News

Tesla Begins Robot Production – 20,000 a Week

October 25, 2026
editor

October 25, 2026 09:00 AM PST

(PenniesToSave.com) – Tesla has started building its Optimus humanoid robot on a production line inside its Fremont, California, factory, in space that turned out the Model S and Model X until this spring [5]. The company told shareholders that the robots coming off the line are the first production-quality units it has ever made [1]. CEO Elon Musk now says Tesla plans to begin delivering Optimus to outside companies around the end of 2027, starting with dangerous, repetitive, and around-the-clock jobs [5].

The announcement arrived alongside a quarter that showed the cost of chasing that future. Tesla reported third-quarter revenue of about $28 billion, roughly unchanged from a year earlier, while profit fell 43% as spending on robotics, computing, and autonomous driving climbed [3]. Shares rose 3.41% during Wednesday’s regular session, then slid 3.32% in after-hours trading once the results were out [4]. For anyone brushing up on the basics of investing in individual stocks, that kind of swing is a useful reminder that bold promises and steady returns are not the same thing.

Behind the headlines sit practical questions. How will these robots learn to do useful work, how close are the competitors, and what could a robot workforce mean for jobs, households, and the investors betting billions on it?

What Did Tesla Actually Announce on Its Earnings Call?

Tesla’s third-quarter shareholder letter described the Fremont output as the company’s first production-quality Optimus robots and said Tesla had proven its current manufacturing process is fit for a commercial product [5]. The line sits where Tesla ended Model S and Model X production earlier this year, and it is designed to eventually build a million robots a year [1]. Teslarati reports that Tesla cleared the old vehicle line in just 46 days [5]. On Wednesday’s earnings call, Musk said most of the equipment needed for that capacity is already in the building [5].

Reaching full speed will take longer. Musk said getting to 20,000 robots a week will take at least a year and possibly a little longer [4]. Other outlets framed the timeline slightly differently: Interesting Engineering reported the target as 20,000 units per week by the end of 2027 [3], while TechRadar described it simply as by 2027 [2]. The fuller wording from the call points to a ramp that could stretch beyond twelve months.

Texas is central to the longer-term plan. A far larger Optimus factory is rising at Giga Texas, where drone footage showed the steel frame nearing completion, and Musk has set a long-term goal of 10 million robots a year there [5]. Business Insider likewise reported an Austin line planned for 10 million robots annually [1]. Separately, Interesting Engineering reported that equipment commissioning at Tesla’s Austin fabrication facility should begin in 2027, though that report does not make clear whether it refers to the same robot plant [3].

The supply chain is taking shape too. In September, Tesla certified Tuopu, Joyson, and Sanhua as mass production partners for Optimus components, and executives said Optimus production costs are trending toward the expected manufacturing cost of the Cybercab [5]. Teslarati read that as a sign the robot is being engineered for volume from the start [5].

How Will Optimus Learn to Do Real Work?

Tesla’s cars give the company a built-in advantage in artificial intelligence, since millions of vehicles send real driving data back to its engineers. Optimus has no such fleet, a gap Musk acknowledged directly on the call [5]. Tesla’s answer is what it calls Optimus Academy, where 10,000 to 20,000 robots will practice a wide range of tasks and feed what they learn into the next version of the AI model [5][6]. That training group is a separate figure from the 20,000-per-week production target, even though the numbers happen to match. Musk also suggested that at some point the robots would be used to help build other robots [6].

New hardware is part of the plan. Tesla aims to begin installing its AI5 chip in Optimus around April 2027 [6], and Musk said the chip delivers roughly ten times the computing power of the current AI4 [4]. The company also expects to reach 400 megawatts of training computing capacity by the end of 2026, shared between its vehicles and its robots [3].

Musk was frank about the limits. He acknowledged that Optimus currently struggles to complete varied tasks on its own [2], and Interesting Engineering reported that considerable work remains before the robot can handle a broad range of activities independently [3]. The hands remain the toughest piece of the design and the focus of much of the current engineering effort [1][5].

Tesla is not the only company chasing robots that teach themselves. TechRadar, citing the South China Morning Post, reports that rival Unitree is developing an AI system meant to let its humanoids improve by writing their own code and simulations, an effort called Unitree Rehearsal [2]. Founder Wang Xingxing argued that hand-coding by engineers is relatively inefficient, though he said people will still provide feedback and judge performance [2].

