Product2 publishers3 min readPublished
Optimus production reaches hundreds a week, far short of Tesla's 20,000 goal
Tesla produced several hundred Optimus robots a week last month, The Information reported, against a stated goal of about 20,000. Anyone budgeting a humanoid pilot for 2027 should plan on a small number of robots, each doing one closely supervised job.
The Product Desk · Product desk
What happened
- Each Optimus hand and forearm contains more than 100 screws and other small parts that must be assembled manually, one person familiar with the matter told The Information.
- Equipment at several stations, including those building hands and joints, does not always line up parts precisely, so more robots come off the line needing repair.
- Tesla has developed a replaceable sensing glove that holds the hand's touch sensors and plans to put it in next year's model.
- Tesla is targeting more than 1,000 robots a week by the end of 2026, according to Ars Technica's account of The Information's reporting.
- Tesla plans to lease the robots to commercial customers first, and Musk has said sales could begin by late 2027.
Compiled by The Product DeskSomething wrong?How this is made
Why it matters
- cost Setup cost grows with the length of each job's task list, since every distinct task adds several days of training before the robot does useful work.
- decision Early lessees have to decide whether to take hands built before the replaceable sensing glove arrives or wait for next year's model.
- exposure A lease turns the customer's factory or warehouse into a source of Tesla training data, so data-use terms belong in the contract review.
Inside a Tesla factory, an Optimus robot works in what The Information described as "tightly controlled areas where their movements and tasks can be closely supervised" [10]. Most of the units Tesla builds stay in-house, where they are used for testing and training [10]. Each job has to be programmed and trained separately, and basic tasks take several days to learn, people familiar with the matter told The Information [9].
The pitch is much broader. Musk has described Optimus as a general-purpose assistant that could eventually do household chores and even babysit [13]. On Tesla's second-quarter 2026 earnings call he said it could be "the biggest product ever" [11]. On the same call he acknowledged that a humanoid able to handle many different tasks is "one of the hardest things to solve" [12].
The volume gap is wide on any reading. If "several hundred" means 200 to 500 a week [1], the 20,000 goal [2] is 40 to 100 times current output [1]. The nearer target of more than 1,000 a week by the end of 2026 is one-twentieth of the goal [2]. It is also two to five times what the line built last month [3]. Output did climb from a few dozen a week only a few months earlier [1].
The hands are one constraint among several. The Information also describes limits on how fast the line can run [6]. Once built, the hands are not durable enough for sustained use, and some touch sensors have had reliability problems [7]. The reporting does not say how much of the production gap the hands account for.
People are part of the ramp too. Tesla ended Model S and Model X production at Fremont in May 2026 and moved line workers and engineers onto Optimus [14]. Some employees are pushing back against training their robotic replacements, Ars Technica reported [17].
Under the lease model, Tesla plans to use data from robots working in customers' factories and warehouses to improve its AI systems [16]. Tesla did not immediately respond to Gizmodo's request for comment [19].
For an operator weighing a pilot, I'd sort candidate jobs on two axes. One is how many distinct tasks the job contains. The other is whether the robot can work in a controlled, supervised area or has to share open floor with people. On the current reporting, Optimus fits one quadrant: a single task in a supervised area [9][10]. Every other quadrant multiplies several days of training by the number of tasks [9]. It also assumes hands that hold up, when Tesla's replaceable sensor glove is planned only for next year's model [8]. The lease adds a separate check on what floor data leaves the building [16].
What to watch
- Whether Tesla's weekly Optimus output passes 1,000 by the end of 2026, the first public test of the ramp beyond several hundred.
- Whether next year's Optimus model ships with the replaceable sensing glove, and whether hand durability improves with it.
- The first commercial lease terms, especially how Tesla's plan to train its AI on customer-site data is written into contracts.