Robotics / News

Figure 02 at BMW Spartanburg: 11 months of results, then Figure 03

Figure AI's Figure 02 humanoid worked at BMW's Spartanburg plant for 11 months, logging 1,250 hours and loading more than 90,000 sheet metal parts, according to a report on the deployment. BMW now plans to use the newer Figure 03. The numbers are real work, but small next to a plant's total output.

RoboHub editorialPublished: Source date:

At a glance · reported claims

11

Months of deployment

1,250 h

Reported operating hours

90,000+

Sheet metal parts loaded

30,000+

BMW X3 vehicles it contributed to

What was reported

An industrial IoT publication (IIoT World) summarizes the Figure 02 deployment at BMW's Spartanburg, South Carolina plant: 11 months, 1,250 hours of operation, more than 90,000 sheet metal parts loaded, and a contribution to the production of more than 30,000 BMW X3 vehicles. The excerpt we could check does not name the original source of the figures or describe the loading task in detail.

The Robot Report's June 2026 roundup confirms the follow-up: after the successful Figure 02 deployment at Spartanburg, BMW will deploy Figure 03, announced in June 2026.

Putting the numbers in scale

A simple division (our arithmetic, not a reported figure): 1,250 hours over 11 months is roughly 114 hours a month, and 90,000 parts over 1,250 hours is about 72 parts an hour. That is steady, useful work for a first humanoid on a real line, though it points to partial shifts rather than around-the-clock operation. Whether it was one robot or several is not stated in what we reviewed.

Contributing to 30,000 vehicles means the robot handled one step among many; it did not build them. The reports give no figures on downtime, interventions, cost per part or comparison with the manual process, which are the numbers that decide whether a deployment pays off.

Why Figure 03 follows

According to the photo caption in The Robot Report, Figure 03 adds tactile-sensor hands, palm cameras, wireless charging and speech-to-speech audio compared with its predecessor. Hands, sensing and charging are the parts that matter most on a line: they affect grip on awkward parts, recovery from a misgrasp and how many hours a day a robot can actually work.

The announcement is a plan to deploy Figure 03. The text we reviewed gives no start date, number of units or targets for hours or parts.

What this means for robot buyers

RoboHub's view: this is one of the longer public deployments of a humanoid in a car plant, which makes it a useful reference, but it is a single case from a vendor-friendly story. Before treating it as a benchmark, ask for the same facts for your use case: hours per shift, parts per hour, interventions per day, and what the robot costs per month including integration.

Also ask about generation changes. Figure 02 gave way to Figure 03 within about a year, so a unit bought today may be superseded quickly. Check what upgrade path, trade-in or lease terms exist, and whether training data and workflows carry over to the next model. Figure robots are sold to businesses and no public prices are given in these sources.

Sources & editorial note

An original RoboHub summary and analysis based on the sources listed above. Figures and plans are attributed to the companies and media that published them; RoboHub has not independently tested the robots. This is not a sponsored article.

Explore robots for your task