In September 2025, a team from TIME magazine visited a weekend home in the Bay Area of California to observe a deployment that the publication would later feature as part of its Best Inventions of 2025 coverage. The subject was the Figure 03, a humanoid robot developed by California-based Figure, and the setting was a domestic one—a deliberate choice that signals a shift in how the company wants its machines to be perceived.
The visit was not a scripted demonstration in a laboratory. According to the source material, five Figure 03 units took turns performing tasks on camera while company founder Brett Adcock and his team played croquet on the lawn outside. The scene is almost pastoral, but the work being documented was decidedly practical. One robot loaded dishes into a dishwasher with notable accuracy. Another loaded laundry into a washer-dryer, though it did not retrieve an item it dropped. A third unit struggled with folding T-shirts.
The TIME team witnessed the Figure 03 successfully load items into a dishwasher and clear clutter from a table. The folding task, however, proved more difficult. This mix of competence and limitation is characteristic of the current state of humanoid robotics, and it is worth examining closely.
The Figure 03 itself is a hybrid design. It features a humanoid upper body with dual seven-degree-of-freedom arms and a four-degree-of-freedom articulated torso. Its vertical reach extends from ground level to 1.9 meters. Instead of walking legs, the robot sits on a holonomic wheeled mobile base, which allows it to move in any direction—including sideways and diagonally—without first rotating. This is a significant operational advantage in constrained environments like restaurant kitchens or hotel service corridors, where turning radius is a practical concern.
The wheeled base also affects deployment timelines. According to the source material, the design trades stair-climbing capability for stability, lower cost, and faster deployment. The wheeled architecture is what enables an eight-to-twelve-week deployment timeline, in part because the robot does not require pre-mapped facility layouts or GPS. Its navigation AI operates in real time from what its sensors actually see, built on FieldAI's physics-first Field Foundation Models.
This is not a concept robot. The source material indicates that Figure is producing the third-generation Figure 03 at a rate of one unit per hour as of 2026, and that the company has deployed robots in logistics, kitchen work, and manufacturing tasks. The benchmark cited is four to five hours of continuous neural network operation in those settings. The stated 2026 goal is more ambitious: drop a robot into an unseen home and have it perform useful work for days with minimal human intervention.
The TIME feature was published under the banner of Best Inventions of 2025, and the domestic launch of the 03 was a central theme. The publication's observation of the robot's dishwasher-loading and table-clearing abilities, alongside its T-shirt-folding struggles, provides a grounded picture of where the technology stands.
Why it matters for European robot service
For readers of Robot Service Map, the Figure 03 deployment is not merely a product announcement. It is a data point in a larger trend that directly affects the European robotics ecosystem: the commercialization of general-purpose humanoid platforms and the service infrastructure required to support them.
The source material notes that Figure 03 is one of several humanoid robots scheduled to hit the market in the next twelve months. Rivals include Boston Dynamics's Atlas, Apptronik's Apollo, and Tesla's Optimus. The source material also references Agility's Digit, which has a square head and inverted kneecaps, and Neura's 4NE-1. This is a crowded field, and Europe has its own players in this space. The question for European operators is not whether humanoids will arrive, but which platforms will be serviceable, maintainable, and cost-effective in local conditions.
The Figure 03's wheeled base is a particularly relevant detail for European deployments. Many European service environments—hotels, hospitals, warehouses, and catering facilities—are located in older buildings with narrow corridors, tight corners, and multiple floor levels. A bipedal robot that can climb stairs is attractive in theory, but the source material explicitly states that the wheeled design trades stair-climbing for stability, lower cost, and faster deployment. For a facility manager in, say, a historic hotel in Lisbon or a hospital in Rotterdam, the ability to move omnidirectionally in a constrained corridor may be more valuable than the ability to climb a staircase that an elevator can handle anyway.
The deployment timeline of eight to twelve weeks is also significant. The source material attributes this to the robot's reliance on real-time sensor data rather than pre-mapped layouts or GPS. For European service providers, this means less site preparation time and potentially lower integration costs. However, it also raises questions about network reliability, edge computing requirements, and the availability of local technical support—details that the source material does not disclose.
Another point of relevance is the production rate. The source material states that Figure is producing the third-generation Figure 03 at one unit per hour as of 2026. This is a manufacturing cadence, not a service commitment. It tells us that the company is scaling production, but it does not tell us about spare-part availability, field service response times, or maintenance contracts. European buyers should be cautious about extrapolating from production volume to service quality.
