Figure AI's humanoid robots began operating inside a BMW manufacturing facility in Spartanburg, South Carolina in 2024. Not in a demonstration cage. On the production line.
The tasks assigned to Figure 01 — and to subsequent deployments — were specific and deliberately chosen: sheet metal transfer, part loading, inspection of components in repetitive sequences. Tasks that are physically demanding, ergonomically damaging over time, and cognitively simple. The kind of work that wears people out over years and is genuinely hard to fill. Figure's robot doesn't think about where it's going next in life. It moves sheet metal, precisely, without fatigue.
Tesla's Optimus program is further behind in deployment terms but ahead in public visibility. The version demonstrated in late 2024 showed a robot capable of folding laundry, sorting objects by touch, and navigating factory floor environments without operator guidance. Tesla's internal timeline targets Optimus performing useful manufacturing work inside Tesla's own facilities — an unusual arrangement where the robot-builder is also the first customer. Boston Dynamics' Atlas, the research robot that's been flipping and running for years, transitioned to an electric version designed explicitly for industrial deployment.
What's happening here is a compression of the cycle from lab to factory floor. Two years ago, the honest assessment of humanoid robots was: technically impressive, economically premature. Today that's changed for a narrow but real category of tasks — the ones that are physically repetitive, spatially constrained, and don't require social judgment.
The more interesting story is what's happening to the humans in these facilities. They're not leaving. They're changing roles. At BMW Spartanburg, engineers and line technicians are working on robot supervision: monitoring sensor feeds, intervening when the robot encounters an edge case it can't handle, maintaining the mechanical and software systems that keep the units running. Factories that deploy robots need people who understand both manufacturing process and robotics integration. That skill profile didn't exist at scale five years ago. It's being built now, in real facilities, by people who were already working there.
The honest version of the robotics story in 2025 is not that machines are taking over. It's that a specific category of physically unpleasant, repetitive work is being automated — and the people who were doing it are being asked to move into supervision, programming, and maintenance roles that pay better and require more skill. That transition is not frictionless. Retraining takes time, geographic constraints are real, and not every worker in a facility doing one kind of job can easily transition to a different one. Those challenges are worth taking seriously.
But the direction is clear. The robots are no longer just in demos. They're on the floor. The next few years will determine how well the industry, policymakers, and workers themselves manage the transition that follows.