Is the Promise Running Ahead of the Proof?

Musk has been unusually candid about the road ahead. He said early production will be somewhat bumpy because roughly 10,000 things have to go right to build each robot [1]. He described the growth of Optimus as an S curve, adding that “the initial curve is a shallow slope” [5]. Tesla says it is still refining the robot’s AI, electromechanical design, reliability, and service life, work Musk expects to wrap up by the end of 2027 [5]. Interesting Engineering went further, reporting that Tesla has yet to establish when Optimus will be commercially available at scale [3].

The newest version of Optimus also has not been shown in public. Musk has said Tesla kept it under wraps to prevent competitors from copying the design [1]. He now promises a video within a month or two showing the robot performing tasks that are not pre-programmed [6]. Earlier Optimus demos featured the robots dancing and handing out candy, which Investopedia notes did not involve advanced AI [6]. Musk has criticized rival demos that run scripted routines [4], and that same standard will apply just as fairly to Tesla’s own footage.

A few other details call for caution. A report syndicated on Yahoo Finance says Tesla placed its first large-scale component order for roughly 5,000 Optimus units, a claim Tesla has not confirmed [4]. TechRadar also notes that the company’s recent flying Roadster demo drew questions about how realistic those features are [2]. Announcements generate headlines, but delivered products are what ultimately earn trust and justify the money behind them.

Who Else Is Racing to Build Humanoid Robots?

Tesla is entering a crowded field. Agility Robotics, which plans to go public, has deployed its Digit robot across nine customer facilities [1]. Figure AI has announced deployments in logistics and distribution centers this year [1]. According to TechRadar, Figure’s production run of its Figure 03 robot has reached four figures, work continues on a fourth-generation model, and its fingertip tracking now reaches millimeter-level precision [2]. CEO Brett Adcock told employees the company’s roadmap has completely changed over the past year [2].

Figure is also gathering training data at scale. TechRadar reports that about 37.5 million videos have been uploaded to Figure’s Index app, which records human activities to teach robots, though the outlet presented that figure with a hedge [2]. One Figure 03 recently walked on a treadmill for nine straight days at company headquarters [2].

Some competitors are aiming straight for the home. Sunday Robotics, 1X, and Weave Robotics are preparing to ship robots into homes this fall [1], and Unitree’s Wang said his company’s robots could be cleaning rooms as early as next year [2].

Musk said he would be “stunned” if any rival had a robot that could compete with Optimus when deliveries begin.

Q3 2026 earnings call, via Yahoo Finance [4]

Musk dismissed the competition outright, saying he does not expect any rival to field a comparable robot by the time Optimus reaches customers [4]. Confidence aside, a field this busy works in the public’s favor. When privately funded companies compete for the same customers, quality tends to rise and costs tend to fall faster than any single player could manage alone.

What Could Humanoid Robots Mean for Jobs and Daily Life?

Tesla is pointing Optimus first at the toughest work. Musk said the earliest customers are likely industrial companies with dangerous or repetitive jobs [4], and Teslarati reports that deliveries will target work that is dangerous, repetitive, and runs around the clock [5]. If robots take on hazardous tasks, the most immediate benefit may be fewer people doing work that puts their health and safety at risk.

For households, expectations should stay grounded. Tesla’s first deliveries are aimed at businesses, not living rooms [4][5]. TechRadar’s writer suggested that a year from now, consumers might be able to buy a robot to tidy up and do laundry, but that is the author’s own forecast rather than a company commitment [2]. The nearer-term home robots are coming from smaller players such as 1X and Weave [1].

There is also a domestic manufacturing story here. Tesla repurposed a California electric vehicle plant for robot production and is building a massive new factory in Texas [1][3][5]. That kind of investment keeps advanced manufacturing, and the skilled jobs that come with designing, building, and servicing machines, on American soil.

The long-term vision is enormous. Musk predicted on the call that humanoid robots could one day outnumber humans on Earth [6]. Not everyone finds that reassuring. Online commenters reacting to Unitree’s self-improving robot plans warned of a future in which machines run code people do not understand, comparing it to the film The Matrix [2]. Those concerns deserve a hearing, and the answer will come from how carefully these companies test and supervise what they build.

How Is Wall Street Pricing Tesla’s Robot Bet?

