Tesla has started Optimus production at its Fremont, California factory, and it is leaning on the humanoid robot just as the core business gets squeezed. The Times of India described the moment as Tesla pressing the accelerator on the product Elon Musk predicted would deliver 80% of the company’s value, with costs jumping and profit falling 43%.
We covered the first announcement in our report on the Optimus production start and the 43% profit drop. This piece takes the hardware angle. Tesla Optimus production is now a factory problem, not a demo problem, and the numbers Musk has attached to it are easy to test against basic manufacturing arithmetic.
Below we separate what Tesla has confirmed from what remains a prediction, run the math on the 20,000-robots-per-week goal, look at the parts that usually break humanoid programs, and set out five predictions for the next 14 months. Our category hub for hardware coverage tracks the wider story.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What Tesla Announced About Optimus Production in Fremont
Tesla said it has started producing Optimus at its Fremont factory. TechCrunch and Business Insider both reported on October 21, 2026 that the line is running during the third quarter, and CNBC covered the same earnings release. Tesla’s own description, as reproduced by Tesla North, is that Optimus production began on high-volume lines at Fremont.
In the same Q3 update, as reproduced by Tesla North, Tesla said: “The bots coming off this line are the first production-quality bots that we have manufactured.” That phrase matters. Tesla has shown prototypes for years, but a production-quality unit means a part built to a repeatable drawing, on tooling meant to make thousands more.
Musk framed the start as modest. On the Q3 earnings call, as carried by Yahoo Finance, he said: “Weâve started what Iâd call early production on our high-volume line in Fremont.” He added that the ramp would be slow and “somewhat bumpy because it is such a radically new product.”
Tesla gave no output figure. It published no unit count, no weekly rate, no price, and no customer delivery date for units leaving the line today. Reports also leave open which Optimus generation Fremont is building. Some coverage points to Gen 3 as the intended first mass-production version, but Tesla has not said so publicly.
The 80% Prediction: What Musk Said and When
The headline number comes from an X post dated September 1, 2025. Musk wrote: “~80% of Teslaâs value will be Optimus.” Fortune covered the post the next day and pointed out that the robots were still not on the market. Axios ran the same claim under a headline that stated it as a direct forecast.
Musk made the remark in connection with Master Plan IV, a roadmap that sits alongside scaling FSD and Optimus. It was a forward-looking view of what Tesla might become. It was not a statement of current financials, and no filing or earnings release has shown Optimus contributing revenue at anything near that scale.
That distinction gets lost in headlines. Optimus today has no disclosed revenue, no disclosed profit, and no disclosed valuation. The fact sheet we worked from is explicit on this point: any claim that Optimus already represents 80% of Tesla’s value is unconfirmed. The statement describes a destination, and the Fremont line is the first physical step toward it.
It is also a striking claim on its own terms. If 80% of the company’s value sits in one product, then the cars, the energy business, and FSD together carry the remaining 20%. Tesla currently earns its money from those other lines, which is why the production ramp is under such close scrutiny.
Confirmed vs Unconfirmed: A Fact Check on Optimus
Readers searching for Tesla Optimus production news will find a mix of confirmed events and executive aspiration. The table below sorts the main claims by how well they are supported, using the reporting from Tesla’s Q3 2026 update and coverage by TechCrunch, Business Insider, CNBC, and Fortune.
| Claim | Status | Basis |
|---|---|---|
| Optimus production started at Fremont | Confirmed | Tesla statement, reported by TechCrunch and Business Insider |
| Musk predicts ~80% of Tesla’s value will be Optimus | Prediction | Musk X post, September 1, 2025 |
| Optimus already makes up 80% of Tesla’s value | Unconfirmed | No filing or earnings data supports it |
| Target of 20,000 robots per week by end of 2027 | Stated goal | Musk, per Q3 2026 coverage |
| Current weekly output or total units built | Not disclosed | No figure published |
| Retail or enterprise price | Not disclosed | No figure published |
| External deliveries near end of 2027 | Stated expectation | Musk, industrial use first |
| Generation being built in Fremont | Not specified | Tesla has not said publicly |
Four of the eight rows are non-answers. That is normal for a program in its first weeks, but it limits how much weight any investor or reader can put on the 80% framing today.
Reading the 20,000-Per-Week Target
Musk said Tesla could produce 20,000 Optimus robots per week by the end of 2027. The number sounds abstract until you convert it. Across 52 weeks it works out to about 1.04 million robots a year, which matches the one-million-unit annual capacity Tesla has described for the line. We covered the target earlier in our piece on why 20,000 a week is the number Optimus sales depend on.
The table below breaks the target into shop-floor terms. These are our own calculations from the 20,000 figure, not Tesla-published rates, and they assume the weekly total is spread evenly.
| Time basis | Robots produced | Assumption |
|---|---|---|
| Per week | 20,000 | Musk’s stated target |
| Per year | About 1,040,000 | 52 weeks at target rate |
| Per day (7-day week) | About 2,857 | Continuous operation |
| Per day (5-day week) | 4,000 | Weekday-only operation |
| Per hour (24/7) | About 119 | 20,000 divided by 168 hours |
| Time per robot (24/7) | About 30 seconds | Across the whole plant, not one station |
One finished humanoid every 30 seconds, around the clock, is a pace that most car plants reach only after years of tuning. Each robot has far more actuated joints than a car has moving subassemblies, so the part count behind that cadence is large. Tesla says it needs at least a year or longer to ramp, and that timeline looks realistic against this math.
