BMW is about to try something that’ll make every factory worker in America sit up a little straighter: humanoid robots on the shop floor.
Not in some far-off German lab, either. The experiment is slated for BMW’s massive Spartanburg, South Carolina, plant—the company’s U.S. crown jewel that already employs about 11,000 people. Over the next two years, BMW says it’ll roll out these robots gradually for “difficult, long, or dangerous” tasks. How many bots? BMW isn’t saying.
Call it the next step after a year of AI hype. Chatbots write emails. Humanoid robots move the stuff that keeps factories running.
Spartanburg becomes the test kitchen
BMW’s pitch is familiar: productivity up, costs down, fewer humans doing the miserable jobs. The company says it’ll first study which tasks fit the robots’ capabilities, then train the machines specifically for those jobs. Early guesses inside the industry: warehouse work and sheet-metal areas—places where repetitive handling and awkward lifting are constant.
Robert Engelhorn, chairman and CEO of BMW Manufacturing, framed it as survival and speed. “The automotive industry and with it vehicle production are evolving rapidly,” he said in comments carried by PR Newswire, adding that the company wants to integrate innovative technologies into production “to secure its future.”
Translation: if BMW can automate more of the grunt work without blowing up quality or safety, it will.
Meet Figure 01, the robot BMW wants on the line
The humanoid BMW is working with comes from Figure, a startup led by founder and CEO Brett Adcock. The robot is called Figure 01, and the specs are the kind that make you realize this isn’t a cute science-fair project:
It stands about 5 feet 7 inches tall (1.70 meters), weighs roughly 132 pounds (60 kg), and can lift up to about 44 pounds (20 kg). Battery life tops out around five hours. It moves at about 2.7 miles per hour (1.2 meters per second)—not exactly sprinting, but fast enough to be useful in a controlled industrial space.
The real selling point is the brain. Figure says the robot runs an integrated AI system “similar to ChatGPT,” meaning it can learn new tasks quickly—potentially by watching instructional videos rather than going through months of custom programming.
Yes, it learned to make coffee. That’s not the point—but it matters.
Figure went viral recently after posting images of the humanoid making coffee. Cute, sure. But the detail that should make factory managers—and factory workers—lean in is this: Figure claimed the robot needed less than 10 hours of training to pull it off.
Making coffee isn’t the future of manufacturing. Rapid training is.
If a general-purpose humanoid can be taught a new routine in half a day, that’s a very different world than the old industrial-robot model—where machines are bolted into cages and programmed to do one thing forever.
Where the robots could land first—and what BMW isn’t saying
BMW says these robots could be used across manufacturing processes, including body construction, body shops, and warehouses. Those are exactly the places where injuries happen, turnover is high, and “we can’t find enough people” has become management’s favorite excuse.
But let’s not kid ourselves: once a robot can reliably move parts, stock shelves, or handle repetitive shop tasks, the conversation shifts from “safety” to headcount. BMW hasn’t promised these machines won’t replace workers. It’s also not disclosing how many it plans to deploy, what the per-unit cost is, or what success looks like beyond corporate buzzwords.
Adcock, for his part, says the robot was designed to be safe around people, and he called the BMW partnership a major validation. He also took a swipe at traditional automation: “Single-purpose robotics has saturated the commercial market for decades,” he said, arguing that multi-purpose robotics is the untapped prize.
He’s right about one thing: if this works, every automaker—and every major U.S. manufacturer—will want a piece of it.
