÜ (Estonia).