← All articles

Tesla says it started producing the Optimus robot in California as profit fell 43%: what is confirmed and what is Musk's forecast

Topics: Robots, Small business

A single neutral-shaped humanoid robot stands on a bright car-factory assembly line at an early stage of production

Short answer: Tesla says it has started producing the Optimus humanoid robot at its Fremont factory, in space that used to build electric cars, and the company itself calls these units its "first production-quality bots." Elon Musk told investors on the earnings call that this is "early production" and said Tesla should be able to make 20,000 robots a week by the end of 2027: that is a forecast from the company's CEO, not a confirmed fact. In the same quarter, Tesla's net income fell 43%, to $787 million, while vehicle revenue stayed roughly flat. Below: what in the announcement is backed by numbers in the report, what rests only on Musk's words, and how to treat the timelines Tesla names, if you are weighing robotics for your own company.

Where this comes from

Source: the TechCrunch article "Tesla starts Optimus robot production in California as profit falls 43%," published October 21, 2026, and updated with details from Tesla's third-quarter earnings call.

What Tesla announced about production

Tesla wrote to shareholders in its third-quarter letter: "The bots coming off this line are the first production-quality bots that we have manufactured... While we will continue to iterate on the hardware and manufacturing process, we have proven that the current manufacturing process for Optimus is suitable for a commercial product" (the ellipsis here marks a cut made in TechCrunch's own quote). Production is happening at the Fremont, California factory, in a space once dedicated to building electric vehicles.

On the call, Musk called the current stage "early production" and warned that the ramp will be slow and "somewhat bumpy because it is such a radically new product." He also said: "There's still some things we need to tweak about the design, because you need to deliver a robot that is safe and reliable and has long duration, so it can't, like, wear out after a month. And there's still quite a lot of work to do with regard to the AI and the robot for it to be truly useful for open-ended tasks." TechCrunch does not say how many Optimus units Tesla had actually built by the time of the announcement, does not name a purchase price for Optimus, does not disclose manufacturing cost, and does not name a single customer.

The 20,000-robots-a-week forecast: whose number is it

The figure of 20,000 robots a week by the end of 2027 belongs to Musk personally; he gave it on the same call. It is a statement by the company's CEO about the future, not a report of a finished plan, and not a number TechCrunch verified independently. Timelines Tesla has named on these calls have a history of slipping, though in a different Tesla project, not Optimus. Back in April, on the call for the previous quarter, Musk said the unsupervised version of the autopilot software for personally owned vehicles would "probably" ship to consumers in the fourth quarter of 2026. On the October 21 call, Tesla's VP of AI software, Ashok Elluswamy, named a different, narrower milestone instead: turning the feature on for the first customer vehicles joining the robotaxi fleet, which he expects "somewhere around Q2 next year." That is not the same milestone, and TechCrunch does not claim the April date was directly pushed to this one. But both times the topic was the same unsupervised autopilot software, and both times Tesla named a date that, at the time of the announcement, had not yet arrived. This does not prove the 20,000-robots forecast will also miss its mark, but it shows that timelines Tesla names on these calls are worth treating as the company's plan, not as something that has already happened.

What the financial results showed

Tesla's third-quarter revenue came to $28.04 billion: a touch below the $28.09 billion from the same quarter of 2025, and about $200 million below the previous quarter's $28.23 billion. EV sales stayed flat year over year at $20 billion: Tesla delivered 486,532 vehicles, about 6,000 more than the previous quarter but fewer than the same quarter a year earlier. Solar and energy storage brought in 16% less year over year, $2.8 billion: the company blamed tariff-related customer refunds. Services revenue, which includes Full Self-Driving subscriptions among other things, grew 33% year over year to $4.6 billion and delivered a record $734 million in gross profit for that segment.

At the same time, Tesla's net income came to $787 million: a 43% drop from the same quarter of 2025, when profit was $1.37 billion. TechCrunch writes that none of the growth in services revenue outpaced the company's combined spending on AI projects, the Optimus robot, and autonomous driving, and that this spending "took a big bite" out of profit. The article does not break the spending growth down by each of the three areas separately, so it is not possible, from this article alone, to say how much Optimus specifically costs. Lower sales of so-called regulatory credits added further pressure on profitability: this is one of Tesla's revenue sources, and there was less of it this quarter than before. TechCrunch does not explain the mechanics of these credits in more detail, and I won't attempt to describe it more precisely than the source does.

What to take from this if you are thinking about robots for your company

If you follow humanoid-robot news to decide when, or whether, to look at them for your own business, this story is worth splitting into two layers. The first layer: Tesla itself, in its shareholder letter, states that it has started the production line and that the first units have come off it. That is a documented statement by the company about itself, not independent confirmation from outside: TechCrunch does not name a single outside source who saw these units or verified their production. The second layer: the pace and the timeline, named only by Musk on the call, which for now rest on nothing but his words, and, as the autopilot example shows, similar Tesla timelines have been named before and, at the time they were announced, had not yet arrived.

A practical takeaway for any company looking at any robot vendor, not just Tesla: an announcement that production has started, and a forecast for production volume, are not by themselves a reason to plan a purchase. Before budgeting for a robot like this, ask the vendor for four things in writing, not just verbally at a presentation. First: the current price per unit, today. Second: a delivery date for your specific order, not an abstract date for the start of mass production. Third: the terms of service if the robot breaks down. Fourth, separate from the first three: a demo or pilot on your actual task, using your own items, in your own space, with your own people present, not just a general demo of the robot doing someone else's task that you saw at a presentation or in a video. Before watching that demo, agree with the vendor in writing on what will count as success: how many of your items in a row the robot must handle without an error, in how much time, and what counts as an error (dropped it, damaged it, failed to pick it up, stopped and waited for a person). Only count the result if the robot met exactly those conditions on its own, without hidden help from the vendor's engineer, and you were there to see it: a demo shown only on video, or only described by the vendor, does not substitute for this kind of check. If the robot fails your task, that is a reason not to buy it now, not a reason to trust a promise that it will learn later. If a vendor cannot answer all four questions in writing, that is a reasonable sign the offer is not yet ready for an actual purchase, and the decision is worth postponing until the next announcement, not forcing to fit your own deadline.

Bottom line

Tesla itself stated in its shareholder letter that it has started producing Optimus at the Fremont factory and considers the current process suitable for a commercial product; the article offers no independent confirmation of this. The forecast of 20,000 robots a week by the end of 2027 belongs to Musk personally and is not confirmed beyond his own words. Timelines for another Tesla project, autonomous driving, already have a history of naming dates that had not arrived by the time they were announced. In the same quarter, Tesla's net income fell 43%, to $787 million, while vehicle revenue stayed roughly flat.

I work on AI agents and automation. If you'd like to see my projects or talk about your own task, take a look at my portfolio.

Sources