Tesla’s quarter can be read two ways. Interesting Engineering reported revenue of $28.04 billion, nearly unchanged from $28.09 billion a year earlier [3]. Yahoo Finance reported $28.05 billion and noted that it topped the Wall Street consensus of $27.88 billion, with adjusted earnings of 48 cents per share beating expectations of 43 cents [4]. At the same time, profit fell 43% as expenses rose [3], and automotive revenue slipped 3% to $20.56 billion [4]. Tesla cleared a modest bar while its core car business shrank.

Tesla’s third-quarter profit fell 43% as spending on robotics, computing, and autonomous driving climbed.

Interesting Engineering [3]

Batteries remain a constraint on vehicle growth. Tesla is expanding battery pack capacity in Berlin, cathode production and lithium refining in Texas, and lithium iron phosphate cell manufacturing in Nevada, and it has shipped its first Megapack storage systems built entirely with its own cells [3]. Production of its 4680 cells is rising to support the Cybertruck, the Semi, and higher Model Y volumes [3].

Investors are being asked to look past the car business. The week before earnings, Musk said Optimus and the Cybercab will make up the vast majority of Tesla’s value [4]. Gene Munster of Deepwater Asset Management said most investors still treat Optimus as optional upside, which he described as a call option carrying enormous value, and he estimated the ramp could imply about 1 million robots in 2028 [4]. Investopedia also reports that special committees at Tesla and SpaceX are working out the terms of a possible merger, while Musk has pledged to stay out of the talks [6].

A valuation that rests increasingly on a product not yet for sale is a bet, not a certainty. Spreading money across many companies through low-cost exchange-traded funds is one way to keep any single bet from dominating a portfolio.

Final Thoughts

Tesla’s Optimus milestone is real. Robots are coming off a production line in Fremont, a larger factory is going up in Texas, and suppliers have been lined up for volume [1][5]. What remains unproven is whether the robot can do useful, unscripted work reliably enough for paying customers.

The next markers are clear. Musk has promised a video of Optimus performing tasks that are not pre-programmed within a month or two [6]. First deliveries to outside companies are targeted for the end of 2027 [5], and reaching 20,000 robots a week will take at least a year, by Musk’s own estimate [4]. Each quarterly report between now and then will show how much Tesla is spending to get there and whether its car business can carry the load [3].

For most households, the practical takeaway is simple. A Tesla robot is not headed for the living room in 2027 under the current plan. The nearer impact will show up in industrial workplaces and in the portfolios of anyone holding the stock. The sensible approach is to judge Optimus the way any product should be judged: by what it delivers, not by what is promised.

Works Cited

[1] Jetha, Rya. “Tesla Targets Shipping Optimus to Customers by the End of 2027.” Business Insider, 21 Oct. 2026, www.businessinsider.com/tesla-starts-producing-optimus-robots-at-its-fremont-factory-2026-10.

[2] Nield, David. “Unitree Says Its ‘Self-Evolving Robots’ Are the Next Big Leap for Humanoids, as Tesla Optimus Production Ramps Up to Hit 20,000 Robots a Week.” TechRadar, 24 Oct. 2026, www.techradar.com/ai-platforms-assistants/unitree-says-its-self-evolving-robots-are-the-next-big-leap-for-humanoids-as-tesla-optimus-production-ramps-up-to-hit-20-000-robots-a-week.

[3] Khollam, Aamir. “Tesla Fires Up Optimus Robot Production with Musk Eyeing 20,000 Units per Week by 2027.” Interesting Engineering, 21 Oct. 2026, interestingengineering.com/ai-robotics/tesla-optimus-robot-enters-production.

[4] Nadig, Radhika Anilkumar. “Elon Musk Says He’d Be ‘Stunned’ If Any Robot Rivals Optimus at Launch: Gene Munster Calls It a Tesla ‘Call Option’ with ‘Massive Value.'” Yahoo Finance, 21 Oct. 2026, finance.yahoo.com/technology/ai/articles/elon-musk-says-hed-stunned-052803880.html.

[5] Gene. “Tesla Reveals Optimus Academy, Where Robots Teach Themselves.” Teslarati, 23 Oct. 2026, www.teslarati.com/tesla-reveals-optimus-academy-where-robots-teach-themselves/.

[6] Kim, Crystal. “Elon Musk Hints at Exciting New Capabilities for Tesla’s Optimus Robots.” Investopedia, 22 Oct. 2026, www.investopedia.com/market-update-musk-teases-new-sizzle-for-teslas-optimus-robots-tsla-spcx-12188240.