Inside the Fremont Line: Repurposed Car Space
Reports say the Optimus line occupies factory space that previously built electric vehicles. That choice gives Tesla a head start on power, floor space, logistics, and trained staff. It also signals that Tesla treats the robot as a manufacturing product from day one, with the same line-balancing discipline it applies to cars.
Musk said on the earnings call, as transcribed by the Singju Post: “We do have the vast majority of a million-unit-per-line production line landed in Fremont right now.” In plain terms, most of the equipment for a line sized at one million robots a year is already on site. Equipment on the floor is not output, however. Tooling has to be commissioned, validated, and tuned, and each of those steps usually surfaces defects nobody planned for.
Tesla’s first-line approach also carries a trade-off. A line sized for one million units a year is expensive to build before demand is proven. Musk has said outside deliveries should start near the end of 2027, with industrial buyers first. Until then, early Optimus units mostly serve Tesla’s own factories, which gives the company a captive test bed and a way to find reliability problems before any customer does.
Where Humanoid Hardware Gets Hard
The Q3 2026 discussion kept returning to the same few engineering areas. Tesla identified electromechanical design, reliability, operating duration, and AI performance as issues it expects to address around the end of 2027. That list is a useful map of where schedule risk lives.
Hands: The Hardest Subsystem
Tesla’s Optimus program lead said the hand is among the most challenging hardware areas, even though much of the rest of the electromechanical system has reached relatively high confidence. Hands pack many small actuators, sensors, and tendons into a tiny volume, and they take the most wear. A robot that cannot grip reliably cannot do the factory work Tesla has promised to outside buyers.
Actuators and Batteries
The earnings discussion also stressed actuator complexity, and robot operating duration puts pressure on battery integration. A humanoid draws power at every joint while it stands still, so runtime depends on both pack size and motor efficiency. The reports we reviewed gave no confirmed component volumes or named suppliers, so any supplier-by-supplier forecast would be guesswork.
A Supply Chain Still Being Built
Musk described a large supply chain that is being industrialized while design changes, production-system work, AI software, and factory scale-up all run at the same time. Changing a part design after the tooling is cut is expensive. Tesla is choosing speed over a frozen design, which explains both the ambition and the warning about bumps.
Timeline: From a 2025 Post to 2027 Deliveries
The sequence of dates below shows how little time separates the first Fremont units from the 20,000-per-week goal. All dates come from Musk’s statements and the Q3 2026 coverage, and each future date depends on Tesla’s plan holding.
| Date | Milestone | Status |
|---|---|---|
| September 1, 2025 | Musk posts that ~80% of Tesla’s value will be Optimus | Done |
| Q3 2026 | Early production begins on Fremont high-volume line | Done |
| October 21, 2026 | Q3 earnings coverage discloses production start | Done |
| Around April 2027 | AI5 chip expected to be integrated into the robot | Planned |
| Near end of 2027 | First deliveries to outside companies, industrial use first | Planned |
| End of 2027 | 20,000 robots per week target | Target |
From the October 2026 start to the end-of-2027 target is roughly 14 months. Tesla would need to go from early production, with undisclosed output, to a rate near 20,000 a week in that window. Few manufacturers have achieved a ramp of that steepness with a brand-new product category.
The Money Side: Profit Down 43% While Optimus Ramps
TechCrunch’s headline on the Fremont start carried the profit figure with it: Tesla began Optimus production as profit fell 43%. The Times of India added that costs jumped. The two facts belong together, because a program this large spends cash well before it earns any.
The spending picture is mixed. We reported that Tesla was trimming capital spending while a SpaceX merger loomed, which makes the one-million-unit Optimus line a bigger share of a tighter budget. Tesla has not said how much the Fremont line costs, and we found no confirmed figure in the Q3 reporting.
Musk’s own framing points the same way. The Autopian wrote that Musk does not see cars as a particularly valuable part of Tesla’s future, and TIKR ran a preview in which Optimus and Cybercab were described as the vast majority of Tesla’s value heading into earnings. Investors are being asked to value future robots against present-day margin pressure, and the Q3 numbers show how large that gap has become.
AI5 and the Brain Problem
Hardware is only half of Optimus. Musk said Tesla expects to integrate its AI5 chip into the robot around April 2027, subject to the development plan staying on schedule. That date lands about eight months before the end-of-2027 target, which leaves little room for slips in either the chip or the software that runs on it.
The chip story connects to Musk’s wider silicon ambitions. We examined the Terafab naming dispute that clouds Musk’s chip plan, which shows how much of the robot and vehicle roadmap depends on custom compute arriving on time. A humanoid that cannot perceive and plan reliably is an expensive set of motors, so AI performance sits on the same critical path as hands and actuators.
