Tesla has converted its former Model S and Model X line at its Fremont factory into a production line for the Optimus Gen 3 humanoid robot, a line reported to be built for roughly one million units a year, and Elon Musk is warning that the start will be extremely slow. The move puts Tesla’s best known robot program on a factory floor that built its most recognizable cars for well over a decade. In a July 1 post on X, Musk set expectations low for the opening phase even as the tooling around him points at enormous scale.
“Optimus production will be extremely slow at first, as everything is new. This is not like making a car,” Musk wrote, as reported by Electrek. He was specific about why. Optimus has roughly 10,000 unique parts, he said, and this year’s production rate is essentially impossible to predict. A complex machine moves at the pace of its slowest component, and when 10,000 distinct parts each need a reliable source, the number of things that can stall a line runs high. Building a car draws on more than a century of established supply chains. A humanoid robot at volume has no comparable history to lean on.
The line itself carries weight. Tesla tore down the Model S and Model X line at Fremont in about 46 days after the last of those cars were built in early May, ending a 14 year Model S run, according to Electrek. Teslarati documented the rapid teardown that cleared the floor and paved the way for Optimus. Handing over space that once built two of Tesla’s flagship vehicles is a real signal of how seriously the company is treating the robot, and it is not a swap a carmaker makes on a whim.
Set the reported capacity against Musk’s caution and the tension is plain. A line tooled to make a million machines a year ranks among the most ambitious manufacturing bets in consumer technology. Designed capacity and actual output are different things, though, and the distance between them is what Musk was flagging. Building the tooling for a million units describes what the factory could manage at full stride years from now. It says very little about what actually rolls off the line in the opening months, which is where Musk aimed his warning.
What the robots are matters as much as how many get built. The Optimus units produced in 2026 are meant for internal factory testing and data collection, not commercial sale. This year’s output amounts to a learning exercise. The robots exist to generate real world data, expose weak points in the Gen 3 design, and stress the young supply chain Musk described. None of them are going out the door as products. That reframes the ramp, because the point of 2026 is validation over volume, and volume targets mean less when nothing built is for sale.
Tesla is not stopping at one site. The company is also building a second Optimus factory at Giga Texas, with production expected around the summer of 2027. Read together, the two sites show how much weight Tesla is putting behind the program even while Musk tamps down near term expectations. Executives usually talk their timelines up rather than down, so a founder openly coaching expectations lower carries real weight, because it cuts against the normal incentive to sell the future hard.
The humanoid robot field has been running hot on expectations and cooler on delivered results, and Optimus is not being sold to anyone this year. A repurposed flagship line, a 46 day teardown, and a second factory already under construction all point to commitment. Commitment is not the same as a shipped product at scale, and Musk saying so out loud is the part that stands out. The picture for mid July 2026 holds together once the layers separate. The ambition is physical and visible on the Fremont floor. The reality Musk described is a slow start gated by 10,000 parts and a supply chain that is not yet mature, with this year’s robots staying inside Tesla’s own factories. Which of those two stories wins out over the coming months is the thing worth watching.