Tesla can reportedly build hundreds of Optimus robots a week — it can't yet give them a general job

Tesla's Optimus production went from a few dozen units a week during small-batch testing in the second quarter to several hundred a week in August, The Information reported, a roughly tenfold increase. Managers are aiming for a continuous automated line capable of more than 1,000 robots a week by the end of the year, and eventually about 20,000 a week. The reporting also described what the ramp has not solved. The hand and forearm contain more than 100 screws and small components and are still assembled manually. Suppliers have delivered misaligned parts. Most units are used inside Tesla for testing, training and data collection; the ones working in its factories are confined to tightly controlled, supervised areas and programmed for specific tasks rather than running as general-purpose machines. Tesla acknowledged the robots remain far from handling generalised tasks. Separately, Ars Technica reported that some Tesla workers have balked at training the robots that are meant to replace them. The tracker's Optimus entry carries the figures, marked as reported rather than published by Tesla.

Why it matters

A humanoid business needs two curves to cross: how many robots you can build, and how many useful things each one can do. For most of the industry the first curve has been the hard one. Tesla's reported numbers suggest it is solving manufacturing faster than anyone — hundreds a week is more than most humanoid companies have built in total — while the second curve has barely moved. Building a robot that cannot yet hold a general job faster does not make it a general robot; it makes a larger fleet waiting for one.

That gap has a practical consequence for how to read the output figure. A robot that spends its life in testing and data collection is not idle — the data may be exactly what closes the capability gap — but it is not a unit of labour, and it should not be counted as one when the rate is compared with a car line or with rivals' deployments. The honest unit for humanoids right now is robots doing productive work outside a supervised pen, and nobody, Tesla included, is publishing that number.

The hand detail is the other thing worth keeping. A component with over a hundred parts that is still assembled by people sits on the critical path of an "automated line" targeting 20,000 a week. Until the hand is automated, the line's rate is the hand's rate.

What to watch

Whether Tesla publishes an Optimus production or deployment figure itself, when the hand assembly is automated, and the first report of Optimus doing unsupervised productive work at a customer site.