Tesla listed AI performance among the problems it expects to resolve around the end of 2027. That timing is the same moment it hopes to ship to outside buyers, so there is no buffer between solving the problem and selling the product.
Competitive Comparison: Who Else Is Building Humanoids
Tesla is not alone. The relevant comparison set includes Figure, Unitree, Boston Dynamics, Agility Robotics, and XPeng’s humanoid effort. Our research found no comparable October 2026 production volumes, deployment counts, or analyst estimates for any of them, so we will not rank them or invent market-share numbers.
What we can compare is the type of disclosure each camp gives. Tesla has an announced factory, a stated weekly target, and a dated chip plan, but no unit count. Startups tend to publish pilot deployments instead and rarely commit to a factory-scale number. Capital also keeps flowing into the field, as shown by Enact Intelligence’s $30M pre-seed for robotics, so Tesla’s rivals are being funded to attack the same gaps in hands and AI.
Tesla’s distinct edge is manufacturing infrastructure. It already runs high-volume factories, buys components in bulk for cars, and can repurpose floor space. Its distinct weakness is that the target is public and dated, which makes every miss visible.
Historical Context: Musk Timelines and the FSD Parallel
Musk’s schedules have a record, and readers weighing the 2027 goal should look at it. Tesla’s own FSD program offers the closest parallel. We reported that Tesla pushed unsupervised FSD to Q2 2027 and limited it to robotaxi zones, a reminder that autonomy promises have moved before.
Musk’s language about Optimus is more guarded than his FSD talk has sometimes been. He said production is early, the ramp is slow, and it will be somewhat bumpy. That is a different tone than a flat promise, and it gives Tesla room to report progress against a lower bar than 20,000 a week.
Fortune’s September 2025 framing is still the right baseline. When Musk made the 80% claim, the robots were not on the market. In October 2026 they are being built, but they are still not on the market. The gap between building and selling is where most of the risk remains.
Market Impact: What Investors and Suppliers Should Watch
For investors, the Fremont start changes the question from “will Tesla build it” to “how fast, and at what cost.” Tesla has confirmed neither a unit count nor a price, so the stock case rests on management’s long-range framing rather than on reported results. The shift to Optimus as the core thesis also raises the stakes of every quarterly update.
For suppliers, the signal is a possible demand spike in actuators, sensors, batteries, and precision machining if the ramp holds. The reports gave no named suppliers or component volumes, so we cannot name winners. Vendors that sell motors, reducers, and compact power electronics to robotics customers will still watch Tesla’s order flow closely.
The corporate backdrop adds another layer. Our coverage of the SpaceX-Tesla merger and its hardware stack shows how many large decisions are in motion at once. Any merger outcome could change how capital is allocated to the robot program.
Five Predictions for Optimus Through 2027
These are our judgments, not reported facts. We built them from the confirmed timeline and the engineering issues Tesla itself has named.
- Tesla will report progress in milestones, not unit counts. Expect statements about line commissioning and internal deployment well before any weekly figure appears.
- The hand will set the schedule. Tesla’s own program lead called it among the hardest areas, so a hand redesign is the most likely source of delay.
- The AI5 integration date near April 2027 will be a key checkpoint. A slip there would compress the final months before the end-of-2027 target.
- The 20,000-per-week rate will not be reached on the stated date. A 14-month jump from early production to that pace would be unusual for any new product. Tesla may still claim partial progress.
- The 80% claim will stay a narrative, not a financial fact. Unless Tesla starts reporting Optimus revenue separately, analysts will keep treating the figure as a long-range vision.
The takeaway: Tesla Optimus production is real, early, and unquantified. The company has proven it can start a line. It has not yet proven it can fill one.
FAQ: Tesla Optimus Production and the 80% Claim
Has Tesla started producing Optimus robots? Yes. Tesla said it has begun production at its Fremont, California factory, and TechCrunch and Business Insider reported the start on October 21, 2026. Musk called it early production with a slow, bumpy ramp.
Does Optimus make up 80% of Tesla’s value? No evidence says so. Musk predicted in a September 1, 2025 X post that about 80% of Tesla’s value will be Optimus. That is a forecast, and Tesla has disclosed no Optimus revenue, profit, or valuation.
How many Optimus robots will Tesla build? Musk said Tesla could produce 20,000 per week by the end of 2027, which is about one million a year. Tesla has not disclosed how many it has built so far.
When can companies buy Optimus? Musk said deliveries to outside companies should begin near the end of 2027, starting with industrial uses. No price or order process has been published.
What is the hardest part of building Optimus? Tesla’s Optimus program lead said the hand is among the most challenging hardware areas. The company also named electromechanical design, reliability, operating duration, and AI performance as open issues.
Which Optimus generation is being built in Fremont? Tesla has not said. Some reports describe Gen 3 as the intended first mass-production version, but that is not an official confirmation.
What chip will Optimus use? Musk said Tesla expects to integrate its AI5 chip into the robot around April 2027, if the development plan stays on schedule.
