Tesla has begun producing its Optimus humanoid robot at its factory in Fremont, California, the company said Wednesday in its third-quarter letter to shareholders, describing the machines as "the first production-quality bots that we have manufactured." The disclosure landed alongside earnings that showed the cost of that ambition: net income fell 43% to $787 million from $1.37 billion a year earlier, as revenue stayed roughly flat at $28.04 billion while spending on AI, autonomy and robots climbed.
Chief executive Elon Musk called this "early production" and said the ramp would be slow and "somewhat bumpy because it is such a radically new product." He nonetheless said Tesla aims to build 20,000 Optimus units a week by the end of 2027, a target that should be read as a company projection rather than an established schedule; Musk also said the robot still needs design changes, that building its hands is a central difficulty, and that "there's quite a lot of work to do" on the AI before it is useful for open-ended tasks. The latest version has not been shown publicly, with a video promised in one to two months.
The Fremont line occupies space where Tesla until this year built the Model S and Model X, a switch that signals how firmly Musk is staking the company's future on robots over cars. He has described an "Optimus Academy" where robots practice tasks to generate training data, and said a next-generation AI5 chip due in April would deliver a "big step change in intelligence." A second line under construction in Austin is planned for far higher volume.
Tesla is moving into an increasingly crowded field. Agility Robotics has deployed its Digit robot across nine customer facilities and plans to go public; Figure AI, valued at $39 billion last year, has announced logistics deployments; and Sunday Robotics, 1X and Weave Robotics are preparing to ship robots into homes this fall. China's Unitree, the world's largest robotics maker by volume, this week showed off a martial-arts "guardian" humanoid, though its shares have slid 55% from their post-IPO peak and Reuters reports Beijing has slowed a queue of further robotics listings.
Underwriting the whole sector is a surge of capital. Robotics and "physical AI" companies raised $33.4 billion globally in the first half of this year, according to PitchBook, already more than the $28.7 billion raised in all of 2024 and far above 2020's sub-$5 billion. Nvidia, which supplies much of the training and simulation software behind these robots, expanded its Isaac platform this week at its Berlin conference, with chief executive Jensen Huang repeating that "the ChatGPT moment for robotics is fast approaching."
Whether that moment has arrived is exactly what the money is betting on, and what Tesla's numbers do not yet settle. Investors have long funded demos; the harder test, as several robotics veterans note, is dexterous manipulation and reliable real-world deployment, neither of which a first production line proves. Musk has floated merging Tesla with his rocket company SpaceX into a roughly $3.5 trillion group to pool AI, chips and robotics work, a sign of how much he is willing to reorganize around the bet even as this quarter's profit shows its near-term cost.