Musk tempers Optimus timeline: outside deliveries 'possible' by end of 2027, 20,000-a-week ramp years away
Musk said 20,000 Optimus units a week is 'at least a year, maybe a little longer' out - and downplayed imminent outside sales on Tesla's Q3 call.
Elon Musk moved to temper expectations for Tesla's Optimus humanoid robot on the company's third-quarter earnings call on October 21, telling investors that deliveries to customers outside Tesla are only 'possible' by the end of 2027 - not a firm commitment. Tesla shares reversed earlier after-hours gains and ended about 4% lower as Musk pushed the commercial timeline further out and cut the year's capital spending guidance to roughly USD 20 billion.
The headline production target is 20,000 Optimus units a week - about a million a year - but Musk said reaching that pace will take 'at least a year, maybe a little longer.' He described the ramp as an S-curve 'with a shallow slope' at the start, warning that the robot contains roughly 10,000 components and that the ramp 'will grow as fast as the slowest component.'
The hardware starting point is real: Tesla has begun 'early production' of production-quality Optimus units on its high-volume line in Fremont, and says the vast majority of equipment for a million-unit-a-year line is already installed. But the AI, the electromechanical design, safety and durability all still need work - issues Musk said he expects 'to be solved by the end of next year.'
Musk called the scaling challenge 'very substantial,' describing Optimus as 'going to be the hardest product to scale manufacturing that we've ever made at Tesla, because everything on the robot is new.' Unlike Tesla's vehicles, Optimus has no existing fleet of millions of units generating training data, so the company must bootstrap learning from scratch with a self-play training fleet.
The shift in tone is notable. Musk has repeatedly called Optimus the greatest product in human history and a future industry larger than cars. With early production underway but outside sales still only 'possible' fourteen months out, the question hanging over the program is no longer whether Tesla can build a humanoid robot - it is whether anyone can build one at car-industry scale, on a car-industry schedule.