Tesla has officially moved its Optimus humanoid robot from the lab to the assembly line. In its third-quarter 2026 results, the company said early production has started on a high-volume Optimus line at its Fremont, California factory, and described the units coming off it as the first production-quality bots it has ever built. The milestone arrives alongside a sharp drop in profit, underlining how much money Tesla is now pouring into AI and robotics.
What Tesla Announced
Tesla disclosed the production start in its Q3 shareholder letter on October 21, and Elon Musk expanded on it during the earnings call the same day. Musk characterised the current phase as early production and warned that the ramp would be slow and somewhat bumpy, describing an S-curve with a shallow early slope because the product is new and parts of the design are still changing.
The shareholder materials went further than past Optimus updates. Tesla said the manufacturing process now in place is suitable for a commercial product, a claim it had not been able to make during earlier prototype phases. Production begins with Optimus Gen 3, the long-teased redesign whose public unveil had been held back for what the company called finishing touches.
Key details from the announcement and call:
- Location: the line sits in space freed by dismantling the Model S and Model X assembly lines at Fremont earlier this year, a teardown Tesla completed in 46 days.
- Planned capacity: the Fremont line is designed for roughly 1 million robots per year, and Musk said most of its equipment has already arrived.
- Ramp target: reaching about 20,000 units per week is expected to take a year or more.
- Next step: a far larger line at Giga Texas, planned for 10 million units a year, with Tesla aiming to shift to a Gen 4 design by the time it comes online.
Inside Optimus Gen 3
Reporting on the production design points to cleaner joint covers and more human-like hands with 22 degrees of freedom, a significant step given that dexterous manipulation has been one of the hardest problems for every humanoid developer. Musk acknowledged that the robot still needs design tweaks to guarantee durability and safety, and that substantial AI work remains before Optimus can reliably handle open-ended tasks rather than scripted routines.
Compute is another piece of the roadmap. Tesla expects to integrate its AI5 chip into Optimus around April 2027. AI5 is pitched as delivering roughly five times the compute of a full two-chip AI4 computer, with trial production already under way at Samsung's Texas fab and mass production slated for 2027. Musk also said a new Optimus video could arrive within one to two months, and that he wants it to show the robot doing tasks that are not pre-programmed rather than a choreographed demo.
Timeline to Customers
Despite the production headline, outside buyers will have to wait. Tesla expects remaining issues across AI, electromechanical design, reliability and durability to be resolved around the end of 2027, and plans to begin delivering robots to external customers at roughly the same time. The first buyers are expected to be industrial operators that want machines for dangerous, repetitive or round-the-clock work, with home robots following once volumes scale.
Until then, the early Fremont output is likely to be used internally, giving Tesla real-world data from its own factories, a strategy other humanoid makers have also pursued before opening commercial sales. Tesla has not published a price for Optimus.
The Financial Backdrop
The robot news landed in a weak quarter for Tesla's core business. According to TechCrunch, net income fell 43% to $787 million, down from $1.37 billion a year earlier, as spending on AI and Optimus rose and regulatory credit sales declined. Total revenue was essentially flat at $28.04 billion. Automotive revenue held around $20 billion on 486,532 deliveries, energy storage and solar revenue fell 16% to $2.8 billion, and services revenue climbed 33% to $4.6 billion.
That mix makes Optimus a central part of Tesla's investment story. Musk has repeatedly argued that humanoid robots could eventually outweigh the car business, and committing a former flagship vehicle line to robot production is the most concrete evidence yet that the company is acting on that belief.
Why It Matters
Tesla starting a dedicated, high-volume humanoid line is a turning point for the humanoid robotics industry. Most rivals, from Agility Robotics to Figure AI and Boston Dynamics, are still operating in pilot or low-volume phases, and the sector has faced scrutiny over the gap between announced orders and real deployments. A factory engineered for a million units a year signals that at least one major manufacturer believes the cost curve for humanoids can follow the path of electric vehicles.
The caveats are substantial. Tesla's own timeline puts commercial deliveries more than a year away, the hardware is still being refined, and general-purpose autonomy remains unproven. Musk's past forecasts for Optimus and self-driving have frequently slipped. Still, manufacturing at scale is where many robotics ventures stall, and Tesla has deep experience there. If the Fremont ramp holds, the debate over humanoid robots will shift from whether they can be built to whether they can do useful work reliably enough to justify the volume.
