Figure’s newest humanoid robot, the Figure 03, has begun performing logistics work inside BMW Group Plant Spartanburg in South Carolina, the automaker’s largest assembly facility by production volume. The deployment advances a commercial agreement between Figure and BMW Manufacturing that started with early trials of the smaller Figure 02 model and has now graduated to a defined, shift-ready workflow. For an auto industry under pressure to cut costs and fill factory roles, the question is whether a humanoid machine sorting parts on a live production floor can prove reliable enough to spread across other plants.
Why Spartanburg is the testing ground for humanoid factory labor
The Figure 03 is not running a demo loop or performing staged tasks for cameras. According to BMW’s corporate communications division, the robot is executing a specific logistics sequencing use case at Spartanburg, picking unsorted components and arranging them for assembly lines. That is a real production task, not a research exercise, and it sits inside a plant that builds SUVs for global export.
BMW Manufacturing VP Ulrich Wieland provided an on-the-record statement about the project, framing it as part of the company’s broader push to integrate what BMW calls “Physical AI” into production. Figure CEO Brett Adcock described the deployment as a step in moving general-purpose robots from controlled lab settings into working factories. Both executives spoke publicly about the effort through BMW Group’s official channels and via a joint announcement that outlined how humanoid systems would be introduced at Spartanburg.
The immediate consequence is straightforward. If Figure 03 can handle repetitive logistics sorting across full shifts without significant downtime or error, Spartanburg becomes a reference site. Other BMW plants, and potentially other automakers, would have a concrete case study rather than a pitch deck. The hypothesis worth tracking: a successful six-month run at Spartanburg could shorten the sales cycle for Figure’s machines at additional facilities by removing the biggest objection buyers raise, which is whether the technology actually works in a messy, fast-moving factory.
From Figure 02 trials to Figure 03 shift work at BMW
The current deployment did not appear overnight. Figure and BMW Manufacturing first announced a commercial framework structured around a milestone-based approach, disclosed through the corporate newswire channels that BMW and Figure use for investor- and media-facing updates. That framework laid out a sequence: identify initial use cases, stage robots into Spartanburg operations, then explore deeper AI-driven robot control, virtualization, and integration with existing systems.
Before the Figure 03 arrived, BMW ran trials with the earlier Figure 02 in a real production environment at the same plant. That robot stands approximately 170 cm tall, weighs 70 kg, and can carry loads up to 20 kg, according to BMW’s own descriptions of its early work with humanoids in manufacturing. During those initial tests, Figure 02 performed a specific task: placing sheet metal parts into fixtures for body-in-white production. BMW board member Milan Nedeljkovic and Figure CEO Brett Adcock both commented publicly on the early work, signaling executive-level attention from both companies and helping clear internal hurdles for a second phase.
The progression from sheet metal handling to logistics sequencing represents a meaningful expansion. Sorting unsorted parts for assembly lines requires the robot to identify different components, orient them correctly, and place them in the right order. That is a step up in complexity from loading identical metal pieces into a fixed jig. It also sits closer to the kind of flexible, judgment-dependent work that humanoid robots are supposed to handle better than traditional fixed-arm automation.
For BMW, the appeal is not just novelty. A humanoid platform can, in theory, move through existing aisles, reach into bins designed for human workers, and interface with tools and equipment already on the floor. That makes it easier to retrofit automation into older sections of the plant without rebuilding line layouts. For Figure, the Spartanburg deployment is a proving ground for how well its hardware, perception systems, and AI control stack can adapt to a complex, human-centric workspace that was never designed for robots.
What the Spartanburg rollout still has to prove
The public record so far comes entirely from corporate announcements by BMW and Figure. No independent data on actual shift hours completed, error rates, or throughput comparisons with human workers has been released. There are no published statements from plant-floor employees or union representatives about how the robots fit into daily operations, how job assignments have changed, or whether any roles have been displaced.
The milestone-based structure of the agreement suggests BMW is not committing to a full-scale rollout until specific performance thresholds are met, but neither company has disclosed what those thresholds are. Without public benchmarks, outside observers have no way to assess whether the Figure 03 is meeting, exceeding, or falling short of expectations. Key unknowns include how often human technicians need to intervene, how robust the robots are to variations in parts presentation, and what happens when upstream processes introduce unexpected delays or errors.
There is also an open question about economics. Humanoid robots are expensive to build and maintain. Traditional industrial robots, the kind already common in auto plants, are proven and relatively cheap per unit of output. For Figure’s machines to earn a permanent place on the floor, they need to handle tasks that fixed automation cannot do cost-effectively, which generally means work that requires mobility, dexterity, and the ability to adapt to changing conditions. Logistics sequencing fits that profile in theory, but the financial math has not been made public, and neither company has broken out capital or operating costs associated with the Spartanburg deployment.
From a labor perspective, the lack of detail leaves room for speculation. BMW and Figure have emphasized that humanoid robots can take on repetitive, ergonomically challenging tasks that are hard to staff and contribute to injuries. That narrative positions the Figure 03 as a supplement to the existing workforce rather than a direct replacement. Until there is more clarity on how job roles evolve around the robots, however, it will be difficult to separate long-term workforce strategy from short-term pilot rhetoric.
What comes after a successful pilot
The next development to watch is whether BMW expands the number of Figure 03 units at Spartanburg or announces plans for a second plant. The commercial agreement explicitly includes exploration of broader integration, but no commitments have been made public on how many robots BMW might eventually deploy or which production lines would be prioritized. A move from one or a handful of units to dozens would signal that the technology has cleared internal hurdles on reliability, safety, and cost.
Another indicator will be how deeply the robots are tied into BMW’s digital infrastructure. The original agreement referenced not just physical deployment but also virtualization and AI-driven control, implying that Figure’s systems could be linked into BMW’s planning, simulation, and quality monitoring tools. If that integration happens, it would shift the robots from being stand-alone machines to becoming nodes in a broader production network that can be monitored and optimized in real time.
On Figure’s side, the Spartanburg project will likely influence how the company approaches other customers. A documented use case in a high-volume automotive plant can serve as a reference for manufacturers evaluating whether to bring humanoid robots into their own facilities. The same communications infrastructure that carried the initial announcement, including the distribution tools used by corporate communications teams, will be central to how Figure presents performance data, new milestones, and any follow-on contracts.
For now, the Figure 03 at Spartanburg stands as a live experiment in whether general-purpose humanoid robots can move from glossy concept videos to unglamorous factory work. The outcome will not hinge on how human the machine looks, but on whether it can show up every day, handle the variability that comes with a real production line, and do so at a cost that makes sense for one of the world’s most efficiency-obsessed industries. Until BMW and Figure share more than carefully curated footage and executive quotes, the broader implications for factory labor will remain a work in progress rather than a settled story.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
