Figure AI’s factory line for its Figure 03 humanoid robot has reached a production pace of one unit per hour, a 24-fold jump from one robot per day achieved in under 120 days. The company’s dedicated manufacturing facility, called BotQ, has delivered more than 350 third-generation humanoids so far, and its first-generation lines are designed to produce up to 12,000 units per year at full capacity. The speed increase arrives as Figure’s earlier model, the Figure 02, already operates inside a BMW plant in Spartanburg, South Carolina, where it helped build more than 30,000 BMW X3 vehicles over ten months.
Why one robot per hour changes the math for automakers
The production milestone matters because it shifts the conversation from pilot programs to volume deployment. A single humanoid rolling off the line every hour means BotQ could, in theory, produce thousands of robots annually, enough to staff multiple factory floors rather than a single test cell. That throughput target of up to 12,000 humanoids per year would represent a supply large enough to serve dozens of automotive and logistics customers simultaneously.
The hypothesis that Figure’s installed base could exceed BMW’s annual U.S. vehicle output by the end of 2026 requires scrutiny. BMW’s Spartanburg plant is one of the largest vehicle exporters in the United States, producing hundreds of thousands of SUVs each year. Even at a sustained one-per-hour clip running around the clock, BotQ’s theoretical annual output of roughly 8,760 units falls well short of Spartanburg’s vehicle volume alone. The 12,000-unit annual capacity ceiling, which Figure describes as a first-generation target, would need to scale several times over before the robot population rivaled a major automaker’s production numbers. The hypothesis overstates the near-term impact. What the pace does signal is that Figure has moved past handbuilt prototyping and into repeatable manufacturing, a step that most humanoid robotics competitors have not publicly demonstrated at comparable scale.
BotQ output and the BMW factory record
Figure reported that BotQ delivered over 350 third-generation humanoids and accelerated from one Figure 03 per day to one per hour in under 120 days. The company said it transitioned from CNC prototyping to injection-molded and tooled processes across more than 150 manufacturing stations, a shift designed to cut per-unit cost and cycle time as volumes grow.
Separately, BMW’s own communications confirm that the earlier Figure 02 model worked ten-hour shifts Monday through Friday at the Spartanburg plant during 2025. Over that ten-month stretch, the robot supported production of more than 30,000 BMW X3 vehicles and moved 90,000-plus components on the assembly line. Those numbers come directly from BMW Group corporate communications, not from Figure, which gives them an independent layer of credibility. The Spartanburg deployment grew out of a formal commercial agreement between Figure and BMW Manufacturing announced through a joint press release, which described a milestone-based approach to bringing general purpose robots into automotive production.
The gap between the Figure 02 deployment at Spartanburg and the Figure 03 production ramp at BotQ is significant. The 02 model proved it could handle repetitive component-moving tasks inside a real factory over sustained shifts. The 03 model is the version Figure now manufactures at volume, suggesting the company views the Spartanburg pilot as validation enough to commit to high-rate production of a newer design. For automakers, that progression-from single-robot trials to standardized hardware rolling off a line-matters more than any individual performance anecdote.
What the production numbers do not yet prove
Several questions sit between Figure’s announced pace and a proven industrial product. First, the one-per-hour rate and the 350-unit delivery count come entirely from Figure’s own disclosures. No third-party audit, independent production log, or supplier verification has been published to confirm sustained throughput at that speed. Manufacturing startups frequently cite peak rates that differ from average sustained output, and without external validation, the number represents a company claim rather than an audited benchmark.
Second, BMW’s Spartanburg data, while useful, lacks the granularity that would let outside analysts judge robot reliability. The automaker’s release provides aggregate vehicle counts and total components moved but does not include cycle-time breakdowns, error rates, unplanned downtime, or maintenance intervals for the Figure 02 units. Those metrics are standard in automotive quality reporting, and their absence makes it difficult to compare Figure’s robots against conventional industrial automation on a like-for-like basis.
Third, Figure’s stated BotQ capacity of 12,000 units per year is a design target for first-generation lines, not a demonstrated annual output. No supporting supplier contracts, floor-plan documentation, or ramp schedules have been made public. The distance between a line “capable of” a rate and a line consistently hitting that rate is where most hardware startups encounter their biggest setbacks. Until Figure publishes longer-term production data, investors and customers will have to treat the 12,000-unit figure as aspirational.
How Figure’s claims fit into the broader robotics ecosystem
The context around these announcements also matters. Industrial and technology companies increasingly route their major partnership news through large distribution services such as PR Newswire’s media network, which ensures filings and corporate statements reach reporters, investors, and industry analysts at the same time. Figure and BMW chose that path for their manufacturing agreement, signaling that both sides expected market scrutiny and wanted a formal record of the terms they were willing to publicize.
Behind the scenes, communications and investor-relations teams use secure portals like PR Newswire’s client login to stage, approve, and schedule those announcements. The reliance on structured distribution does not independently verify the underlying technical claims, but it does create a documented trail of who said what, and when. For emerging robotics companies, that trail can later be compared against actual performance as deployments scale.
Figure’s decision to highlight both the BMW deployment and the BotQ ramp through formal channels reflects a broader shift in how humanoid robotics is being positioned. Rather than emphasizing speculative, open-ended “general intelligence,” the company is foregrounding specific, countable outcomes: robots per hour, components moved, vehicles supported. That framing aligns more closely with how automakers evaluate capital expenditures, where metrics like mean time between failures and cost per unit of throughput ultimately determine whether a new technology survives procurement reviews.
What comes next for humanoid robots in factories
If Figure can sustain its reported production rate and translate the BMW pilot into additional factory contracts, the company could help define a new category of automation somewhere between fixed industrial robots and human line workers. Humanoids capable of using existing tools, navigating current layouts, and learning new tasks through software updates promise to reduce the need for expensive retooling when product mixes change.
Yet the open questions around durability, safety certification, and lifetime operating costs remain substantial. Automakers will want multi-year data on how often the robots fail, how quickly they can be repaired, and how their performance degrades over time. Labor regulators and plant safety officers will scrutinize human-robot interaction zones, emergency stop behavior, and edge cases that rarely appear in marketing videos.
For now, Figure’s one-robot-per-hour milestone is best understood as a marker that the company has crossed an important internal threshold: it can build its latest humanoid on a repeatable, tool-based line rather than in a lab. That does not guarantee market dominance or even long-term viability, but it does separate Figure from peers still showing only hand-assembled prototypes. As more data emerges from BotQ and from deployments like BMW’s Spartanburg plant, the industry will get a clearer view of whether humanoid robots are ready to move from headline-grabbing demos to the unglamorous, demanding reality of factory life.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
