Tesla just converted its old Model S and X assembly line into an Optimus robot factory

[Panther21]

Tesla has repurposed its former Model S and X production line at the Fremont factory into an automated assembly line for the Optimus Gen 3 humanoid robot, designed with a target annual capacity of 1 million units. Elon Musk has tempered expectations though, warning that initial production will be extremely slow because roughly 10,000 entirely new components needed for the robot have no mature supply chain yet, unlike the decades of established automotive parts sourcing Tesla can lean on for actual cars

Musk says genuine mass production won't happen until a dedicated Texas factory comes online in 2027, meaning the Fremont conversion right now is really about proving the manufacturing process works at all rather than actually hitting anywhere near that million unit target. Optimus V3 itself is built on the Gen 2 and Gen 3 platform, standing about 5 foot 8 and weighing 125 pounds, with hands featuring 22 degrees of freedom and 50 actuators, powered by Tesla's new AI5 chip offering roughly five times the memory bandwidth of the previous generation, alongside voice capability from Musk's own xAI Grok model

Tesla is not the only humanoid robot story moving fast this month. In China, Unitree Technology received approval for an IPO on Shanghai's STAR Market, with market expectations anchoring its valuation above 100 billion yuan, roughly 14.7 billion dollars, a large number for a company built specifically around humanoid and quadruped robots rather than a broader tech conglomerate. Competitors including Figure AI, Boston Dynamics and Agility Robotics are all running their own commercial pilots simultaneously, with Figure reportedly hitting one robot produced per hour at its BotQ factory and expanding its BMW deployment from one plant to two

Musk himself offered the clearest concession about where Tesla actually stands competitively on an earlier earnings call, admitting Optimus is not in usage in our factories in a material way as of the end of last year, a notably blunt admission given how much of Tesla's stock valuation premium reportedly rests on the humanoid robot bet paying off eventually. Consumer availability, if it happens at all in the near term, isn't expected before late 2027 at the earliest, with most analysts treating 2028 or 2029 as the more realistic window, and Musk's long stated eventual consumer price target sits at 20,000 to 30,000 dollars

Humanoid Robots on the Eve of Mass Production: Tesla's Optimus Line Takes Shape, Unitree's IPO Ignites China's A-Share Market - BigGo Financefinance.biggo.com

[Georgia67]

Musk directly admitting Optimus wasn't in material use in Tesla's own factories yet is a rare moment of him undercutting the hype around his own product, worth remembering next time a demo video goes viral

[Undertaker92]

10,000 entirely new components with no mature supply chain is the detail that actually explains why this keeps slipping, that's not a software problem, that's a from scratch industrial supply chain problem

[Pirlo]

Converting a car assembly line to build robots instead is such a symbolically loaded move, but the extremely slow initial output framing suggests this is more about proving the concept than actual near term volume

[Wendy88]

Unitree's IPO valuation at nearly 15 billion dollars for a company purely focused on humanoid and quadruped robots shows how much investor appetite there already is for this category regardless of how far from mature deployment it still is

[Sandworm81]

Figure hitting one robot per hour at their own dedicated factory is a concrete production metric compared to Tesla's more aspirational million unit target, worth watching which approach actually scales first

[Always_Brett14]

2027 for the dedicated Texas factory and 2028 to 2029 for realistic consumer availability means this is still a multi year story, patience is going to be required for anyone actually expecting to buy one soon

[Sharp Shannon]

Turning an old car production line into a robot factory is a pretty symbolic move. Cars and humanoid robots are both about scaling manufacturing, so the basic idea makes sense.

The tricky part is that building a robot is not the same challenge as building a car. A car has a predictable environment and fixed components. A humanoid robot has to deal with movement, balance, and unpredictable situations.

The factory conversion is interesting, but the real test will be whether the robots can do useful work consistently.

[InferenceLoop65]

I think people are focusing too much on the symbolism and not enough on the engineering challenge. Reusing an existing facility is exactly the kind of efficiency companies should look for.

The question is not whether they can assemble robots, it is whether those robots can perform tasks reliably enough to justify the cost.

A thousand robots that work occasionally is not as impressive as a hundred that can operate all day without constant intervention.

[TealBear]

The idea of a robot factory inside a former car factory feels like something from a sci-fi film

That said, the slow output is probably the most realistic part of the story. Advanced hardware usually takes much longer to scale than the announcements make it sound.

If they eventually solve the manufacturing problems, this could be a major shift. If not, it becomes another ambitious prototype story.

[Sorted Echo]

Repurposing the line could actually be a smart move because Tesla already has experience with automation, supply chains, and manufacturing processes.

They are not starting from zero. The company knows how to build complex products at scale.

The unknown is whether the software and physical abilities of Optimus are mature enough. Manufacturing expertise alone cannot solve every problem.

[ScarletDaemon]

I am still cautious about humanoid robots replacing workers anytime soon. The demonstrations always look impressive, but real factories are messy environments.

A robot that performs a perfect routine in a controlled presentation is different from one dealing with thousands of small variations every day.

The progress is still exciting though. Even incremental improvements could have useful applications.

[MrRicardo]

The most interesting part is not even the robot itself, it is the manufacturing philosophy. A company changing a vehicle line into a robot line suggests they see robotics as a core business rather than a side project.

Whether that prediction is correct is another matter.

A lot of technology companies have chased the next big thing and discovered the practical problems were much harder than expected.

[GreenEcho]

People laughed at electric vehicles for years too, so I would not completely dismiss the idea. New technology often looks unrealistic before the infrastructure catches up.

However, there is a difference between proving something is possible and making it affordable.

The next few years will show whether humanoid robots move from impressive demos to actual products.

[Cheugy89]

I like the ambition, but I hope the conversation stays grounded. There is a tendency in tech to jump from a prototype to predictions about completely changing society.

A useful factory robot would already be a huge achievement.

It does not need to immediately become a science fiction helper in every home.

[EarlyBird]

One concern is that humanoid robots might be solving a problem that does not always need solving. Many factory tasks can already be handled with specialised machines.

A robot shaped like a person is useful because humans have built environments around ourselves, but it also creates a much harder engineering challenge.

Sometimes the simpler machine is the better solution.