Humanoids Move From Labs To Factories

The real shift is that humanoids have moved from lab demos to factory insertion, which means the bottleneck is no longer whether a robot can move, but how fast companies can collect enough real work data to make fleets useful. Foundation is already planning first deliveries to an auto OEM in April and May 2025, and the wider category is now organized around narrow industrial jobs where robots can step into existing lines without a year long retrofit.

• What makes this different from the Boston Dynamics era is the stack. Cheap reliable motors, cameras, processors, and GPUs arrived at the same time as transformer based models that can turn raw sensor input into actions in messy environments, instead of needing every movement hard coded in advance.
  1 sacra
• The first wave is not a home robot story. It is auto plants, warehouses, and other sites with repetitive, dangerous jobs, high turnover, and layouts already built for human hands and legs. That is why humanoids can win before fully custom automation, because customers can swap them in beside people instead of rebuilding the building.
  1 sacra 2 sacra
• The market is close in the sense that serious players are now spending against deployment, not just prototypes. Figure has raised about $1.75B, Agility about $641M, Apptronik partnered with Google DeepMind, and Tesla said first generation Optimus production lines were being installed in anticipation of volume production. Even so, revenue is still mostly pilots and the main race is building the data flywheel.
  2 sacra 3 sacra 4 sacra 5 tesla

Over the next few years, the winners will be the companies that turn a few narrowly useful robots into large installed fleets, then use intervention and operating data to widen the job list. If that happens, humanoids stop being a robotics category and start becoming a new labor layer for industry first, then consumer markets later.