The source material also references a broader vision: the idea that by 2030, every household in the developed world has access to a humanoid robot, leased for $300 per month, performing laundry, cleaning, kitchen organization, and errands. This is a speculative projection, not a confirmed roadmap. The source material itself frames it as a future scenario. European operators should treat such projections as directional, not contractual.
What is more concrete is the current benchmark: four to five hours of continuous neural network operation in logistics, kitchen work, and manufacturing tasks. This is a measurable capability, but it also implies a duty cycle. A robot that can operate for four to five hours continuously may need recharging, cooling, or maintenance after that period. The source material does not specify the charging time, the battery capacity, or the recommended shift structure. These are operational details that European service integrators will need to clarify directly with the manufacturer.
What buyers and operators should know
The Figure 03 is a real product with demonstrated capabilities, but the source material provides a nuanced picture that should inform procurement decisions.
First, the robot's strengths are in structured, repetitive tasks. The TIME observation showed successful dishwasher loading and table clearing. These are tasks with clear physical constraints: dishes have a defined place, tables have a defined surface. The robot's dual seven-degree-of-freedom arms and articulated torso give it a wide reach and dexterity within its workspace. The holonomic base allows it to reposition without awkward rotations. For a kitchen or a warehouse aisle, this is a practical advantage.
Second, the robot's weaknesses are in tasks requiring fine manipulation of deformable objects. The T-shirt folding struggle is a clear example. Folding fabric requires precise force control and an understanding of material behavior that current neural networks may not fully capture. Buyers should not expect the Figure 03 to handle all household tasks equally well. The source material does not specify which other tasks were tested or how the robot performed in them.
Third, the deployment model is designed for speed. The eight-to-twelve-week timeline, enabled by real-time navigation AI and the absence of pre-mapped layouts, is a significant operational benefit. However, the source material does not disclose the cost of the robot, the terms of the lease, or the service-level agreements. It also does not specify whether the robot requires a dedicated charging station, a network connection, or on-site supervision. These are critical unknowns for any procurement decision.
Fourth, the production rate of one unit per hour is a supply-side indicator, not a demand-side guarantee. It suggests that Figure is investing in manufacturing capacity, but it does not indicate how many units are actually deployed, where they are deployed, or how they are performing in the field. The source material mentions that five robots were present at the TIME visit, but it does not state the total fleet size.
Fifth, the competitive landscape is crowded. The source material lists at least six humanoid platforms scheduled to hit the market in the next twelve months. This is good news for buyers in terms of choice, but it also means that the service ecosystem—spare parts, trained technicians, software updates—may be fragmented. European operators should evaluate not just the robot's capabilities but the manufacturer's commitment to the European market, including local support infrastructure.
Sixth, the source material includes a cautionary note about the difference between a product and a remote-controlled toy. The quote in the source material—"If you can't do this, you don't have a product. You have a very expensive remote-controlled toy"—refers to the ability to operate autonomously for extended periods. The Figure 03's current benchmark of four to five hours of continuous neural network operation is a step toward that goal, but the 2026 goal of days-long autonomous operation in unseen homes is not yet achieved. Buyers should verify the current autonomy level against their specific use case.
Seventh, the source material references a specific industrial deployment: BMW X3 vehicles in a Spartanburg, South Carolina plant. This is a manufacturing context, not a service context. It demonstrates that Figure has experience in industrial settings, but it does not imply that the robot is ready for all service environments. The source material also notes that neither Figure nor its competitors have attempted six verticals simultaneously from a commercial launch. This suggests a cautious, vertical-by-vertical approach to market entry.
Finally, the source material includes a speculative scenario about biometric data and meal preparation. This is not a current capability. It is a vision of the future. European buyers should not make purchasing decisions based on speculative features that are not yet available or specified.
In summary, the Figure 03 is a credible, production-ready humanoid robot with demonstrated capabilities in kitchen and logistics tasks. Its wheeled base and real-time navigation AI make it suitable for constrained indoor environments. However, the source material leaves several important questions unanswered: the total cost of ownership, the service and maintenance terms, the battery and charging specifications, and the roadmap for software updates. European operators should engage directly with Figure to obtain these details before making any commitments.
Sources
https://time.com/7324233/figure-03-robot-humanoid-reveal/
Published by Vigla Media OÜ (Estonia).