Tesla scales humanoid to 24/day
- Tesla’s Optimus story is shifting from demo videos to factory math: Elon Musk said on April 22 production starts in Fremont in late July or August. - The key constraint is scale, not concept — Musk said Optimus has about 10,000 unique parts and early output will be “quite slow.” - That matters because Tesla is now treating humanoids like a manufacturing program, not a lab project — but volume still looks like a 2027 story.
Humanoid robots are finally hitting the part of the story that matters — factories, yields, bottlenecks, and ugly ramp charts. That’s why the leaked Tesla Optimus roadmap is interesting. Not because “24 robots a day” is a huge number by itself, but because it suggests Tesla is trying to industrialize a humanoid the same way it industrialized cars. The timing also lines up with what Elon Musk said on Tesla’s April 22, 2026 earnings call: production is supposed to start in Fremont in late July or August, and the first output will be slow because this is a brand-new product with a brand-new supply chain. ### What actually changed? The new piece is the manufacturing cadence. The leaked notes describe a 120-day ramp from roughly 1 robot per day to 24 per day. That turns Optimus from a vague “someday” project into a concrete factory plan with throughput targets, yield assumptions, and weekly pain points. Tesla has talked big about Optimus before — in January the more realistic outcome might be only a few thousand. ### Why is 24 per day a real threshold? Because it’s enough to prove a line, not enough to prove a business. Twenty-four a day is around 168 a week if the line runs every day — far below car-scale manufacturing, but high enough that every weak link starts showing up fast. Early prototype work can hide fragile assembly steps behind stable suppliers, test coverage, and rework that doesn’t choke the line. That’s the real meaning of the roadmap. ### Why does vertical integration matter so much? Because actuators are basically the robot. A humanoid is not hard in just one way. It is hard in thousands of tiny mechanical ways — motors, gearboxes, screws, sensors, wiring, thermal limits, hand dexterity, and control loops all stacked together. Tesla’s pitch is that it already knows how to design better than startups that buy more of those parts from outside vendors. The leaked notes lean hard on that idea, and it fits Tesla’s long-running playbook in cars and energy products. ### So what’s the catch? The catch is that a humanoid is not a car with legs. Musk said early production will be “quite slow” because Optimus has around 10,000 unique parts and almost nothing in the supply chain carries over from Tesla’s vehicle programs. That means even if the leaked roadmap is real, the headaches rise, every missing connector, bad calibration, or flaky actuator turns into line-wide friction. ### Why use Fremont? Because Tesla appears to be giving Optimus real factory space, not just skunkworks attention. Reporting around the April 22 call says Tesla is converting the old Model S/X line in Fremont for Optimus production. That is symbolically important. It says Tesla is willing to trade mature vehicle capacity for a bet on humanoids — at least at the margin. ### Does this make Tesla the humanoid leader? Maybe in ambition, not yet in shipped volume. The leaked roadmap shows Tesla thinking like a mass manufacturer earlier than most rivals. But the whole field still has the same problem: flashy demos are easy to overread. The winner will be the company that can build reliable robots at repeatable cost, the knows how to ramp hard things. But for now, the story is still about whether the line works — not whether the robot has won.
