Tesla's Optimus Ramp Reaches Hundreds Per Week, With Hands as the Bottleneck
Tesla produced several hundred Optimus robots per week last month as it pushes toward a stated goal of 20,000 units per week, with hand assembly and parts fit now pacing the ramp. The Verge
The current build rate follows a repurposing earlier this year of Tesla's Model S and Model X production lines for Optimus production. That conversion gave Tesla floor space and line infrastructure for humanoids, but the manufacturing problems since have been specific to humanoids rather than carryover automotive process issues.
Hands are the constraint. The V3 Optimus hands and forearms contain more than 100 screws and other small components that currently require manual assembly by a worker. Tesla is also lining up parts precisely at a scale its vehicle operations did not have to manage, because Optimus parts are much smaller and must fit together more precisely than car parts.
V3 hardware is lighter and carries more cameras than previous versions. It still lacks a sensing glove under development that would improve durability by allowing touch sensors to be replaced without swapping out the entire hand. That missing subassembly matters for serviceability. A hand replacement is a field swap of a complex end-effector. A glove replacement would be a lower-level service operation.
Deployment remains narrow. Tesla's Optimus robots are currently used internally for specific tasks in limited areas, not for general use or with remote teleoperation. The plan, as reported, is to initially lease Optimus robots rather than sell them and use collected training data to improve the robots. For practitioners, the implication is a closed fleet learning loop before any open distribution.
From small batches to line production
During the second quarter, Tesla produced a few dozen Optimus robots per week while conducting small batch testing, according to background reporting from The Information. The step to several hundred per week last month indicates line processes at its Fremont factory have moved beyond bench builds, where Tesla says it develops manufacturing lines and processes for Optimus.
That transition has not been linear. Tesla halted purchasing components for the Optimus humanoid robot while making changes to the robot's hardware and software, a pause reported in July 2025. The Information Separate September 2026 coverage cited a report saying Tesla Optimus robot hands break and suppliers can't keep up with production. Electrek
Hiring points to the same bottleneck. Tesla lists a Manufacturing Engineer, Optimus Hands Launch role to lead the launch and ramp of production for Optimus humanoid robot hands, and a Robotics Manufacturing Engineer role to launch production for Optimus humanoid robots involving Full Bot Validation Testing.
Personnel has also shifted. Milan Kovac, the head of Tesla's Optimus humanoid robot program, announced plans to step down in a post on X in June 2025. Reuters
Timelines and product definition
Elon Musk has said Optimus will be Tesla's biggest product of all time. Public timelines for external availability have moved. In April 2024, Musk said Tesla could start selling Optimus robots by the end of next year. Reuters In January 2026, Musk said Tesla will probably sell its Optimus robots to the public by the end of next year. Bloomberg The current reported plan for initial leases rather than sales would keep early units under Tesla control while training data accumulates.
Tesla defines its Optimus robot as a "general-purpose, bi-pedal, autonomous humanoid robot capable of performing unsafe, repetitive or boring tasks." Tesla Current internal use for specific tasks in limited areas is consistent with that definition as an aspiration rather than a present capability claim. No general-purpose deployment or teleoperated operation is described in the recent reporting.
The broader context here is a manufacturing regime change, not simply a volume ramp. Automotive assembly tolerates millimeter variation across large stampings and castings with compliance built into fastening and sealing. Dexterous hands invert that logic. Tolerance stack-up across dozens of fasteners, miniature bearings, tendon routings and sensor flexes compounds quickly, and manual assembly hides variation that automated stations will later have to control or design out.
In my view, the lease-first detail is the most telling operational choice. It keeps failure analysis, touch-sensor wear data and grasp failure cases inside Tesla's engineering loop while avoiding a service network for a system whose most failure-prone subsystem, the hand, still requires manual build and full-hand replacement. The gap is large. Several hundred per week to 20,000 per week is nearly two orders of magnitude, and that multiple will not close through labor scaling.
Worth flagging for humanoid teams watching this: Tesla's problem set now looks familiar to anyone who has shipped precision mechatronics. Design for assembly, fastener count reduction, modular sensor packaging, and supplier process control for micro-mechanical parts tend to decide schedules more than policy models or fleet size. If Tesla lands the sensing glove, automates hand build, and holds supplier yield, leased fleets doing constrained tasks could generate the manipulation datasets that general-purpose claims will eventually need. That outcome would leave Optimus as factory equipment first and product second, which over the long arc is still how useful robots usually reach the public.
